feat: Add webhook trigger test and snapshot

Co-authored-by: nicholas.tindle <nicholas.tindle@agpt.co>

.branchlet.json

@@ -1,37 +0,0 @@
-{
-  "worktreeCopyPatterns": [
-    ".env*",
-    ".vscode/**",
-    ".auth/**",
-    ".claude/**",
-    "autogpt_platform/.env*",
-    "autogpt_platform/backend/.env*",
-    "autogpt_platform/frontend/.env*",
-    "autogpt_platform/frontend/.auth/**",
-    "autogpt_platform/db/docker/.env*"
-  ],
-  "worktreeCopyIgnores": [
-    "**/node_modules/**",
-    "**/dist/**",
-    "**/.git/**",
-    "**/Thumbs.db",
-    "**/.DS_Store",
-    "**/.next/**",
-    "**/__pycache__/**",
-    "**/.ruff_cache/**",
-    "**/.pytest_cache/**",
-    "**/*.pyc",
-    "**/playwright-report/**",
-    "**/logs/**",
-    "**/site/**"
-  ],
-  "worktreePathTemplate": "$BASE_PATH.worktree",
-  "postCreateCmd": [
-    "cd autogpt_platform/autogpt_libs && poetry install",
-    "cd autogpt_platform/backend && poetry install && poetry run prisma generate",
-    "cd autogpt_platform/frontend && pnpm install",
-    "cd docs && pip install -r requirements.txt"
-  ],
-  "terminalCommand": "code .",
-  "deleteBranchWithWorktree": false
-}

.dockerignore

@@ -1,9 +1,6 @@
 # Ignore everything by default, selectively add things to context
 *
 
-# Documentation (for embeddings/search)
-!docs/
-
 # Platform - Libs
 !autogpt_platform/autogpt_libs/autogpt_libs/
 !autogpt_platform/autogpt_libs/pyproject.toml
@@ -19,7 +16,6 @@
 !autogpt_platform/backend/poetry.lock
 !autogpt_platform/backend/README.md
 !autogpt_platform/backend/.env
-!autogpt_platform/backend/gen_prisma_types_stub.py
 
 # Platform - Market
 !autogpt_platform/market/market/

.github/workflows/claude-dependabot.yml

@@ -74,7 +74,7 @@ jobs:
 
       - name: Generate Prisma Client
         working-directory: autogpt_platform/backend
-        run: poetry run prisma generate && poetry run gen-prisma-stub
+        run: poetry run prisma generate
 
       # Frontend Node.js/pnpm setup (mirrors platform-frontend-ci.yml)
       - name: Set up Node.js

.github/workflows/claude.yml

@@ -90,7 +90,7 @@ jobs:
 
       - name: Generate Prisma Client
         working-directory: autogpt_platform/backend
-        run: poetry run prisma generate && poetry run gen-prisma-stub
+        run: poetry run prisma generate
 
       # Frontend Node.js/pnpm setup (mirrors platform-frontend-ci.yml)
       - name: Set up Node.js

.github/workflows/copilot-setup-steps.yml

@@ -72,7 +72,7 @@ jobs:
 
       - name: Generate Prisma Client
         working-directory: autogpt_platform/backend
-        run: poetry run prisma generate && poetry run gen-prisma-stub
+        run: poetry run prisma generate
 
       # Frontend Node.js/pnpm setup (mirrors platform-frontend-ci.yml)
       - name: Set up Node.js
@@ -108,16 +108,6 @@ jobs:
       #   run: pnpm playwright install --with-deps chromium
 
       # Docker setup for development environment
-      - name: Free up disk space
-        run: |
-          # Remove large unused tools to free disk space for Docker builds
-          sudo rm -rf /usr/share/dotnet
-          sudo rm -rf /usr/local/lib/android
-          sudo rm -rf /opt/ghc
-          sudo rm -rf /opt/hostedtoolcache/CodeQL
-          sudo docker system prune -af
-          df -h
-
       - name: Set up Docker Buildx
         uses: docker/setup-buildx-action@v3
 

.github/workflows/platform-backend-ci.yml

@@ -134,7 +134,7 @@ jobs:
         run: poetry install
 
       - name: Generate Prisma Client
-        run: poetry run prisma generate && poetry run gen-prisma-stub
+        run: poetry run prisma generate
 
       - id: supabase
         name: Start Supabase
@@ -176,7 +176,7 @@ jobs:
           }
 
       - name: Run Database Migrations
-        run: poetry run prisma migrate deploy
+        run: poetry run prisma migrate dev --name updates
         env:
           DATABASE_URL: ${{ steps.supabase.outputs.DB_URL }}
           DIRECT_URL: ${{ steps.supabase.outputs.DB_URL }}

.github/workflows/platform-frontend-ci.yml

@@ -11,7 +11,6 @@ on:
       - ".github/workflows/platform-frontend-ci.yml"
       - "autogpt_platform/frontend/**"
   merge_group:
-  workflow_dispatch:
 
 concurrency:
   group: ${{ github.workflow }}-${{ github.event_name == 'merge_group' && format('merge-queue-{0}', github.ref) || format('{0}-{1}', github.ref, github.event.pull_request.number || github.sha) }}
@@ -152,14 +151,6 @@ jobs:
         run: |
           cp ../.env.default ../.env
 
-      - name: Copy backend .env and set OpenAI API key
-        run: |
-          cp ../backend/.env.default ../backend/.env
-          echo "OPENAI_INTERNAL_API_KEY=${{ secrets.OPENAI_API_KEY }}" >> ../backend/.env
-        env:
-          # Used by E2E test data script to generate embeddings for approved store agents
-          OPENAI_API_KEY: ${{ secrets.OPENAI_API_KEY }}
-
       - name: Set up Docker Buildx
         uses: docker/setup-buildx-action@v3
 
@@ -235,25 +226,13 @@ jobs:
 
       - name: Run Playwright tests
         run: pnpm test:no-build
-        continue-on-error: false
 
-      - name: Upload Playwright report
-        if: always()
+      - name: Upload Playwright artifacts
+        if: failure()
         uses: actions/upload-artifact@v4
         with:
           name: playwright-report
           path: playwright-report
-          if-no-files-found: ignore
-          retention-days: 3
-
-      - name: Upload Playwright test results
-        if: always()
-        uses: actions/upload-artifact@v4
-        with:
-          name: playwright-test-results
-          path: test-results
-          if-no-files-found: ignore
-          retention-days: 3
 
       - name: Print Final Docker Compose logs
         if: always()

.github/workflows/repo-close-stale-issues.yml

@@ -11,7 +11,7 @@ jobs:
   stale:
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/stale@v10
+      - uses: actions/stale@v9
         with:
           # operations-per-run: 5000
           stale-issue-message: >

.github/workflows/repo-pr-label.yml

@@ -61,6 +61,6 @@ jobs:
       pull-requests: write
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/labeler@v6
+      - uses: actions/labeler@v5
         with:
           sync-labels: true

autogpt_platform/Makefile

@@ -6,14 +6,12 @@ start-core:
 
 # Stop core services
 stop-core:
-	docker compose stop 
+	docker compose stop deps
 
 reset-db:
-	docker compose stop db
 	rm -rf db/docker/volumes/db/data
 	cd backend && poetry run prisma migrate deploy
 	cd backend && poetry run prisma generate
-	cd backend && poetry run gen-prisma-stub
 	
 # View logs for core services
 logs-core:
@@ -35,7 +33,6 @@ init-env:
 migrate:
 	cd backend && poetry run prisma migrate deploy
 	cd backend && poetry run prisma generate
-	cd backend && poetry run gen-prisma-stub
 
 run-backend:
 	cd backend && poetry run app
@@ -61,4 +58,4 @@ help:
 	@echo "  run-backend - Run the backend FastAPI server"
 	@echo "  run-frontend - Run the frontend Next.js development server"
 	@echo "  test-data - Run the test data creator"
-	@echo "  load-store-agents - Load store agents from agents/ folder into test database"
+	@echo "  load-store-agents - Load store agents from agents/ folder into test database"

autogpt_platform/autogpt_libs/autogpt_libs/api_key/keysmith.py

@@ -57,9 +57,6 @@ class APIKeySmith:
 
     def hash_key(self, raw_key: str) -> tuple[str, str]:
         """Migrate a legacy hash to secure hash format."""
-        if not raw_key.startswith(self.PREFIX):
-            raise ValueError("Key without 'agpt_' prefix would fail validation")
-
         salt = self._generate_salt()
         hash = self._hash_key_with_salt(raw_key, salt)
         return hash, salt.hex()

autogpt_platform/autogpt_libs/autogpt_libs/auth/helpers.py

@@ -1,25 +1,29 @@
 from fastapi import FastAPI
+from fastapi.openapi.utils import get_openapi
 
 from .jwt_utils import bearer_jwt_auth
 
 
 def add_auth_responses_to_openapi(app: FastAPI) -> None:
     """
-    Patch a FastAPI instance's `openapi()` method to add 401 responses
+    Set up custom OpenAPI schema generation that adds 401 responses
     to all authenticated endpoints.
 
     This is needed when using HTTPBearer with auto_error=False to get proper
     401 responses instead of 403, but FastAPI only automatically adds security
     responses when auto_error=True.
     """
-    # Wrap current method to allow stacking OpenAPI schema modifiers like this
-    wrapped_openapi = app.openapi
 
     def custom_openapi():
         if app.openapi_schema:
             return app.openapi_schema
 
-        openapi_schema = wrapped_openapi()
+        openapi_schema = get_openapi(
+            title=app.title,
+            version=app.version,
+            description=app.description,
+            routes=app.routes,
+        )
 
         # Add 401 response to all endpoints that have security requirements
         for path, methods in openapi_schema["paths"].items():

autogpt_platform/backend/.env.default

@@ -58,13 +58,6 @@ V0_API_KEY=
 OPEN_ROUTER_API_KEY=
 NVIDIA_API_KEY=
 
-# Langfuse Prompt Management
-# Used for managing the CoPilot system prompt externally
-# Get credentials from https://cloud.langfuse.com or your self-hosted instance
-LANGFUSE_PUBLIC_KEY=
-LANGFUSE_SECRET_KEY=
-LANGFUSE_HOST=https://cloud.langfuse.com
-
 # OAuth Credentials
 # For the OAuth callback URL, use <your_frontend_url>/auth/integrations/oauth_callback,
 # e.g. http://localhost:3000/auth/integrations/oauth_callback

autogpt_platform/backend/.gitignore

@@ -18,4 +18,3 @@ load-tests/results/
 load-tests/*.json
 load-tests/*.log
 load-tests/node_modules/*
-migrations/*/rollback*.sql

autogpt_platform/backend/Dockerfile

@@ -48,8 +48,7 @@ RUN poetry install --no-ansi --no-root
 # Generate Prisma client
 COPY autogpt_platform/backend/schema.prisma ./
 COPY autogpt_platform/backend/backend/data/partial_types.py ./backend/data/partial_types.py
-COPY autogpt_platform/backend/gen_prisma_types_stub.py ./
-RUN poetry run prisma generate && poetry run gen-prisma-stub
+RUN poetry run prisma generate
 
 FROM debian:13-slim AS server_dependencies
 
@@ -100,7 +99,6 @@ COPY autogpt_platform/backend/migrations /app/autogpt_platform/backend/migration
 FROM server_dependencies AS server
 
 COPY autogpt_platform/backend /app/autogpt_platform/backend
-COPY docs /app/docs
 RUN poetry install --no-ansi --only-root
 
 ENV PORT=8000

autogpt_platform/backend/TESTING.md

@@ -108,7 +108,7 @@ import fastapi.testclient
 import pytest
 from pytest_snapshot.plugin import Snapshot
 
-from backend.api.features.myroute import router
+from backend.server.v2.myroute import router
 
 app = fastapi.FastAPI()
 app.include_router(router)
@@ -149,7 +149,7 @@ These provide the easiest way to set up authentication mocking in test modules:
 import fastapi
 import fastapi.testclient
 import pytest
-from backend.api.features.myroute import router
+from backend.server.v2.myroute import router
 
 app = fastapi.FastAPI()
 app.include_router(router)

autogpt_platform/backend/backend/api/external/fastapi_app.py

@@ -1,25 +0,0 @@
-from fastapi import FastAPI
-
-from backend.api.middleware.security import SecurityHeadersMiddleware
-from backend.monitoring.instrumentation import instrument_fastapi
-
-from .v1.routes import v1_router
-
-external_api = FastAPI(
-    title="AutoGPT External API",
-    description="External API for AutoGPT integrations",
-    docs_url="/docs",
-    version="1.0",
-)
-
-external_api.add_middleware(SecurityHeadersMiddleware)
-external_api.include_router(v1_router, prefix="/v1")
-
-# Add Prometheus instrumentation
-instrument_fastapi(
-    external_api,
-    service_name="external-api",
-    expose_endpoint=True,
-    endpoint="/metrics",
-    include_in_schema=True,
-)

autogpt_platform/backend/backend/api/external/middleware.py

@@ -1,107 +0,0 @@
-from fastapi import HTTPException, Security, status
-from fastapi.security import APIKeyHeader, HTTPAuthorizationCredentials, HTTPBearer
-from prisma.enums import APIKeyPermission
-
-from backend.data.auth.api_key import APIKeyInfo, validate_api_key
-from backend.data.auth.base import APIAuthorizationInfo
-from backend.data.auth.oauth import (
-    InvalidClientError,
-    InvalidTokenError,
-    OAuthAccessTokenInfo,
-    validate_access_token,
-)
-
-api_key_header = APIKeyHeader(name="X-API-Key", auto_error=False)
-bearer_auth = HTTPBearer(auto_error=False)
-
-
-async def require_api_key(api_key: str | None = Security(api_key_header)) -> APIKeyInfo:
-    """Middleware for API key authentication only"""
-    if api_key is None:
-        raise HTTPException(
-            status_code=status.HTTP_401_UNAUTHORIZED, detail="Missing API key"
-        )
-
-    api_key_obj = await validate_api_key(api_key)
-
-    if not api_key_obj:
-        raise HTTPException(
-            status_code=status.HTTP_401_UNAUTHORIZED, detail="Invalid API key"
-        )
-
-    return api_key_obj
-
-
-async def require_access_token(
-    bearer: HTTPAuthorizationCredentials | None = Security(bearer_auth),
-) -> OAuthAccessTokenInfo:
-    """Middleware for OAuth access token authentication only"""
-    if bearer is None:
-        raise HTTPException(
-            status_code=status.HTTP_401_UNAUTHORIZED,
-            detail="Missing Authorization header",
-        )
-
-    try:
-        token_info, _ = await validate_access_token(bearer.credentials)
-    except (InvalidClientError, InvalidTokenError) as e:
-        raise HTTPException(status_code=status.HTTP_401_UNAUTHORIZED, detail=str(e))
-
-    return token_info
-
-
-async def require_auth(
-    api_key: str | None = Security(api_key_header),
-    bearer: HTTPAuthorizationCredentials | None = Security(bearer_auth),
-) -> APIAuthorizationInfo:
-    """
-    Unified authentication middleware supporting both API keys and OAuth tokens.
-
-    Supports two authentication methods, which are checked in order:
-    1. X-API-Key header (existing API key authentication)
-    2. Authorization: Bearer <token> header (OAuth access token)
-
-    Returns:
-        APIAuthorizationInfo: base class of both APIKeyInfo and OAuthAccessTokenInfo.
-    """
-    # Try API key first
-    if api_key is not None:
-        api_key_info = await validate_api_key(api_key)
-        if api_key_info:
-            return api_key_info
-        raise HTTPException(
-            status_code=status.HTTP_401_UNAUTHORIZED, detail="Invalid API key"
-        )
-
-    # Try OAuth bearer token
-    if bearer is not None:
-        try:
-            token_info, _ = await validate_access_token(bearer.credentials)
-            return token_info
-        except (InvalidClientError, InvalidTokenError) as e:
-            raise HTTPException(status_code=status.HTTP_401_UNAUTHORIZED, detail=str(e))
-
-    # No credentials provided
-    raise HTTPException(
-        status_code=status.HTTP_401_UNAUTHORIZED,
-        detail="Missing authentication. Provide API key or access token.",
-    )
-
-
-def require_permission(permission: APIKeyPermission):
-    """
-    Dependency function for checking specific permissions
-    (works with API keys and OAuth tokens)
-    """
-
-    async def check_permission(
-        auth: APIAuthorizationInfo = Security(require_auth),
-    ) -> APIAuthorizationInfo:
-        if permission not in auth.scopes:
-            raise HTTPException(
-                status_code=status.HTTP_403_FORBIDDEN,
-                detail=f"Missing required permission: {permission.value}",
-            )
-        return auth
-
-    return check_permission

autogpt_platform/backend/backend/api/features/analytics_test.py

@@ -1,340 +0,0 @@
-"""Tests for analytics API endpoints."""
-
-import json
-from unittest.mock import AsyncMock, Mock
-
-import fastapi
-import fastapi.testclient
-import pytest
-import pytest_mock
-from pytest_snapshot.plugin import Snapshot
-
-from .analytics import router as analytics_router
-
-app = fastapi.FastAPI()
-app.include_router(analytics_router)
-
-client = fastapi.testclient.TestClient(app)
-
-
-@pytest.fixture(autouse=True)
-def setup_app_auth(mock_jwt_user):
-    """Setup auth overrides for all tests in this module."""
-    from autogpt_libs.auth.jwt_utils import get_jwt_payload
-
-    app.dependency_overrides[get_jwt_payload] = mock_jwt_user["get_jwt_payload"]
-    yield
-    app.dependency_overrides.clear()
-
-
-# =============================================================================
-# /log_raw_metric endpoint tests
-# =============================================================================
-
-
-def test_log_raw_metric_success(
-    mocker: pytest_mock.MockFixture,
-    configured_snapshot: Snapshot,
-    test_user_id: str,
-) -> None:
-    """Test successful raw metric logging."""
-    mock_result = Mock(id="metric-123-uuid")
-    mock_log_metric = mocker.patch(
-        "backend.data.analytics.log_raw_metric",
-        new_callable=AsyncMock,
-        return_value=mock_result,
-    )
-
-    request_data = {
-        "metric_name": "page_load_time",
-        "metric_value": 2.5,
-        "data_string": "/dashboard",
-    }
-
-    response = client.post("/log_raw_metric", json=request_data)
-
-    assert response.status_code == 200, f"Unexpected response: {response.text}"
-    assert response.json() == "metric-123-uuid"
-
-    mock_log_metric.assert_called_once_with(
-        user_id=test_user_id,
-        metric_name="page_load_time",
-        metric_value=2.5,
-        data_string="/dashboard",
-    )
-
-    configured_snapshot.assert_match(
-        json.dumps({"metric_id": response.json()}, indent=2, sort_keys=True),
-        "analytics_log_metric_success",
-    )
-
-
-@pytest.mark.parametrize(
-    "metric_value,metric_name,data_string,test_id",
-    [
-        (100, "api_calls_count", "external_api", "integer_value"),
-        (0, "error_count", "no_errors", "zero_value"),
-        (-5.2, "temperature_delta", "cooling", "negative_value"),
-        (1.23456789, "precision_test", "float_precision", "float_precision"),
-        (999999999, "large_number", "max_value", "large_number"),
-        (0.0000001, "tiny_number", "min_value", "tiny_number"),
-    ],
-)
-def test_log_raw_metric_various_values(
-    mocker: pytest_mock.MockFixture,
-    configured_snapshot: Snapshot,
-    metric_value: float,
-    metric_name: str,
-    data_string: str,
-    test_id: str,
-) -> None:
-    """Test raw metric logging with various metric values."""
-    mock_result = Mock(id=f"metric-{test_id}-uuid")
-    mocker.patch(
-        "backend.data.analytics.log_raw_metric",
-        new_callable=AsyncMock,
-        return_value=mock_result,
-    )
-
-    request_data = {
-        "metric_name": metric_name,
-        "metric_value": metric_value,
-        "data_string": data_string,
-    }
-
-    response = client.post("/log_raw_metric", json=request_data)
-
-    assert response.status_code == 200, f"Failed for {test_id}: {response.text}"
-
-    configured_snapshot.assert_match(
-        json.dumps(
-            {"metric_id": response.json(), "test_case": test_id},
-            indent=2,
-            sort_keys=True,
-        ),
-        f"analytics_metric_{test_id}",
-    )
-
-
-@pytest.mark.parametrize(
-    "invalid_data,expected_error",
-    [
-        ({}, "Field required"),
-        ({"metric_name": "test"}, "Field required"),
-        (
-            {"metric_name": "test", "metric_value": "not_a_number", "data_string": "x"},
-            "Input should be a valid number",
-        ),
-        (
-            {"metric_name": "", "metric_value": 1.0, "data_string": "test"},
-            "String should have at least 1 character",
-        ),
-        (
-            {"metric_name": "test", "metric_value": 1.0, "data_string": ""},
-            "String should have at least 1 character",
-        ),
-    ],
-    ids=[
-        "empty_request",
-        "missing_metric_value_and_data_string",
-        "invalid_metric_value_type",
-        "empty_metric_name",
-        "empty_data_string",
-    ],
-)
-def test_log_raw_metric_validation_errors(
-    invalid_data: dict,
-    expected_error: str,
-) -> None:
-    """Test validation errors for invalid metric requests."""
-    response = client.post("/log_raw_metric", json=invalid_data)
-
-    assert response.status_code == 422
-    error_detail = response.json()
-    assert "detail" in error_detail, f"Missing 'detail' in error: {error_detail}"
-
-    error_text = json.dumps(error_detail)
-    assert (
-        expected_error in error_text
-    ), f"Expected '{expected_error}' in error response: {error_text}"
-
-
-def test_log_raw_metric_service_error(
-    mocker: pytest_mock.MockFixture,
-    test_user_id: str,
-) -> None:
-    """Test error handling when analytics service fails."""
-    mocker.patch(
-        "backend.data.analytics.log_raw_metric",
-        new_callable=AsyncMock,
-        side_effect=Exception("Database connection failed"),
-    )
-
-    request_data = {
-        "metric_name": "test_metric",
-        "metric_value": 1.0,
-        "data_string": "test",
-    }
-
-    response = client.post("/log_raw_metric", json=request_data)
-
-    assert response.status_code == 500
-    error_detail = response.json()["detail"]
-    assert "Database connection failed" in error_detail["message"]
-    assert "hint" in error_detail
-
-
-# =============================================================================
-# /log_raw_analytics endpoint tests
-# =============================================================================
-
-
-def test_log_raw_analytics_success(
-    mocker: pytest_mock.MockFixture,
-    configured_snapshot: Snapshot,
-    test_user_id: str,
-) -> None:
-    """Test successful raw analytics logging."""
-    mock_result = Mock(id="analytics-789-uuid")
-    mock_log_analytics = mocker.patch(
-        "backend.data.analytics.log_raw_analytics",
-        new_callable=AsyncMock,
-        return_value=mock_result,
-    )
-
-    request_data = {
-        "type": "user_action",
-        "data": {
-            "action": "button_click",
-            "button_id": "submit_form",
-            "timestamp": "2023-01-01T00:00:00Z",
-            "metadata": {"form_type": "registration", "fields_filled": 5},
-        },
-        "data_index": "button_click_submit_form",
-    }
-
-    response = client.post("/log_raw_analytics", json=request_data)
-
-    assert response.status_code == 200, f"Unexpected response: {response.text}"
-    assert response.json() == "analytics-789-uuid"
-
-    mock_log_analytics.assert_called_once_with(
-        test_user_id,
-        "user_action",
-        request_data["data"],
-        "button_click_submit_form",
-    )
-
-    configured_snapshot.assert_match(
-        json.dumps({"analytics_id": response.json()}, indent=2, sort_keys=True),
-        "analytics_log_analytics_success",
-    )
-
-
-def test_log_raw_analytics_complex_data(
-    mocker: pytest_mock.MockFixture,
-    configured_snapshot: Snapshot,
-) -> None:
-    """Test raw analytics logging with complex nested data structures."""
-    mock_result = Mock(id="analytics-complex-uuid")
-    mocker.patch(
-        "backend.data.analytics.log_raw_analytics",
-        new_callable=AsyncMock,
-        return_value=mock_result,
-    )
-
-    request_data = {
-        "type": "agent_execution",
-        "data": {
-            "agent_id": "agent_123",
-            "execution_id": "exec_456",
-            "status": "completed",
-            "duration_ms": 3500,
-            "nodes_executed": 15,
-            "blocks_used": [
-                {"block_id": "llm_block", "count": 3},
-                {"block_id": "http_block", "count": 5},
-                {"block_id": "code_block", "count": 2},
-            ],
-            "errors": [],
-            "metadata": {
-                "trigger": "manual",
-                "user_tier": "premium",
-                "environment": "production",
-            },
-        },
-        "data_index": "agent_123_exec_456",
-    }
-
-    response = client.post("/log_raw_analytics", json=request_data)
-
-    assert response.status_code == 200
-
-    configured_snapshot.assert_match(
-        json.dumps(
-            {"analytics_id": response.json(), "logged_data": request_data["data"]},
-            indent=2,
-            sort_keys=True,
-        ),
-        "analytics_log_analytics_complex_data",
-    )
-
-
-@pytest.mark.parametrize(
-    "invalid_data,expected_error",
-    [
-        ({}, "Field required"),
-        ({"type": "test"}, "Field required"),
-        (
-            {"type": "test", "data": "not_a_dict", "data_index": "test"},
-            "Input should be a valid dictionary",
-        ),
-        ({"type": "test", "data": {"key": "value"}}, "Field required"),
-    ],
-    ids=[
-        "empty_request",
-        "missing_data_and_data_index",
-        "invalid_data_type",
-        "missing_data_index",
-    ],
-)
-def test_log_raw_analytics_validation_errors(
-    invalid_data: dict,
-    expected_error: str,
-) -> None:
-    """Test validation errors for invalid analytics requests."""
-    response = client.post("/log_raw_analytics", json=invalid_data)
-
-    assert response.status_code == 422
-    error_detail = response.json()
-    assert "detail" in error_detail, f"Missing 'detail' in error: {error_detail}"
-
-    error_text = json.dumps(error_detail)
-    assert (
-        expected_error in error_text
-    ), f"Expected '{expected_error}' in error response: {error_text}"
-
-
-def test_log_raw_analytics_service_error(
-    mocker: pytest_mock.MockFixture,
-    test_user_id: str,
-) -> None:
-    """Test error handling when analytics service fails."""
-    mocker.patch(
-        "backend.data.analytics.log_raw_analytics",
-        new_callable=AsyncMock,
-        side_effect=Exception("Analytics DB unreachable"),
-    )
-
-    request_data = {
-        "type": "test_event",
-        "data": {"key": "value"},
-        "data_index": "test_index",
-    }
-
-    response = client.post("/log_raw_analytics", json=request_data)
-
-    assert response.status_code == 500
-    error_detail = response.json()["detail"]
-    assert "Analytics DB unreachable" in error_detail["message"]
-    assert "hint" in error_detail

autogpt_platform/backend/backend/api/features/chat/db.py

@@ -1,249 +0,0 @@
-"""Database operations for chat sessions."""
-
-import asyncio
-import logging
-from datetime import UTC, datetime
-from typing import Any, cast
-
-from prisma.models import ChatMessage as PrismaChatMessage
-from prisma.models import ChatSession as PrismaChatSession
-from prisma.types import (
-    ChatMessageCreateInput,
-    ChatSessionCreateInput,
-    ChatSessionUpdateInput,
-    ChatSessionWhereInput,
-)
-
-from backend.data.db import transaction
-from backend.util.json import SafeJson
-
-logger = logging.getLogger(__name__)
-
-
-async def get_chat_session(session_id: str) -> PrismaChatSession | None:
-    """Get a chat session by ID from the database."""
-    session = await PrismaChatSession.prisma().find_unique(
-        where={"id": session_id},
-        include={"Messages": True},
-    )
-    if session and session.Messages:
-        # Sort messages by sequence in Python - Prisma Python client doesn't support
-        # order_by in include clauses (unlike Prisma JS), so we sort after fetching
-        session.Messages.sort(key=lambda m: m.sequence)
-    return session
-
-
-async def create_chat_session(
-    session_id: str,
-    user_id: str,
-) -> PrismaChatSession:
-    """Create a new chat session in the database."""
-    data = ChatSessionCreateInput(
-        id=session_id,
-        userId=user_id,
-        credentials=SafeJson({}),
-        successfulAgentRuns=SafeJson({}),
-        successfulAgentSchedules=SafeJson({}),
-    )
-    return await PrismaChatSession.prisma().create(
-        data=data,
-        include={"Messages": True},
-    )
-
-
-async def update_chat_session(
-    session_id: str,
-    credentials: dict[str, Any] | None = None,
-    successful_agent_runs: dict[str, Any] | None = None,
-    successful_agent_schedules: dict[str, Any] | None = None,
-    total_prompt_tokens: int | None = None,
-    total_completion_tokens: int | None = None,
-    title: str | None = None,
-) -> PrismaChatSession | None:
-    """Update a chat session's metadata."""
-    data: ChatSessionUpdateInput = {"updatedAt": datetime.now(UTC)}
-
-    if credentials is not None:
-        data["credentials"] = SafeJson(credentials)
-    if successful_agent_runs is not None:
-        data["successfulAgentRuns"] = SafeJson(successful_agent_runs)
-    if successful_agent_schedules is not None:
-        data["successfulAgentSchedules"] = SafeJson(successful_agent_schedules)
-    if total_prompt_tokens is not None:
-        data["totalPromptTokens"] = total_prompt_tokens
-    if total_completion_tokens is not None:
-        data["totalCompletionTokens"] = total_completion_tokens
-    if title is not None:
-        data["title"] = title
-
-    session = await PrismaChatSession.prisma().update(
-        where={"id": session_id},
-        data=data,
-        include={"Messages": True},
-    )
-    if session and session.Messages:
-        # Sort in Python - Prisma Python doesn't support order_by in include clauses
-        session.Messages.sort(key=lambda m: m.sequence)
-    return session
-
-
-async def add_chat_message(
-    session_id: str,
-    role: str,
-    sequence: int,
-    content: str | None = None,
-    name: str | None = None,
-    tool_call_id: str | None = None,
-    refusal: str | None = None,
-    tool_calls: list[dict[str, Any]] | None = None,
-    function_call: dict[str, Any] | None = None,
-) -> PrismaChatMessage:
-    """Add a message to a chat session."""
-    # Build input dict dynamically rather than using ChatMessageCreateInput directly
-    # because Prisma's TypedDict validation rejects optional fields set to None.
-    # We only include fields that have values, then cast at the end.
-    data: dict[str, Any] = {
-        "Session": {"connect": {"id": session_id}},
-        "role": role,
-        "sequence": sequence,
-    }
-
-    # Add optional string fields
-    if content is not None:
-        data["content"] = content
-    if name is not None:
-        data["name"] = name
-    if tool_call_id is not None:
-        data["toolCallId"] = tool_call_id
-    if refusal is not None:
-        data["refusal"] = refusal
-
-    # Add optional JSON fields only when they have values
-    if tool_calls is not None:
-        data["toolCalls"] = SafeJson(tool_calls)
-    if function_call is not None:
-        data["functionCall"] = SafeJson(function_call)
-
-    # Run message create and session timestamp update in parallel for lower latency
-    _, message = await asyncio.gather(
-        PrismaChatSession.prisma().update(
-            where={"id": session_id},
-            data={"updatedAt": datetime.now(UTC)},
-        ),
-        PrismaChatMessage.prisma().create(data=cast(ChatMessageCreateInput, data)),
-    )
-    return message
-
-
-async def add_chat_messages_batch(
-    session_id: str,
-    messages: list[dict[str, Any]],
-    start_sequence: int,
-) -> list[PrismaChatMessage]:
-    """Add multiple messages to a chat session in a batch.
-
-    Uses a transaction for atomicity - if any message creation fails,
-    the entire batch is rolled back.
-    """
-    if not messages:
-        return []
-
-    created_messages = []
-
-    async with transaction() as tx:
-        for i, msg in enumerate(messages):
-            # Build input dict dynamically rather than using ChatMessageCreateInput
-            # directly because Prisma's TypedDict validation rejects optional fields
-            # set to None. We only include fields that have values, then cast.
-            data: dict[str, Any] = {
-                "Session": {"connect": {"id": session_id}},
-                "role": msg["role"],
-                "sequence": start_sequence + i,
-            }
-
-            # Add optional string fields
-            if msg.get("content") is not None:
-                data["content"] = msg["content"]
-            if msg.get("name") is not None:
-                data["name"] = msg["name"]
-            if msg.get("tool_call_id") is not None:
-                data["toolCallId"] = msg["tool_call_id"]
-            if msg.get("refusal") is not None:
-                data["refusal"] = msg["refusal"]
-
-            # Add optional JSON fields only when they have values
-            if msg.get("tool_calls") is not None:
-                data["toolCalls"] = SafeJson(msg["tool_calls"])
-            if msg.get("function_call") is not None:
-                data["functionCall"] = SafeJson(msg["function_call"])
-
-            created = await PrismaChatMessage.prisma(tx).create(
-                data=cast(ChatMessageCreateInput, data)
-            )
-            created_messages.append(created)
-
-        # Update session's updatedAt timestamp within the same transaction.
-        # Note: Token usage (total_prompt_tokens, total_completion_tokens) is updated
-        # separately via update_chat_session() after streaming completes.
-        await PrismaChatSession.prisma(tx).update(
-            where={"id": session_id},
-            data={"updatedAt": datetime.now(UTC)},
-        )
-
-    return created_messages
-
-
-async def get_user_chat_sessions(
-    user_id: str,
-    limit: int = 50,
-    offset: int = 0,
-) -> list[PrismaChatSession]:
-    """Get chat sessions for a user, ordered by most recent."""
-    return await PrismaChatSession.prisma().find_many(
-        where={"userId": user_id},
-        order={"updatedAt": "desc"},
-        take=limit,
-        skip=offset,
-    )
-
-
-async def get_user_session_count(user_id: str) -> int:
-    """Get the total number of chat sessions for a user."""
-    return await PrismaChatSession.prisma().count(where={"userId": user_id})
-
-
-async def delete_chat_session(session_id: str, user_id: str | None = None) -> bool:
-    """Delete a chat session and all its messages.
-
-    Args:
-        session_id: The session ID to delete.
-        user_id: If provided, validates that the session belongs to this user
-            before deletion. This prevents unauthorized deletion of other
-            users' sessions.
-
-    Returns:
-        True if deleted successfully, False otherwise.
-    """
-    try:
-        # Build typed where clause with optional user_id validation
-        where_clause: ChatSessionWhereInput = {"id": session_id}
-        if user_id is not None:
-            where_clause["userId"] = user_id
-
-        result = await PrismaChatSession.prisma().delete_many(where=where_clause)
-        if result == 0:
-            logger.warning(
-                f"No session deleted for {session_id} "
-                f"(user_id validation: {user_id is not None})"
-            )
-            return False
-        return True
-    except Exception as e:
-        logger.error(f"Failed to delete chat session {session_id}: {e}")
-        return False
-
-
-async def get_chat_session_message_count(session_id: str) -> int:
-    """Get the number of messages in a chat session."""
-    count = await PrismaChatMessage.prisma().count(where={"sessionId": session_id})
-    return count

autogpt_platform/backend/backend/api/features/chat/model.py

@@ -1,597 +0,0 @@
-import asyncio
-import logging
-import uuid
-from datetime import UTC, datetime
-from typing import Any
-from weakref import WeakValueDictionary
-
-from openai.types.chat import (
-    ChatCompletionAssistantMessageParam,
-    ChatCompletionDeveloperMessageParam,
-    ChatCompletionFunctionMessageParam,
-    ChatCompletionMessageParam,
-    ChatCompletionSystemMessageParam,
-    ChatCompletionToolMessageParam,
-    ChatCompletionUserMessageParam,
-)
-from openai.types.chat.chat_completion_assistant_message_param import FunctionCall
-from openai.types.chat.chat_completion_message_tool_call_param import (
-    ChatCompletionMessageToolCallParam,
-    Function,
-)
-from prisma.models import ChatMessage as PrismaChatMessage
-from prisma.models import ChatSession as PrismaChatSession
-from pydantic import BaseModel
-
-from backend.data.redis_client import get_redis_async
-from backend.util import json
-from backend.util.exceptions import DatabaseError, RedisError
-
-from . import db as chat_db
-from .config import ChatConfig
-
-logger = logging.getLogger(__name__)
-config = ChatConfig()
-
-
-def _parse_json_field(value: str | dict | list | None, default: Any = None) -> Any:
-    """Parse a JSON field that may be stored as string or already parsed."""
-    if value is None:
-        return default
-    if isinstance(value, str):
-        return json.loads(value)
-    return value
-
-
-# Redis cache key prefix for chat sessions
-CHAT_SESSION_CACHE_PREFIX = "chat:session:"
-
-
-def _get_session_cache_key(session_id: str) -> str:
-    """Get the Redis cache key for a chat session."""
-    return f"{CHAT_SESSION_CACHE_PREFIX}{session_id}"
-
-
-# Session-level locks to prevent race conditions during concurrent upserts.
-# Uses WeakValueDictionary to automatically garbage collect locks when no longer referenced,
-# preventing unbounded memory growth while maintaining lock semantics for active sessions.
-# Invalidation: Locks are auto-removed by GC when no coroutine holds a reference (after
-# async with lock: completes). Explicit cleanup also occurs in delete_chat_session().
-_session_locks: WeakValueDictionary[str, asyncio.Lock] = WeakValueDictionary()
-_session_locks_mutex = asyncio.Lock()
-
-
-async def _get_session_lock(session_id: str) -> asyncio.Lock:
-    """Get or create a lock for a specific session to prevent concurrent upserts.
-
-    Uses WeakValueDictionary for automatic cleanup: locks are garbage collected
-    when no coroutine holds a reference to them, preventing memory leaks from
-    unbounded growth of session locks.
-    """
-    async with _session_locks_mutex:
-        lock = _session_locks.get(session_id)
-        if lock is None:
-            lock = asyncio.Lock()
-            _session_locks[session_id] = lock
-        return lock
-
-
-class ChatMessage(BaseModel):
-    role: str
-    content: str | None = None
-    name: str | None = None
-    tool_call_id: str | None = None
-    refusal: str | None = None
-    tool_calls: list[dict] | None = None
-    function_call: dict | None = None
-
-
-class Usage(BaseModel):
-    prompt_tokens: int
-    completion_tokens: int
-    total_tokens: int
-
-
-class ChatSession(BaseModel):
-    session_id: str
-    user_id: str
-    title: str | None = None
-    messages: list[ChatMessage]
-    usage: list[Usage]
-    credentials: dict[str, dict] = {}  # Map of provider -> credential metadata
-    started_at: datetime
-    updated_at: datetime
-    successful_agent_runs: dict[str, int] = {}
-    successful_agent_schedules: dict[str, int] = {}
-
-    @staticmethod
-    def new(user_id: str) -> "ChatSession":
-        return ChatSession(
-            session_id=str(uuid.uuid4()),
-            user_id=user_id,
-            title=None,
-            messages=[],
-            usage=[],
-            credentials={},
-            started_at=datetime.now(UTC),
-            updated_at=datetime.now(UTC),
-        )
-
-    @staticmethod
-    def from_db(
-        prisma_session: PrismaChatSession,
-        prisma_messages: list[PrismaChatMessage] | None = None,
-    ) -> "ChatSession":
-        """Convert Prisma models to Pydantic ChatSession."""
-        messages = []
-        if prisma_messages:
-            for msg in prisma_messages:
-                messages.append(
-                    ChatMessage(
-                        role=msg.role,
-                        content=msg.content,
-                        name=msg.name,
-                        tool_call_id=msg.toolCallId,
-                        refusal=msg.refusal,
-                        tool_calls=_parse_json_field(msg.toolCalls),
-                        function_call=_parse_json_field(msg.functionCall),
-                    )
-                )
-
-        # Parse JSON fields from Prisma
-        credentials = _parse_json_field(prisma_session.credentials, default={})
-        successful_agent_runs = _parse_json_field(
-            prisma_session.successfulAgentRuns, default={}
-        )
-        successful_agent_schedules = _parse_json_field(
-            prisma_session.successfulAgentSchedules, default={}
-        )
-
-        # Calculate usage from token counts
-        usage = []
-        if prisma_session.totalPromptTokens or prisma_session.totalCompletionTokens:
-            usage.append(
-                Usage(
-                    prompt_tokens=prisma_session.totalPromptTokens or 0,
-                    completion_tokens=prisma_session.totalCompletionTokens or 0,
-                    total_tokens=(prisma_session.totalPromptTokens or 0)
-                    + (prisma_session.totalCompletionTokens or 0),
-                )
-            )
-
-        return ChatSession(
-            session_id=prisma_session.id,
-            user_id=prisma_session.userId,
-            title=prisma_session.title,
-            messages=messages,
-            usage=usage,
-            credentials=credentials,
-            started_at=prisma_session.createdAt,
-            updated_at=prisma_session.updatedAt,
-            successful_agent_runs=successful_agent_runs,
-            successful_agent_schedules=successful_agent_schedules,
-        )
-
-    def to_openai_messages(self) -> list[ChatCompletionMessageParam]:
-        messages = []
-        for message in self.messages:
-            if message.role == "developer":
-                m = ChatCompletionDeveloperMessageParam(
-                    role="developer",
-                    content=message.content or "",
-                )
-                if message.name:
-                    m["name"] = message.name
-                messages.append(m)
-            elif message.role == "system":
-                m = ChatCompletionSystemMessageParam(
-                    role="system",
-                    content=message.content or "",
-                )
-                if message.name:
-                    m["name"] = message.name
-                messages.append(m)
-            elif message.role == "user":
-                m = ChatCompletionUserMessageParam(
-                    role="user",
-                    content=message.content or "",
-                )
-                if message.name:
-                    m["name"] = message.name
-                messages.append(m)
-            elif message.role == "assistant":
-                m = ChatCompletionAssistantMessageParam(
-                    role="assistant",
-                    content=message.content or "",
-                )
-                if message.function_call:
-                    m["function_call"] = FunctionCall(
-                        arguments=message.function_call["arguments"],
-                        name=message.function_call["name"],
-                    )
-                if message.refusal:
-                    m["refusal"] = message.refusal
-                if message.tool_calls:
-                    t: list[ChatCompletionMessageToolCallParam] = []
-                    for tool_call in message.tool_calls:
-                        # Tool calls are stored with nested structure: {id, type, function: {name, arguments}}
-                        function_data = tool_call.get("function", {})
-
-                        # Skip tool calls that are missing required fields
-                        if "id" not in tool_call or "name" not in function_data:
-                            logger.warning(
-                                f"Skipping invalid tool call: missing required fields. "
-                                f"Got: {tool_call.keys()}, function keys: {function_data.keys()}"
-                            )
-                            continue
-
-                        # Arguments are stored as a JSON string
-                        arguments_str = function_data.get("arguments", "{}")
-
-                        t.append(
-                            ChatCompletionMessageToolCallParam(
-                                id=tool_call["id"],
-                                type="function",
-                                function=Function(
-                                    arguments=arguments_str,
-                                    name=function_data["name"],
-                                ),
-                            )
-                        )
-                    m["tool_calls"] = t
-                if message.name:
-                    m["name"] = message.name
-                messages.append(m)
-            elif message.role == "tool":
-                messages.append(
-                    ChatCompletionToolMessageParam(
-                        role="tool",
-                        content=message.content or "",
-                        tool_call_id=message.tool_call_id or "",
-                    )
-                )
-            elif message.role == "function":
-                messages.append(
-                    ChatCompletionFunctionMessageParam(
-                        role="function",
-                        content=message.content,
-                        name=message.name or "",
-                    )
-                )
-        return messages
-
-
-async def _get_session_from_cache(session_id: str) -> ChatSession | None:
-    """Get a chat session from Redis cache."""
-    redis_key = _get_session_cache_key(session_id)
-    async_redis = await get_redis_async()
-    raw_session: bytes | None = await async_redis.get(redis_key)
-
-    if raw_session is None:
-        return None
-
-    try:
-        session = ChatSession.model_validate_json(raw_session)
-        logger.info(
-            f"Loading session {session_id} from cache: "
-            f"message_count={len(session.messages)}, "
-            f"roles={[m.role for m in session.messages]}"
-        )
-        return session
-    except Exception as e:
-        logger.error(f"Failed to deserialize session {session_id}: {e}", exc_info=True)
-        raise RedisError(f"Corrupted session data for {session_id}") from e
-
-
-async def _cache_session(session: ChatSession) -> None:
-    """Cache a chat session in Redis."""
-    redis_key = _get_session_cache_key(session.session_id)
-    async_redis = await get_redis_async()
-    await async_redis.setex(redis_key, config.session_ttl, session.model_dump_json())
-
-
-async def _get_session_from_db(session_id: str) -> ChatSession | None:
-    """Get a chat session from the database."""
-    prisma_session = await chat_db.get_chat_session(session_id)
-    if not prisma_session:
-        return None
-
-    messages = prisma_session.Messages
-    logger.info(
-        f"Loading session {session_id} from DB: "
-        f"has_messages={messages is not None}, "
-        f"message_count={len(messages) if messages else 0}, "
-        f"roles={[m.role for m in messages] if messages else []}"
-    )
-
-    return ChatSession.from_db(prisma_session, messages)
-
-
-async def _save_session_to_db(
-    session: ChatSession, existing_message_count: int
-) -> None:
-    """Save or update a chat session in the database."""
-    # Check if session exists in DB
-    existing = await chat_db.get_chat_session(session.session_id)
-
-    if not existing:
-        # Create new session
-        await chat_db.create_chat_session(
-            session_id=session.session_id,
-            user_id=session.user_id,
-        )
-        existing_message_count = 0
-
-    # Calculate total tokens from usage
-    total_prompt = sum(u.prompt_tokens for u in session.usage)
-    total_completion = sum(u.completion_tokens for u in session.usage)
-
-    # Update session metadata
-    await chat_db.update_chat_session(
-        session_id=session.session_id,
-        credentials=session.credentials,
-        successful_agent_runs=session.successful_agent_runs,
-        successful_agent_schedules=session.successful_agent_schedules,
-        total_prompt_tokens=total_prompt,
-        total_completion_tokens=total_completion,
-    )
-
-    # Add new messages (only those after existing count)
-    new_messages = session.messages[existing_message_count:]
-    if new_messages:
-        messages_data = []
-        for msg in new_messages:
-            messages_data.append(
-                {
-                    "role": msg.role,
-                    "content": msg.content,
-                    "name": msg.name,
-                    "tool_call_id": msg.tool_call_id,
-                    "refusal": msg.refusal,
-                    "tool_calls": msg.tool_calls,
-                    "function_call": msg.function_call,
-                }
-            )
-        logger.info(
-            f"Saving {len(new_messages)} new messages to DB for session {session.session_id}: "
-            f"roles={[m['role'] for m in messages_data]}, "
-            f"start_sequence={existing_message_count}"
-        )
-        await chat_db.add_chat_messages_batch(
-            session_id=session.session_id,
-            messages=messages_data,
-            start_sequence=existing_message_count,
-        )
-
-
-async def get_chat_session(
-    session_id: str,
-    user_id: str | None = None,
-) -> ChatSession | None:
-    """Get a chat session by ID.
-
-    Checks Redis cache first, falls back to database if not found.
-    Caches database results back to Redis.
-
-    Args:
-        session_id: The session ID to fetch.
-        user_id: If provided, validates that the session belongs to this user.
-            If None, ownership is not validated (admin/system access).
-    """
-    # Try cache first
-    try:
-        session = await _get_session_from_cache(session_id)
-        if session:
-            # Verify user ownership if user_id was provided for validation
-            if user_id is not None and session.user_id != user_id:
-                logger.warning(
-                    f"Session {session_id} user id mismatch: {session.user_id} != {user_id}"
-                )
-                return None
-            return session
-    except RedisError:
-        logger.warning(f"Cache error for session {session_id}, trying database")
-    except Exception as e:
-        logger.warning(f"Unexpected cache error for session {session_id}: {e}")
-
-    # Fall back to database
-    logger.info(f"Session {session_id} not in cache, checking database")
-    session = await _get_session_from_db(session_id)
-
-    if session is None:
-        logger.warning(f"Session {session_id} not found in cache or database")
-        return None
-
-    # Verify user ownership if user_id was provided for validation
-    if user_id is not None and session.user_id != user_id:
-        logger.warning(
-            f"Session {session_id} user id mismatch: {session.user_id} != {user_id}"
-        )
-        return None
-
-    # Cache the session from DB
-    try:
-        await _cache_session(session)
-        logger.info(f"Cached session {session_id} from database")
-    except Exception as e:
-        logger.warning(f"Failed to cache session {session_id}: {e}")
-
-    return session
-
-
-async def upsert_chat_session(
-    session: ChatSession,
-) -> ChatSession:
-    """Update a chat session in both cache and database.
-
-    Uses session-level locking to prevent race conditions when concurrent
-    operations (e.g., background title update and main stream handler)
-    attempt to upsert the same session simultaneously.
-
-    Raises:
-        DatabaseError: If the database write fails. The cache is still updated
-            as a best-effort optimization, but the error is propagated to ensure
-            callers are aware of the persistence failure.
-        RedisError: If the cache write fails (after successful DB write).
-    """
-    # Acquire session-specific lock to prevent concurrent upserts
-    lock = await _get_session_lock(session.session_id)
-
-    async with lock:
-        # Get existing message count from DB for incremental saves
-        existing_message_count = await chat_db.get_chat_session_message_count(
-            session.session_id
-        )
-
-        db_error: Exception | None = None
-
-        # Save to database (primary storage)
-        try:
-            await _save_session_to_db(session, existing_message_count)
-        except Exception as e:
-            logger.error(
-                f"Failed to save session {session.session_id} to database: {e}"
-            )
-            db_error = e
-
-        # Save to cache (best-effort, even if DB failed)
-        try:
-            await _cache_session(session)
-        except Exception as e:
-            # If DB succeeded but cache failed, raise cache error
-            if db_error is None:
-                raise RedisError(
-                    f"Failed to persist chat session {session.session_id} to Redis: {e}"
-                ) from e
-            # If both failed, log cache error but raise DB error (more critical)
-            logger.warning(
-                f"Cache write also failed for session {session.session_id}: {e}"
-            )
-
-        # Propagate DB error after attempting cache (prevents data loss)
-        if db_error is not None:
-            raise DatabaseError(
-                f"Failed to persist chat session {session.session_id} to database"
-            ) from db_error
-
-        return session
-
-
-async def create_chat_session(user_id: str) -> ChatSession:
-    """Create a new chat session and persist it.
-
-    Raises:
-        DatabaseError: If the database write fails. We fail fast to ensure
-            callers never receive a non-persisted session that only exists
-            in cache (which would be lost when the cache expires).
-    """
-    session = ChatSession.new(user_id)
-
-    # Create in database first - fail fast if this fails
-    try:
-        await chat_db.create_chat_session(
-            session_id=session.session_id,
-            user_id=user_id,
-        )
-    except Exception as e:
-        logger.error(f"Failed to create session {session.session_id} in database: {e}")
-        raise DatabaseError(
-            f"Failed to create chat session {session.session_id} in database"
-        ) from e
-
-    # Cache the session (best-effort optimization, DB is source of truth)
-    try:
-        await _cache_session(session)
-    except Exception as e:
-        logger.warning(f"Failed to cache new session {session.session_id}: {e}")
-
-    return session
-
-
-async def get_user_sessions(
-    user_id: str,
-    limit: int = 50,
-    offset: int = 0,
-) -> tuple[list[ChatSession], int]:
-    """Get chat sessions for a user from the database with total count.
-
-    Returns:
-        A tuple of (sessions, total_count) where total_count is the overall
-        number of sessions for the user (not just the current page).
-    """
-    prisma_sessions = await chat_db.get_user_chat_sessions(user_id, limit, offset)
-    total_count = await chat_db.get_user_session_count(user_id)
-
-    sessions = []
-    for prisma_session in prisma_sessions:
-        # Convert without messages for listing (lighter weight)
-        sessions.append(ChatSession.from_db(prisma_session, None))
-
-    return sessions, total_count
-
-
-async def delete_chat_session(session_id: str, user_id: str | None = None) -> bool:
-    """Delete a chat session from both cache and database.
-
-    Args:
-        session_id: The session ID to delete.
-        user_id: If provided, validates that the session belongs to this user
-            before deletion. This prevents unauthorized deletion.
-
-    Returns:
-        True if deleted successfully, False otherwise.
-    """
-    # Delete from database first (with optional user_id validation)
-    # This confirms ownership before invalidating cache
-    deleted = await chat_db.delete_chat_session(session_id, user_id)
-
-    if not deleted:
-        return False
-
-    # Only invalidate cache and clean up lock after DB confirms deletion
-    try:
-        redis_key = _get_session_cache_key(session_id)
-        async_redis = await get_redis_async()
-        await async_redis.delete(redis_key)
-    except Exception as e:
-        logger.warning(f"Failed to delete session {session_id} from cache: {e}")
-
-    # Clean up session lock (belt-and-suspenders with WeakValueDictionary)
-    async with _session_locks_mutex:
-        _session_locks.pop(session_id, None)
-
-    return True
-
-
-async def update_session_title(session_id: str, title: str) -> bool:
-    """Update only the title of a chat session.
-
-    This is a lightweight operation that doesn't touch messages, avoiding
-    race conditions with concurrent message updates. Use this for background
-    title generation instead of upsert_chat_session.
-
-    Args:
-        session_id: The session ID to update.
-        title: The new title to set.
-
-    Returns:
-        True if updated successfully, False otherwise.
-    """
-    try:
-        result = await chat_db.update_chat_session(session_id=session_id, title=title)
-        if result is None:
-            logger.warning(f"Session {session_id} not found for title update")
-            return False
-
-        # Invalidate cache so next fetch gets updated title
-        try:
-            redis_key = _get_session_cache_key(session_id)
-            async_redis = await get_redis_async()
-            await async_redis.delete(redis_key)
-        except Exception as e:
-            logger.warning(f"Failed to invalidate cache for session {session_id}: {e}")
-
-        return True
-    except Exception as e:
-        logger.error(f"Failed to update title for session {session_id}: {e}")
-        return False

autogpt_platform/backend/backend/api/features/chat/model_test.py

@@ -1,119 +0,0 @@
-import pytest
-
-from .model import (
-    ChatMessage,
-    ChatSession,
-    Usage,
-    get_chat_session,
-    upsert_chat_session,
-)
-
-messages = [
-    ChatMessage(content="Hello, how are you?", role="user"),
-    ChatMessage(
-        content="I'm fine, thank you!",
-        role="assistant",
-        tool_calls=[
-            {
-                "id": "t123",
-                "type": "function",
-                "function": {
-                    "name": "get_weather",
-                    "arguments": '{"city": "New York"}',
-                },
-            }
-        ],
-    ),
-    ChatMessage(
-        content="I'm using the tool to get the weather",
-        role="tool",
-        tool_call_id="t123",
-    ),
-]
-
-
-@pytest.mark.asyncio(loop_scope="session")
-async def test_chatsession_serialization_deserialization():
-    s = ChatSession.new(user_id="abc123")
-    s.messages = messages
-    s.usage = [Usage(prompt_tokens=100, completion_tokens=200, total_tokens=300)]
-    serialized = s.model_dump_json()
-    s2 = ChatSession.model_validate_json(serialized)
-    assert s2.model_dump() == s.model_dump()
-
-
-@pytest.mark.asyncio(loop_scope="session")
-async def test_chatsession_redis_storage(setup_test_user, test_user_id):
-
-    s = ChatSession.new(user_id=test_user_id)
-    s.messages = messages
-
-    s = await upsert_chat_session(s)
-
-    s2 = await get_chat_session(
-        session_id=s.session_id,
-        user_id=s.user_id,
-    )
-
-    assert s2 == s
-
-
-@pytest.mark.asyncio(loop_scope="session")
-async def test_chatsession_redis_storage_user_id_mismatch(
-    setup_test_user, test_user_id
-):
-
-    s = ChatSession.new(user_id=test_user_id)
-    s.messages = messages
-    s = await upsert_chat_session(s)
-
-    s2 = await get_chat_session(s.session_id, "different_user_id")
-
-    assert s2 is None
-
-
-@pytest.mark.asyncio(loop_scope="session")
-async def test_chatsession_db_storage(setup_test_user, test_user_id):
-    """Test that messages are correctly saved to and loaded from DB (not cache)."""
-    from backend.data.redis_client import get_redis_async
-
-    # Create session with messages including assistant message
-    s = ChatSession.new(user_id=test_user_id)
-    s.messages = messages  # Contains user, assistant, and tool messages
-    assert s.session_id is not None, "Session id is not set"
-    # Upsert to save to both cache and DB
-    s = await upsert_chat_session(s)
-
-    # Clear the Redis cache to force DB load
-    redis_key = f"chat:session:{s.session_id}"
-    async_redis = await get_redis_async()
-    await async_redis.delete(redis_key)
-
-    # Load from DB (cache was cleared)
-    s2 = await get_chat_session(
-        session_id=s.session_id,
-        user_id=s.user_id,
-    )
-
-    assert s2 is not None, "Session not found after loading from DB"
-    assert len(s2.messages) == len(
-        s.messages
-    ), f"Message count mismatch: expected {len(s.messages)}, got {len(s2.messages)}"
-
-    # Verify all roles are present
-    roles = [m.role for m in s2.messages]
-    assert "user" in roles, f"User message missing. Roles found: {roles}"
-    assert "assistant" in roles, f"Assistant message missing. Roles found: {roles}"
-    assert "tool" in roles, f"Tool message missing. Roles found: {roles}"
-
-    # Verify message content
-    for orig, loaded in zip(s.messages, s2.messages):
-        assert orig.role == loaded.role, f"Role mismatch: {orig.role} != {loaded.role}"
-        assert (
-            orig.content == loaded.content
-        ), f"Content mismatch for {orig.role}: {orig.content} != {loaded.content}"
-        if orig.tool_calls:
-            assert (
-                loaded.tool_calls is not None
-            ), f"Tool calls missing for {orig.role} message"
-            assert len(orig.tool_calls) == len(loaded.tool_calls)

autogpt_platform/backend/backend/api/features/chat/response_model.py

@@ -1,144 +0,0 @@
-"""
-Response models for Vercel AI SDK UI Stream Protocol.
-
-This module implements the AI SDK UI Stream Protocol (v1) for streaming chat responses.
-See: https://ai-sdk.dev/docs/ai-sdk-ui/stream-protocol
-"""
-
-from enum import Enum
-from typing import Any
-
-from pydantic import BaseModel, Field
-
-
-class ResponseType(str, Enum):
-    """Types of streaming responses following AI SDK protocol."""
-
-    # Message lifecycle
-    START = "start"
-    FINISH = "finish"
-
-    # Text streaming
-    TEXT_START = "text-start"
-    TEXT_DELTA = "text-delta"
-    TEXT_END = "text-end"
-
-    # Tool interaction
-    TOOL_INPUT_START = "tool-input-start"
-    TOOL_INPUT_AVAILABLE = "tool-input-available"
-    TOOL_OUTPUT_AVAILABLE = "tool-output-available"
-
-    # Other
-    ERROR = "error"
-    USAGE = "usage"
-
-
-class StreamBaseResponse(BaseModel):
-    """Base response model for all streaming responses."""
-
-    type: ResponseType
-
-    def to_sse(self) -> str:
-        """Convert to SSE format."""
-        return f"data: {self.model_dump_json()}\n\n"
-
-
-# ========== Message Lifecycle ==========
-
-
-class StreamStart(StreamBaseResponse):
-    """Start of a new message."""
-
-    type: ResponseType = ResponseType.START
-    messageId: str = Field(..., description="Unique message ID")
-
-
-class StreamFinish(StreamBaseResponse):
-    """End of message/stream."""
-
-    type: ResponseType = ResponseType.FINISH
-
-
-# ========== Text Streaming ==========
-
-
-class StreamTextStart(StreamBaseResponse):
-    """Start of a text block."""
-
-    type: ResponseType = ResponseType.TEXT_START
-    id: str = Field(..., description="Text block ID")
-
-
-class StreamTextDelta(StreamBaseResponse):
-    """Streaming text content delta."""
-
-    type: ResponseType = ResponseType.TEXT_DELTA
-    id: str = Field(..., description="Text block ID")
-    delta: str = Field(..., description="Text content delta")
-
-
-class StreamTextEnd(StreamBaseResponse):
-    """End of a text block."""
-
-    type: ResponseType = ResponseType.TEXT_END
-    id: str = Field(..., description="Text block ID")
-
-
-# ========== Tool Interaction ==========
-
-
-class StreamToolInputStart(StreamBaseResponse):
-    """Tool call started notification."""
-
-    type: ResponseType = ResponseType.TOOL_INPUT_START
-    toolCallId: str = Field(..., description="Unique tool call ID")
-    toolName: str = Field(..., description="Name of the tool being called")
-
-
-class StreamToolInputAvailable(StreamBaseResponse):
-    """Tool input is ready for execution."""
-
-    type: ResponseType = ResponseType.TOOL_INPUT_AVAILABLE
-    toolCallId: str = Field(..., description="Unique tool call ID")
-    toolName: str = Field(..., description="Name of the tool being called")
-    input: dict[str, Any] = Field(
-        default_factory=dict, description="Tool input arguments"
-    )
-
-
-class StreamToolOutputAvailable(StreamBaseResponse):
-    """Tool execution result."""
-
-    type: ResponseType = ResponseType.TOOL_OUTPUT_AVAILABLE
-    toolCallId: str = Field(..., description="Tool call ID this responds to")
-    output: str | dict[str, Any] = Field(..., description="Tool execution output")
-    # Additional fields for internal use (not part of AI SDK spec but useful)
-    toolName: str | None = Field(
-        default=None, description="Name of the tool that was executed"
-    )
-    success: bool = Field(
-        default=True, description="Whether the tool execution succeeded"
-    )
-
-
-# ========== Other ==========
-
-
-class StreamUsage(StreamBaseResponse):
-    """Token usage statistics."""
-
-    type: ResponseType = ResponseType.USAGE
-    promptTokens: int = Field(..., description="Number of prompt tokens")
-    completionTokens: int = Field(..., description="Number of completion tokens")
-    totalTokens: int = Field(..., description="Total number of tokens")
-
-
-class StreamError(StreamBaseResponse):
-    """Error response."""
-
-    type: ResponseType = ResponseType.ERROR
-    errorText: str = Field(..., description="Error message text")
-    code: str | None = Field(default=None, description="Error code")
-    details: dict[str, Any] | None = Field(
-        default=None, description="Additional error details"
-    )

autogpt_platform/backend/backend/api/features/chat/routes.py

@@ -1,362 +0,0 @@
-"""Chat API routes for chat session management and streaming via SSE."""
-
-import logging
-from collections.abc import AsyncGenerator
-from typing import Annotated
-
-from autogpt_libs import auth
-from fastapi import APIRouter, Depends, Query, Security
-from fastapi.responses import StreamingResponse
-from pydantic import BaseModel
-
-from backend.util.exceptions import NotFoundError
-
-from . import service as chat_service
-from .config import ChatConfig
-from .model import ChatSession, create_chat_session, get_chat_session, get_user_sessions
-
-config = ChatConfig()
-
-
-logger = logging.getLogger(__name__)
-
-
-async def _validate_and_get_session(
-    session_id: str,
-    user_id: str | None,
-) -> ChatSession:
-    """Validate session exists and belongs to user."""
-    session = await get_chat_session(session_id, user_id)
-    if not session:
-        raise NotFoundError(f"Session {session_id} not found.")
-    return session
-
-
-router = APIRouter(
-    tags=["chat"],
-)
-
-# ========== Request/Response Models ==========
-
-
-class StreamChatRequest(BaseModel):
-    """Request model for streaming chat with optional context."""
-
-    message: str
-    is_user_message: bool = True
-    context: dict[str, str] | None = None  # {url: str, content: str}
-
-
-class CreateSessionResponse(BaseModel):
-    """Response model containing information on a newly created chat session."""
-
-    id: str
-    created_at: str
-    user_id: str | None
-
-
-class SessionDetailResponse(BaseModel):
-    """Response model providing complete details for a chat session, including messages."""
-
-    id: str
-    created_at: str
-    updated_at: str
-    user_id: str | None
-    messages: list[dict]
-
-
-class SessionSummaryResponse(BaseModel):
-    """Response model for a session summary (without messages)."""
-
-    id: str
-    created_at: str
-    updated_at: str
-    title: str | None = None
-
-
-class ListSessionsResponse(BaseModel):
-    """Response model for listing chat sessions."""
-
-    sessions: list[SessionSummaryResponse]
-    total: int
-
-
-# ========== Routes ==========
-
-
-@router.get(
-    "/sessions",
-    dependencies=[Security(auth.requires_user)],
-)
-async def list_sessions(
-    user_id: Annotated[str, Security(auth.get_user_id)],
-    limit: int = Query(default=50, ge=1, le=100),
-    offset: int = Query(default=0, ge=0),
-) -> ListSessionsResponse:
-    """
-    List chat sessions for the authenticated user.
-
-    Returns a paginated list of chat sessions belonging to the current user,
-    ordered by most recently updated.
-
-    Args:
-        user_id: The authenticated user's ID.
-        limit: Maximum number of sessions to return (1-100).
-        offset: Number of sessions to skip for pagination.
-
-    Returns:
-        ListSessionsResponse: List of session summaries and total count.
-    """
-    sessions, total_count = await get_user_sessions(user_id, limit, offset)
-
-    return ListSessionsResponse(
-        sessions=[
-            SessionSummaryResponse(
-                id=session.session_id,
-                created_at=session.started_at.isoformat(),
-                updated_at=session.updated_at.isoformat(),
-                title=session.title,
-            )
-            for session in sessions
-        ],
-        total=total_count,
-    )
-
-
-@router.post(
-    "/sessions",
-)
-async def create_session(
-    user_id: Annotated[str, Depends(auth.get_user_id)],
-) -> CreateSessionResponse:
-    """
-    Create a new chat session.
-
-    Initiates a new chat session for the authenticated user.
-
-    Args:
-        user_id: The authenticated user ID parsed from the JWT (required).
-
-    Returns:
-        CreateSessionResponse: Details of the created session.
-
-    """
-    logger.info(
-        f"Creating session with user_id: "
-        f"...{user_id[-8:] if len(user_id) > 8 else '<redacted>'}"
-    )
-
-    session = await create_chat_session(user_id)
-
-    return CreateSessionResponse(
-        id=session.session_id,
-        created_at=session.started_at.isoformat(),
-        user_id=session.user_id,
-    )
-
-
-@router.get(
-    "/sessions/{session_id}",
-)
-async def get_session(
-    session_id: str,
-    user_id: Annotated[str | None, Depends(auth.get_user_id)],
-) -> SessionDetailResponse:
-    """
-    Retrieve the details of a specific chat session.
-
-    Looks up a chat session by ID for the given user (if authenticated) and returns all session data including messages.
-
-    Args:
-        session_id: The unique identifier for the desired chat session.
-        user_id: The optional authenticated user ID, or None for anonymous access.
-
-    Returns:
-        SessionDetailResponse: Details for the requested session; raises NotFoundError if not found.
-
-    """
-    session = await get_chat_session(session_id, user_id)
-    if not session:
-        raise NotFoundError(f"Session {session_id} not found")
-
-    messages = [message.model_dump() for message in session.messages]
-    logger.info(
-        f"Returning session {session_id}: "
-        f"message_count={len(messages)}, "
-        f"roles={[m.get('role') for m in messages]}"
-    )
-
-    return SessionDetailResponse(
-        id=session.session_id,
-        created_at=session.started_at.isoformat(),
-        updated_at=session.updated_at.isoformat(),
-        user_id=session.user_id or None,
-        messages=messages,
-    )
-
-
-@router.post(
-    "/sessions/{session_id}/stream",
-)
-async def stream_chat_post(
-    session_id: str,
-    request: StreamChatRequest,
-    user_id: str | None = Depends(auth.get_user_id),
-):
-    """
-    Stream chat responses for a session (POST with context support).
-
-    Streams the AI/completion responses in real time over Server-Sent Events (SSE), including:
-      - Text fragments as they are generated
-      - Tool call UI elements (if invoked)
-      - Tool execution results
-
-    Args:
-        session_id: The chat session identifier to associate with the streamed messages.
-        request: Request body containing message, is_user_message, and optional context.
-        user_id: Optional authenticated user ID.
-    Returns:
-        StreamingResponse: SSE-formatted response chunks.
-
-    """
-    session = await _validate_and_get_session(session_id, user_id)
-
-    async def event_generator() -> AsyncGenerator[str, None]:
-        async for chunk in chat_service.stream_chat_completion(
-            session_id,
-            request.message,
-            is_user_message=request.is_user_message,
-            user_id=user_id,
-            session=session,  # Pass pre-fetched session to avoid double-fetch
-            context=request.context,
-        ):
-            yield chunk.to_sse()
-        # AI SDK protocol termination
-        yield "data: [DONE]\n\n"
-
-    return StreamingResponse(
-        event_generator(),
-        media_type="text/event-stream",
-        headers={
-            "Cache-Control": "no-cache",
-            "Connection": "keep-alive",
-            "X-Accel-Buffering": "no",  # Disable nginx buffering
-            "x-vercel-ai-ui-message-stream": "v1",  # AI SDK protocol header
-        },
-    )
-
-
-@router.get(
-    "/sessions/{session_id}/stream",
-)
-async def stream_chat_get(
-    session_id: str,
-    message: Annotated[str, Query(min_length=1, max_length=10000)],
-    user_id: str | None = Depends(auth.get_user_id),
-    is_user_message: bool = Query(default=True),
-):
-    """
-    Stream chat responses for a session (GET - legacy endpoint).
-
-    Streams the AI/completion responses in real time over Server-Sent Events (SSE), including:
-      - Text fragments as they are generated
-      - Tool call UI elements (if invoked)
-      - Tool execution results
-
-    Args:
-        session_id: The chat session identifier to associate with the streamed messages.
-        message: The user's new message to process.
-        user_id: Optional authenticated user ID.
-        is_user_message: Whether the message is a user message.
-    Returns:
-        StreamingResponse: SSE-formatted response chunks.
-
-    """
-    session = await _validate_and_get_session(session_id, user_id)
-
-    async def event_generator() -> AsyncGenerator[str, None]:
-        async for chunk in chat_service.stream_chat_completion(
-            session_id,
-            message,
-            is_user_message=is_user_message,
-            user_id=user_id,
-            session=session,  # Pass pre-fetched session to avoid double-fetch
-        ):
-            yield chunk.to_sse()
-        # AI SDK protocol termination
-        yield "data: [DONE]\n\n"
-
-    return StreamingResponse(
-        event_generator(),
-        media_type="text/event-stream",
-        headers={
-            "Cache-Control": "no-cache",
-            "Connection": "keep-alive",
-            "X-Accel-Buffering": "no",  # Disable nginx buffering
-            "x-vercel-ai-ui-message-stream": "v1",  # AI SDK protocol header
-        },
-    )
-
-
-@router.patch(
-    "/sessions/{session_id}/assign-user",
-    dependencies=[Security(auth.requires_user)],
-    status_code=200,
-)
-async def session_assign_user(
-    session_id: str,
-    user_id: Annotated[str, Security(auth.get_user_id)],
-) -> dict:
-    """
-    Assign an authenticated user to a chat session.
-
-    Used (typically post-login) to claim an existing anonymous session as the current authenticated user.
-
-    Args:
-        session_id: The identifier for the (previously anonymous) session.
-        user_id: The authenticated user's ID to associate with the session.
-
-    Returns:
-        dict: Status of the assignment.
-
-    """
-    await chat_service.assign_user_to_session(session_id, user_id)
-    return {"status": "ok"}
-
-
-# ========== Health Check ==========
-
-
-@router.get("/health", status_code=200)
-async def health_check() -> dict:
-    """
-    Health check endpoint for the chat service.
-
-    Performs a full cycle test of session creation and retrieval. Should always return healthy
-    if the service and data layer are operational.
-
-    Returns:
-        dict: A status dictionary indicating health, service name, and API version.
-
-    """
-    from backend.data.user import get_or_create_user
-
-    # Ensure health check user exists (required for FK constraint)
-    health_check_user_id = "health-check-user"
-    await get_or_create_user(
-        {
-            "sub": health_check_user_id,
-            "email": "health-check@system.local",
-            "user_metadata": {"name": "Health Check User"},
-        }
-    )
-
-    # Create and retrieve session to verify full data layer
-    session = await create_chat_session(health_check_user_id)
-    await get_chat_session(session.session_id, health_check_user_id)
-
-    return {
-        "status": "healthy",
-        "service": "chat",
-        "version": "0.1.0",
-    }

autogpt_platform/backend/backend/api/features/chat/service.py

@@ -1,904 +0,0 @@
-import asyncio
-import logging
-from collections.abc import AsyncGenerator
-from typing import Any
-
-import orjson
-from langfuse import Langfuse
-from openai import (
-    APIConnectionError,
-    APIError,
-    APIStatusError,
-    AsyncOpenAI,
-    RateLimitError,
-)
-from openai.types.chat import ChatCompletionChunk, ChatCompletionToolParam
-
-from backend.data.understanding import (
-    format_understanding_for_prompt,
-    get_business_understanding,
-)
-from backend.util.exceptions import NotFoundError
-from backend.util.settings import Settings
-
-from . import db as chat_db
-from .config import ChatConfig
-from .model import (
-    ChatMessage,
-    ChatSession,
-    Usage,
-    get_chat_session,
-    update_session_title,
-    upsert_chat_session,
-)
-from .response_model import (
-    StreamBaseResponse,
-    StreamError,
-    StreamFinish,
-    StreamStart,
-    StreamTextDelta,
-    StreamTextEnd,
-    StreamTextStart,
-    StreamToolInputAvailable,
-    StreamToolInputStart,
-    StreamToolOutputAvailable,
-    StreamUsage,
-)
-from .tools import execute_tool, tools
-
-logger = logging.getLogger(__name__)
-
-config = ChatConfig()
-settings = Settings()
-client = AsyncOpenAI(api_key=config.api_key, base_url=config.base_url)
-
-# Langfuse client (lazy initialization)
-_langfuse_client: Langfuse | None = None
-
-
-class LangfuseNotConfiguredError(Exception):
-    """Raised when Langfuse is required but not configured."""
-
-    pass
-
-
-def _is_langfuse_configured() -> bool:
-    """Check if Langfuse credentials are configured."""
-    return bool(
-        settings.secrets.langfuse_public_key and settings.secrets.langfuse_secret_key
-    )
-
-
-def _get_langfuse_client() -> Langfuse:
-    """Get or create the Langfuse client for prompt management and tracing."""
-    global _langfuse_client
-    if _langfuse_client is None:
-        if not _is_langfuse_configured():
-            raise LangfuseNotConfiguredError(
-                "Langfuse is not configured. The chat feature requires Langfuse for prompt management. "
-                "Please set the LANGFUSE_PUBLIC_KEY and LANGFUSE_SECRET_KEY environment variables."
-            )
-        _langfuse_client = Langfuse(
-            public_key=settings.secrets.langfuse_public_key,
-            secret_key=settings.secrets.langfuse_secret_key,
-            host=settings.secrets.langfuse_host or "https://cloud.langfuse.com",
-        )
-    return _langfuse_client
-
-
-def _get_environment() -> str:
-    """Get the current environment name for Langfuse tagging."""
-    return settings.config.app_env.value
-
-
-def _get_langfuse_prompt() -> str:
-    """Fetch the latest production prompt from Langfuse.
-
-    Returns:
-        The compiled prompt text from Langfuse.
-
-    Raises:
-        Exception: If Langfuse is unavailable or prompt fetch fails.
-    """
-    try:
-        langfuse = _get_langfuse_client()
-        # cache_ttl_seconds=0 disables SDK caching to always get the latest prompt
-        prompt = langfuse.get_prompt(config.langfuse_prompt_name, cache_ttl_seconds=0)
-        compiled = prompt.compile()
-        logger.info(
-            f"Fetched prompt '{config.langfuse_prompt_name}' from Langfuse "
-            f"(version: {prompt.version})"
-        )
-        return compiled
-    except Exception as e:
-        logger.error(f"Failed to fetch prompt from Langfuse: {e}")
-        raise
-
-
-async def _is_first_session(user_id: str) -> bool:
-    """Check if this is the user's first chat session.
-
-    Returns True if the user has 1 or fewer sessions (meaning this is their first).
-    """
-    try:
-        session_count = await chat_db.get_user_session_count(user_id)
-        return session_count <= 1
-    except Exception as e:
-        logger.warning(f"Failed to check session count for user {user_id}: {e}")
-        return False  # Default to non-onboarding if we can't check
-
-
-async def _build_system_prompt(user_id: str | None) -> tuple[str, Any]:
-    """Build the full system prompt including business understanding if available.
-
-    Args:
-        user_id: The user ID for fetching business understanding
-                     If "default" and this is the user's first session, will use "onboarding" instead.
-
-    Returns:
-        Tuple of (compiled prompt string, Langfuse prompt object for tracing)
-    """
-
-    langfuse = _get_langfuse_client()
-
-    # cache_ttl_seconds=0 disables SDK caching to always get the latest prompt
-    prompt = langfuse.get_prompt(config.langfuse_prompt_name, cache_ttl_seconds=0)
-
-    # If user is authenticated, try to fetch their business understanding
-    understanding = None
-    if user_id:
-        try:
-            understanding = await get_business_understanding(user_id)
-        except Exception as e:
-            logger.warning(f"Failed to fetch business understanding: {e}")
-            understanding = None
-    if understanding:
-        context = format_understanding_for_prompt(understanding)
-    else:
-        context = "This is the first time you are meeting the user. Greet them and introduce them to the platform"
-
-    compiled = prompt.compile(users_information=context)
-    return compiled, prompt
-
-
-async def _generate_session_title(message: str) -> str | None:
-    """Generate a concise title for a chat session based on the first message.
-
-    Args:
-        message: The first user message in the session
-
-    Returns:
-        A short title (3-6 words) or None if generation fails
-    """
-    try:
-        response = await client.chat.completions.create(
-            model=config.title_model,
-            messages=[
-                {
-                    "role": "system",
-                    "content": (
-                        "Generate a very short title (3-6 words) for a chat conversation "
-                        "based on the user's first message. The title should capture the "
-                        "main topic or intent. Return ONLY the title, no quotes or punctuation."
-                    ),
-                },
-                {"role": "user", "content": message[:500]},  # Limit input length
-            ],
-            max_tokens=20,
-        )
-        title = response.choices[0].message.content
-        if title:
-            # Clean up the title
-            title = title.strip().strip("\"'")
-            # Limit length
-            if len(title) > 50:
-                title = title[:47] + "..."
-            return title
-        return None
-    except Exception as e:
-        logger.warning(f"Failed to generate session title: {e}")
-        return None
-
-
-async def assign_user_to_session(
-    session_id: str,
-    user_id: str,
-) -> ChatSession:
-    """
-    Assign a user to a chat session.
-    """
-    session = await get_chat_session(session_id, None)
-    if not session:
-        raise NotFoundError(f"Session {session_id} not found")
-    session.user_id = user_id
-    return await upsert_chat_session(session)
-
-
-async def stream_chat_completion(
-    session_id: str,
-    message: str | None = None,
-    is_user_message: bool = True,
-    user_id: str | None = None,
-    retry_count: int = 0,
-    session: ChatSession | None = None,
-    context: dict[str, str] | None = None,  # {url: str, content: str}
-) -> AsyncGenerator[StreamBaseResponse, None]:
-    """Main entry point for streaming chat completions with database handling.
-
-    This function handles all database operations and delegates streaming
-    to the internal _stream_chat_chunks function.
-
-    Args:
-        session_id: Chat session ID
-        user_message: User's input message
-        user_id: User ID for authentication (None for anonymous)
-        session: Optional pre-loaded session object (for recursive calls to avoid Redis refetch)
-
-    Yields:
-        StreamBaseResponse objects formatted as SSE
-
-    Raises:
-        NotFoundError: If session_id is invalid
-        ValueError: If max_context_messages is exceeded
-
-    """
-    logger.info(
-        f"Streaming chat completion for session {session_id} for message {message} and user id {user_id}. Message is user message: {is_user_message}"
-    )
-
-    # Check if Langfuse is configured - required for chat functionality
-    if not _is_langfuse_configured():
-        logger.error("Chat request failed: Langfuse is not configured")
-        yield StreamError(
-            errorText="Chat service is not available. Langfuse must be configured "
-            "with LANGFUSE_PUBLIC_KEY and LANGFUSE_SECRET_KEY environment variables."
-        )
-        yield StreamFinish()
-        return
-
-    # Langfuse observations will be created after session is loaded (need messages for input)
-    # Initialize to None so finally block can safely check and end them
-    trace = None
-    generation = None
-
-    # Only fetch from Redis if session not provided (initial call)
-    if session is None:
-        session = await get_chat_session(session_id, user_id)
-        logger.info(
-            f"Fetched session from Redis: {session.session_id if session else 'None'}, "
-            f"message_count={len(session.messages) if session else 0}"
-        )
-    else:
-        logger.info(
-            f"Using provided session object: {session.session_id}, "
-            f"message_count={len(session.messages)}"
-        )
-
-    if not session:
-        raise NotFoundError(
-            f"Session {session_id} not found. Please create a new session first."
-        )
-
-    if message:
-        # Build message content with context if provided
-        message_content = message
-        if context and context.get("url") and context.get("content"):
-            context_text = f"Page URL: {context['url']}\n\nPage Content:\n{context['content']}\n\n---\n\nUser Message: {message}"
-            message_content = context_text
-            logger.info(
-                f"Including page context: URL={context['url']}, content_length={len(context['content'])}"
-            )
-
-        session.messages.append(
-            ChatMessage(
-                role="user" if is_user_message else "assistant", content=message_content
-            )
-        )
-        logger.info(
-            f"Appended message (role={'user' if is_user_message else 'assistant'}), "
-            f"new message_count={len(session.messages)}"
-        )
-
-    logger.info(
-        f"Upserting session: {session.session_id} with user id {session.user_id}, "
-        f"message_count={len(session.messages)}"
-    )
-    session = await upsert_chat_session(session)
-    assert session, "Session not found"
-
-    # Generate title for new sessions on first user message (non-blocking)
-    # Check: is_user_message, no title yet, and this is the first user message
-    if is_user_message and message and not session.title:
-        user_messages = [m for m in session.messages if m.role == "user"]
-        if len(user_messages) == 1:
-            # First user message - generate title in background
-            import asyncio
-
-            # Capture only the values we need (not the session object) to avoid
-            # stale data issues when the main flow modifies the session
-            captured_session_id = session_id
-            captured_message = message
-
-            async def _update_title():
-                try:
-                    title = await _generate_session_title(captured_message)
-                    if title:
-                        # Use dedicated title update function that doesn't
-                        # touch messages, avoiding race conditions
-                        await update_session_title(captured_session_id, title)
-                        logger.info(
-                            f"Generated title for session {captured_session_id}: {title}"
-                        )
-                except Exception as e:
-                    logger.warning(f"Failed to update session title: {e}")
-
-            # Fire and forget - don't block the chat response
-            asyncio.create_task(_update_title())
-
-    # Build system prompt with business understanding
-    system_prompt, langfuse_prompt = await _build_system_prompt(user_id)
-
-    # Build input messages including system prompt for complete Langfuse logging
-    trace_input_messages = [{"role": "system", "content": system_prompt}] + [
-        m.model_dump() for m in session.messages
-    ]
-
-    # Create Langfuse trace for this LLM call (each call gets its own trace, grouped by session_id)
-    # Using v3 SDK: start_observation creates a root span, update_trace sets trace-level attributes
-    try:
-        langfuse = _get_langfuse_client()
-        env = _get_environment()
-        trace = langfuse.start_observation(
-            name="chat_completion",
-            input={"messages": trace_input_messages},
-            metadata={
-                "environment": env,
-                "model": config.model,
-                "message_count": len(session.messages),
-                "prompt_name": langfuse_prompt.name if langfuse_prompt else None,
-                "prompt_version": langfuse_prompt.version if langfuse_prompt else None,
-            },
-        )
-        # Set trace-level attributes (session_id, user_id, tags)
-        trace.update_trace(
-            session_id=session_id,
-            user_id=user_id,
-            tags=[env, "copilot"],
-        )
-    except Exception as e:
-        logger.warning(f"Failed to create Langfuse trace: {e}")
-
-    # Initialize variables that will be used in finally block (must be defined before try)
-    assistant_response = ChatMessage(
-        role="assistant",
-        content="",
-    )
-    accumulated_tool_calls: list[dict[str, Any]] = []
-
-    # Wrap main logic in try/finally to ensure Langfuse observations are always ended
-    try:
-        has_yielded_end = False
-        has_yielded_error = False
-        has_done_tool_call = False
-        has_received_text = False
-        text_streaming_ended = False
-        tool_response_messages: list[ChatMessage] = []
-        should_retry = False
-
-        # Generate unique IDs for AI SDK protocol
-        import uuid as uuid_module
-
-        message_id = str(uuid_module.uuid4())
-        text_block_id = str(uuid_module.uuid4())
-
-        # Yield message start
-        yield StreamStart(messageId=message_id)
-
-        # Create Langfuse generation for each LLM call, linked to the prompt
-        # Using v3 SDK: start_observation with as_type="generation"
-        generation = (
-            trace.start_observation(
-                as_type="generation",
-                name="llm_call",
-                model=config.model,
-                input={"messages": trace_input_messages},
-                prompt=langfuse_prompt,
-            )
-            if trace
-            else None
-        )
-
-        try:
-            async for chunk in _stream_chat_chunks(
-                session=session,
-                tools=tools,
-                system_prompt=system_prompt,
-                text_block_id=text_block_id,
-            ):
-
-                if isinstance(chunk, StreamTextStart):
-                    # Emit text-start before first text delta
-                    if not has_received_text:
-                        yield chunk
-                elif isinstance(chunk, StreamTextDelta):
-                    delta = chunk.delta or ""
-                    assert assistant_response.content is not None
-                    assistant_response.content += delta
-                    has_received_text = True
-                    yield chunk
-                elif isinstance(chunk, StreamTextEnd):
-                    # Emit text-end after text completes
-                    if has_received_text and not text_streaming_ended:
-                        text_streaming_ended = True
-                        yield chunk
-                elif isinstance(chunk, StreamToolInputStart):
-                    # Emit text-end before first tool call, but only if we've received text
-                    if has_received_text and not text_streaming_ended:
-                        yield StreamTextEnd(id=text_block_id)
-                        text_streaming_ended = True
-                    yield chunk
-                elif isinstance(chunk, StreamToolInputAvailable):
-                    # Accumulate tool calls in OpenAI format
-                    accumulated_tool_calls.append(
-                        {
-                            "id": chunk.toolCallId,
-                            "type": "function",
-                            "function": {
-                                "name": chunk.toolName,
-                                "arguments": orjson.dumps(chunk.input).decode("utf-8"),
-                            },
-                        }
-                    )
-                elif isinstance(chunk, StreamToolOutputAvailable):
-                    result_content = (
-                        chunk.output
-                        if isinstance(chunk.output, str)
-                        else orjson.dumps(chunk.output).decode("utf-8")
-                    )
-                    tool_response_messages.append(
-                        ChatMessage(
-                            role="tool",
-                            content=result_content,
-                            tool_call_id=chunk.toolCallId,
-                        )
-                    )
-                    has_done_tool_call = True
-                    # Track if any tool execution failed
-                    if not chunk.success:
-                        logger.warning(
-                            f"Tool {chunk.toolName} (ID: {chunk.toolCallId}) execution failed"
-                        )
-                    yield chunk
-                elif isinstance(chunk, StreamFinish):
-                    if not has_done_tool_call:
-                        # Emit text-end before finish if we received text but haven't closed it
-                        if has_received_text and not text_streaming_ended:
-                            yield StreamTextEnd(id=text_block_id)
-                            text_streaming_ended = True
-                        has_yielded_end = True
-                        yield chunk
-                elif isinstance(chunk, StreamError):
-                    has_yielded_error = True
-                elif isinstance(chunk, StreamUsage):
-                    session.usage.append(
-                        Usage(
-                            prompt_tokens=chunk.promptTokens,
-                            completion_tokens=chunk.completionTokens,
-                            total_tokens=chunk.totalTokens,
-                        )
-                    )
-                else:
-                    logger.error(f"Unknown chunk type: {type(chunk)}", exc_info=True)
-        except Exception as e:
-            logger.error(f"Error during stream: {e!s}", exc_info=True)
-
-            # Check if this is a retryable error (JSON parsing, incomplete tool calls, etc.)
-            is_retryable = isinstance(e, (orjson.JSONDecodeError, KeyError, TypeError))
-
-            if is_retryable and retry_count < config.max_retries:
-                logger.info(
-                    f"Retryable error encountered. Attempt {retry_count + 1}/{config.max_retries}"
-                )
-                should_retry = True
-            else:
-                # Non-retryable error or max retries exceeded
-                # Save any partial progress before reporting error
-                messages_to_save: list[ChatMessage] = []
-
-                # Add assistant message if it has content or tool calls
-                if accumulated_tool_calls:
-                    assistant_response.tool_calls = accumulated_tool_calls
-                if assistant_response.content or assistant_response.tool_calls:
-                    messages_to_save.append(assistant_response)
-
-                # Add tool response messages after assistant message
-                messages_to_save.extend(tool_response_messages)
-
-                session.messages.extend(messages_to_save)
-                await upsert_chat_session(session)
-
-                if not has_yielded_error:
-                    error_message = str(e)
-                    if not is_retryable:
-                        error_message = f"Non-retryable error: {error_message}"
-                    elif retry_count >= config.max_retries:
-                        error_message = f"Max retries ({config.max_retries}) exceeded: {error_message}"
-
-                    error_response = StreamError(errorText=error_message)
-                    yield error_response
-                if not has_yielded_end:
-                    yield StreamFinish()
-                return
-
-        # Handle retry outside of exception handler to avoid nesting
-        if should_retry and retry_count < config.max_retries:
-            logger.info(
-                f"Retrying stream_chat_completion for session {session_id}, attempt {retry_count + 1}"
-            )
-            async for chunk in stream_chat_completion(
-                session_id=session.session_id,
-                user_id=user_id,
-                retry_count=retry_count + 1,
-                session=session,
-                context=context,
-            ):
-                yield chunk
-            return  # Exit after retry to avoid double-saving in finally block
-
-        # Normal completion path - save session and handle tool call continuation
-        logger.info(
-            f"Normal completion path: session={session.session_id}, "
-            f"current message_count={len(session.messages)}"
-        )
-
-        # Build the messages list in the correct order
-        messages_to_save: list[ChatMessage] = []
-
-        # Add assistant message with tool_calls if any
-        if accumulated_tool_calls:
-            assistant_response.tool_calls = accumulated_tool_calls
-            logger.info(
-                f"Added {len(accumulated_tool_calls)} tool calls to assistant message"
-            )
-        if assistant_response.content or assistant_response.tool_calls:
-            messages_to_save.append(assistant_response)
-            logger.info(
-                f"Saving assistant message with content_len={len(assistant_response.content or '')}, tool_calls={len(assistant_response.tool_calls or [])}"
-            )
-
-        # Add tool response messages after assistant message
-        messages_to_save.extend(tool_response_messages)
-        logger.info(
-            f"Saving {len(tool_response_messages)} tool response messages, "
-            f"total_to_save={len(messages_to_save)}"
-        )
-
-        session.messages.extend(messages_to_save)
-        logger.info(
-            f"Extended session messages, new message_count={len(session.messages)}"
-        )
-        await upsert_chat_session(session)
-
-        # If we did a tool call, stream the chat completion again to get the next response
-        if has_done_tool_call:
-            logger.info(
-                "Tool call executed, streaming chat completion again to get assistant response"
-            )
-            async for chunk in stream_chat_completion(
-                session_id=session.session_id,
-                user_id=user_id,
-                session=session,  # Pass session object to avoid Redis refetch
-                context=context,
-            ):
-                yield chunk
-
-    finally:
-        # Always end Langfuse observations to prevent resource leaks
-        # Guard against None and catch errors to avoid masking original exceptions
-        if generation is not None:
-            try:
-                latest_usage = session.usage[-1] if session.usage else None
-                generation.update(
-                    model=config.model,
-                    output={
-                        "content": assistant_response.content,
-                        "tool_calls": accumulated_tool_calls or None,
-                    },
-                    usage_details=(
-                        {
-                            "input": latest_usage.prompt_tokens,
-                            "output": latest_usage.completion_tokens,
-                            "total": latest_usage.total_tokens,
-                        }
-                        if latest_usage
-                        else None
-                    ),
-                )
-                generation.end()
-            except Exception as e:
-                logger.warning(f"Failed to end Langfuse generation: {e}")
-
-        if trace is not None:
-            try:
-                if accumulated_tool_calls:
-                    trace.update_trace(output={"tool_calls": accumulated_tool_calls})
-                else:
-                    trace.update_trace(output={"response": assistant_response.content})
-                trace.end()
-            except Exception as e:
-                logger.warning(f"Failed to end Langfuse trace: {e}")
-
-
-# Retry configuration for OpenAI API calls
-MAX_RETRIES = 3
-BASE_DELAY_SECONDS = 1.0
-MAX_DELAY_SECONDS = 30.0
-
-
-def _is_retryable_error(error: Exception) -> bool:
-    """Determine if an error is retryable."""
-    if isinstance(error, RateLimitError):
-        return True
-    if isinstance(error, APIConnectionError):
-        return True
-    if isinstance(error, APIStatusError):
-        # APIStatusError has a response with status_code
-        # Retry on 5xx status codes (server errors)
-        if error.response.status_code >= 500:
-            return True
-    if isinstance(error, APIError):
-        # Retry on overloaded errors or 500 errors (may not have status code)
-        error_message = str(error).lower()
-        if "overloaded" in error_message or "internal server error" in error_message:
-            return True
-    return False
-
-
-async def _stream_chat_chunks(
-    session: ChatSession,
-    tools: list[ChatCompletionToolParam],
-    system_prompt: str | None = None,
-    text_block_id: str | None = None,
-) -> AsyncGenerator[StreamBaseResponse, None]:
-    """
-    Pure streaming function for OpenAI chat completions with tool calling.
-
-    This function is database-agnostic and focuses only on streaming logic.
-    Implements exponential backoff retry for transient API errors.
-
-    Args:
-        session: Chat session with conversation history
-        tools: Available tools for the model
-        system_prompt: System prompt to prepend to messages
-
-    Yields:
-        SSE formatted JSON response objects
-
-    """
-    model = config.model
-
-    logger.info("Starting pure chat stream")
-
-    # Build messages with system prompt prepended
-    messages = session.to_openai_messages()
-    if system_prompt:
-        from openai.types.chat import ChatCompletionSystemMessageParam
-
-        system_message = ChatCompletionSystemMessageParam(
-            role="system",
-            content=system_prompt,
-        )
-        messages = [system_message] + messages
-
-    # Loop to handle tool calls and continue conversation
-    while True:
-        retry_count = 0
-        last_error: Exception | None = None
-
-        while retry_count <= MAX_RETRIES:
-            try:
-                logger.info(
-                    f"Creating OpenAI chat completion stream..."
-                    f"{f' (retry {retry_count}/{MAX_RETRIES})' if retry_count > 0 else ''}"
-                )
-
-                # Create the stream with proper types
-                stream = await client.chat.completions.create(
-                    model=model,
-                    messages=messages,
-                    tools=tools,
-                    tool_choice="auto",
-                    stream=True,
-                    stream_options={"include_usage": True},
-                )
-
-                # Variables to accumulate tool calls
-                tool_calls: list[dict[str, Any]] = []
-                active_tool_call_idx: int | None = None
-                finish_reason: str | None = None
-                # Track which tool call indices have had their start event emitted
-                emitted_start_for_idx: set[int] = set()
-
-                # Track if we've started the text block
-                text_started = False
-
-                # Process the stream
-                chunk: ChatCompletionChunk
-                async for chunk in stream:
-                    if chunk.usage:
-                        yield StreamUsage(
-                            promptTokens=chunk.usage.prompt_tokens,
-                            completionTokens=chunk.usage.completion_tokens,
-                            totalTokens=chunk.usage.total_tokens,
-                        )
-
-                    if chunk.choices:
-                        choice = chunk.choices[0]
-                        delta = choice.delta
-
-                        # Capture finish reason
-                        if choice.finish_reason:
-                            finish_reason = choice.finish_reason
-                            logger.info(f"Finish reason: {finish_reason}")
-
-                        # Handle content streaming
-                        if delta.content:
-                            # Emit text-start on first text content
-                            if not text_started and text_block_id:
-                                yield StreamTextStart(id=text_block_id)
-                                text_started = True
-                            # Stream the text delta
-                            text_response = StreamTextDelta(
-                                id=text_block_id or "",
-                                delta=delta.content,
-                            )
-                            yield text_response
-
-                        # Handle tool calls
-                        if delta.tool_calls:
-                            for tc_chunk in delta.tool_calls:
-                                idx = tc_chunk.index
-
-                                # Update active tool call index if needed
-                                if (
-                                    active_tool_call_idx is None
-                                    or active_tool_call_idx != idx
-                                ):
-                                    active_tool_call_idx = idx
-
-                                # Ensure we have a tool call object at this index
-                                while len(tool_calls) <= idx:
-                                    tool_calls.append(
-                                        {
-                                            "id": "",
-                                            "type": "function",
-                                            "function": {
-                                                "name": "",
-                                                "arguments": "",
-                                            },
-                                        },
-                                    )
-
-                                # Accumulate the tool call data
-                                if tc_chunk.id:
-                                    tool_calls[idx]["id"] = tc_chunk.id
-                                if tc_chunk.function:
-                                    if tc_chunk.function.name:
-                                        tool_calls[idx]["function"][
-                                            "name"
-                                        ] = tc_chunk.function.name
-                                    if tc_chunk.function.arguments:
-                                        tool_calls[idx]["function"][
-                                            "arguments"
-                                        ] += tc_chunk.function.arguments
-
-                                # Emit StreamToolInputStart only after we have the tool call ID
-                                if (
-                                    idx not in emitted_start_for_idx
-                                    and tool_calls[idx]["id"]
-                                    and tool_calls[idx]["function"]["name"]
-                                ):
-                                    yield StreamToolInputStart(
-                                        toolCallId=tool_calls[idx]["id"],
-                                        toolName=tool_calls[idx]["function"]["name"],
-                                    )
-                                    emitted_start_for_idx.add(idx)
-                logger.info(f"Stream complete. Finish reason: {finish_reason}")
-
-                # Yield all accumulated tool calls after the stream is complete
-                # This ensures all tool call arguments have been fully received
-                for idx, tool_call in enumerate(tool_calls):
-                    try:
-                        async for tc in _yield_tool_call(tool_calls, idx, session):
-                            yield tc
-                    except (orjson.JSONDecodeError, KeyError, TypeError) as e:
-                        logger.error(
-                            f"Failed to parse tool call {idx}: {e}",
-                            exc_info=True,
-                            extra={"tool_call": tool_call},
-                        )
-                        yield StreamError(
-                            errorText=f"Invalid tool call arguments for tool {tool_call.get('function', {}).get('name', 'unknown')}: {e}",
-                        )
-                        # Re-raise to trigger retry logic in the parent function
-                        raise
-
-                yield StreamFinish()
-                return
-            except Exception as e:
-                last_error = e
-                if _is_retryable_error(e) and retry_count < MAX_RETRIES:
-                    retry_count += 1
-                    # Calculate delay with exponential backoff
-                    delay = min(
-                        BASE_DELAY_SECONDS * (2 ** (retry_count - 1)),
-                        MAX_DELAY_SECONDS,
-                    )
-                    logger.warning(
-                        f"Retryable error in stream: {e!s}. "
-                        f"Retrying in {delay:.1f}s (attempt {retry_count}/{MAX_RETRIES})"
-                    )
-                    await asyncio.sleep(delay)
-                    continue  # Retry the stream
-                else:
-                    # Non-retryable error or max retries exceeded
-                    logger.error(
-                        f"Error in stream (not retrying): {e!s}",
-                        exc_info=True,
-                    )
-                    error_response = StreamError(errorText=str(e))
-                    yield error_response
-                    yield StreamFinish()
-                    return
-
-        # If we exit the retry loop without returning, it means we exhausted retries
-        if last_error:
-            logger.error(
-                f"Max retries ({MAX_RETRIES}) exceeded. Last error: {last_error!s}",
-                exc_info=True,
-            )
-            yield StreamError(errorText=f"Max retries exceeded: {last_error!s}")
-            yield StreamFinish()
-            return
-
-
-async def _yield_tool_call(
-    tool_calls: list[dict[str, Any]],
-    yield_idx: int,
-    session: ChatSession,
-) -> AsyncGenerator[StreamBaseResponse, None]:
-    """
-    Yield a tool call and its execution result.
-
-    Raises:
-        orjson.JSONDecodeError: If tool call arguments cannot be parsed as JSON
-        KeyError: If expected tool call fields are missing
-        TypeError: If tool call structure is invalid
-    """
-    tool_name = tool_calls[yield_idx]["function"]["name"]
-    tool_call_id = tool_calls[yield_idx]["id"]
-    logger.info(f"Yielding tool call: {tool_calls[yield_idx]}")
-
-    # Parse tool call arguments - handle empty arguments gracefully
-    raw_arguments = tool_calls[yield_idx]["function"]["arguments"]
-    if raw_arguments:
-        arguments = orjson.loads(raw_arguments)
-    else:
-        arguments = {}
-
-    yield StreamToolInputAvailable(
-        toolCallId=tool_call_id,
-        toolName=tool_name,
-        input=arguments,
-    )
-
-    tool_execution_response: StreamToolOutputAvailable = await execute_tool(
-        tool_name=tool_name,
-        parameters=arguments,
-        tool_call_id=tool_call_id,
-        user_id=session.user_id,
-        session=session,
-    )
-
-    logger.info(f"Yielding Tool execution response: {tool_execution_response}")
-    yield tool_execution_response

autogpt_platform/backend/backend/api/features/chat/tools/__init__.py

@@ -1,55 +0,0 @@
-from typing import TYPE_CHECKING, Any
-
-from openai.types.chat import ChatCompletionToolParam
-
-from backend.api.features.chat.model import ChatSession
-
-from .add_understanding import AddUnderstandingTool
-from .agent_output import AgentOutputTool
-from .base import BaseTool
-from .find_agent import FindAgentTool
-from .find_block import FindBlockTool
-from .find_library_agent import FindLibraryAgentTool
-from .get_doc_page import GetDocPageTool
-from .run_agent import RunAgentTool
-from .run_block import RunBlockTool
-from .search_docs import SearchDocsTool
-
-if TYPE_CHECKING:
-    from backend.api.features.chat.response_model import StreamToolOutputAvailable
-
-# Single source of truth for all tools
-TOOL_REGISTRY: dict[str, BaseTool] = {
-    "add_understanding": AddUnderstandingTool(),
-    "find_agent": FindAgentTool(),
-    "find_block": FindBlockTool(),
-    "find_library_agent": FindLibraryAgentTool(),
-    "run_agent": RunAgentTool(),
-    "run_block": RunBlockTool(),
-    "agent_output": AgentOutputTool(),
-    "search_docs": SearchDocsTool(),
-    "get_doc_page": GetDocPageTool(),
-}
-
-# Export individual tool instances for backwards compatibility
-find_agent_tool = TOOL_REGISTRY["find_agent"]
-run_agent_tool = TOOL_REGISTRY["run_agent"]
-
-# Generated from registry for OpenAI API
-tools: list[ChatCompletionToolParam] = [
-    tool.as_openai_tool() for tool in TOOL_REGISTRY.values()
-]
-
-
-async def execute_tool(
-    tool_name: str,
-    parameters: dict[str, Any],
-    user_id: str | None,
-    session: ChatSession,
-    tool_call_id: str,
-) -> "StreamToolOutputAvailable":
-    """Execute a tool by name."""
-    tool = TOOL_REGISTRY.get(tool_name)
-    if not tool:
-        raise ValueError(f"Tool {tool_name} not found")
-    return await tool.execute(user_id, session, tool_call_id, **parameters)

autogpt_platform/backend/backend/api/features/chat/tools/add_understanding.py

@@ -1,119 +0,0 @@
-"""Tool for capturing user business understanding incrementally."""
-
-import logging
-from typing import Any
-
-from backend.api.features.chat.model import ChatSession
-from backend.data.understanding import (
-    BusinessUnderstandingInput,
-    upsert_business_understanding,
-)
-
-from .base import BaseTool
-from .models import ErrorResponse, ToolResponseBase, UnderstandingUpdatedResponse
-
-logger = logging.getLogger(__name__)
-
-
-class AddUnderstandingTool(BaseTool):
-    """Tool for capturing user's business understanding incrementally."""
-
-    @property
-    def name(self) -> str:
-        return "add_understanding"
-
-    @property
-    def description(self) -> str:
-        return """Capture and store information about the user's business context,
-workflows, pain points, and automation goals. Call this tool whenever the user
-shares information about their business. Each call incrementally adds to the
-existing understanding - you don't need to provide all fields at once.
-
-Use this to build a comprehensive profile that helps recommend better agents
-and automations for the user's specific needs."""
-
-    @property
-    def parameters(self) -> dict[str, Any]:
-        # Auto-generate from Pydantic model schema
-        schema = BusinessUnderstandingInput.model_json_schema()
-        properties = {}
-        for field_name, field_schema in schema.get("properties", {}).items():
-            prop: dict[str, Any] = {"description": field_schema.get("description", "")}
-            # Handle anyOf for Optional types
-            if "anyOf" in field_schema:
-                for option in field_schema["anyOf"]:
-                    if option.get("type") != "null":
-                        prop["type"] = option.get("type", "string")
-                        if "items" in option:
-                            prop["items"] = option["items"]
-                        break
-            else:
-                prop["type"] = field_schema.get("type", "string")
-                if "items" in field_schema:
-                    prop["items"] = field_schema["items"]
-            properties[field_name] = prop
-        return {"type": "object", "properties": properties, "required": []}
-
-    @property
-    def requires_auth(self) -> bool:
-        """Requires authentication to store user-specific data."""
-        return True
-
-    async def _execute(
-        self,
-        user_id: str | None,
-        session: ChatSession,
-        **kwargs,
-    ) -> ToolResponseBase:
-        """
-        Capture and store business understanding incrementally.
-
-        Each call merges new data with existing understanding:
-        - String fields are overwritten if provided
-        - List fields are appended (with deduplication)
-        """
-        session_id = session.session_id
-
-        if not user_id:
-            return ErrorResponse(
-                message="Authentication required to save business understanding.",
-                session_id=session_id,
-            )
-
-        # Check if any data was provided
-        if not any(v is not None for v in kwargs.values()):
-            return ErrorResponse(
-                message="Please provide at least one field to update.",
-                session_id=session_id,
-            )
-
-        # Build input model from kwargs (only include fields defined in the model)
-        valid_fields = set(BusinessUnderstandingInput.model_fields.keys())
-        input_data = BusinessUnderstandingInput(
-            **{k: v for k, v in kwargs.items() if k in valid_fields}
-        )
-
-        # Track which fields were updated
-        updated_fields = [
-            k for k, v in kwargs.items() if k in valid_fields and v is not None
-        ]
-
-        # Upsert with merge
-        understanding = await upsert_business_understanding(user_id, input_data)
-
-        # Build current understanding summary (filter out empty values)
-        current_understanding = {
-            k: v
-            for k, v in understanding.model_dump(
-                exclude={"id", "user_id", "created_at", "updated_at"}
-            ).items()
-            if v is not None and v != [] and v != ""
-        }
-
-        return UnderstandingUpdatedResponse(
-            message=f"Updated understanding with: {', '.join(updated_fields)}. "
-            "I now have a better picture of your business context.",
-            session_id=session_id,
-            updated_fields=updated_fields,
-            current_understanding=current_understanding,
-        )

autogpt_platform/backend/backend/api/features/chat/tools/agent_output.py

@@ -1,446 +0,0 @@
-"""Tool for retrieving agent execution outputs from user's library."""
-
-import logging
-import re
-from datetime import datetime, timedelta, timezone
-from typing import Any
-
-from pydantic import BaseModel, field_validator
-
-from backend.api.features.chat.model import ChatSession
-from backend.api.features.library import db as library_db
-from backend.api.features.library.model import LibraryAgent
-from backend.data import execution as execution_db
-from backend.data.execution import ExecutionStatus, GraphExecution, GraphExecutionMeta
-
-from .base import BaseTool
-from .models import (
-    AgentOutputResponse,
-    ErrorResponse,
-    ExecutionOutputInfo,
-    NoResultsResponse,
-    ToolResponseBase,
-)
-from .utils import fetch_graph_from_store_slug
-
-logger = logging.getLogger(__name__)
-
-
-class AgentOutputInput(BaseModel):
-    """Input parameters for the agent_output tool."""
-
-    agent_name: str = ""
-    library_agent_id: str = ""
-    store_slug: str = ""
-    execution_id: str = ""
-    run_time: str = "latest"
-
-    @field_validator(
-        "agent_name",
-        "library_agent_id",
-        "store_slug",
-        "execution_id",
-        "run_time",
-        mode="before",
-    )
-    @classmethod
-    def strip_strings(cls, v: Any) -> Any:
-        """Strip whitespace from string fields."""
-        return v.strip() if isinstance(v, str) else v
-
-
-def parse_time_expression(
-    time_expr: str | None,
-) -> tuple[datetime | None, datetime | None]:
-    """
-    Parse time expression into datetime range (start, end).
-
-    Supports: "latest", "yesterday", "today", "last week", "last 7 days",
-    "last month", "last 30 days", ISO date "YYYY-MM-DD", ISO datetime.
-    """
-    if not time_expr or time_expr.lower() == "latest":
-        return None, None
-
-    now = datetime.now(timezone.utc)
-    today_start = now.replace(hour=0, minute=0, second=0, microsecond=0)
-    expr = time_expr.lower().strip()
-
-    # Relative time expressions lookup
-    relative_times: dict[str, tuple[datetime, datetime]] = {
-        "yesterday": (today_start - timedelta(days=1), today_start),
-        "today": (today_start, now),
-        "last week": (now - timedelta(days=7), now),
-        "last 7 days": (now - timedelta(days=7), now),
-        "last month": (now - timedelta(days=30), now),
-        "last 30 days": (now - timedelta(days=30), now),
-    }
-    if expr in relative_times:
-        return relative_times[expr]
-
-    # Try ISO date format (YYYY-MM-DD)
-    date_match = re.match(r"^(\d{4})-(\d{2})-(\d{2})$", expr)
-    if date_match:
-        try:
-            year, month, day = map(int, date_match.groups())
-            start = datetime(year, month, day, 0, 0, 0, tzinfo=timezone.utc)
-            return start, start + timedelta(days=1)
-        except ValueError:
-            # Invalid date components (e.g., month=13, day=32)
-            pass
-
-    # Try ISO datetime
-    try:
-        parsed = datetime.fromisoformat(expr.replace("Z", "+00:00"))
-        if parsed.tzinfo is None:
-            parsed = parsed.replace(tzinfo=timezone.utc)
-        return parsed - timedelta(hours=1), parsed + timedelta(hours=1)
-    except ValueError:
-        return None, None
-
-
-class AgentOutputTool(BaseTool):
-    """Tool for retrieving execution outputs from user's library agents."""
-
-    @property
-    def name(self) -> str:
-        return "agent_output"
-
-    @property
-    def description(self) -> str:
-        return """Retrieve execution outputs from agents in the user's library.
-
-        Identify the agent using one of:
-        - agent_name: Fuzzy search in user's library
-        - library_agent_id: Exact library agent ID
-        - store_slug: Marketplace format 'username/agent-name'
-
-        Select which run to retrieve using:
-        - execution_id: Specific execution ID
-        - run_time: 'latest' (default), 'yesterday', 'last week', or ISO date 'YYYY-MM-DD'
-        """
-
-    @property
-    def parameters(self) -> dict[str, Any]:
-        return {
-            "type": "object",
-            "properties": {
-                "agent_name": {
-                    "type": "string",
-                    "description": "Agent name to search for in user's library (fuzzy match)",
-                },
-                "library_agent_id": {
-                    "type": "string",
-                    "description": "Exact library agent ID",
-                },
-                "store_slug": {
-                    "type": "string",
-                    "description": "Marketplace identifier: 'username/agent-slug'",
-                },
-                "execution_id": {
-                    "type": "string",
-                    "description": "Specific execution ID to retrieve",
-                },
-                "run_time": {
-                    "type": "string",
-                    "description": (
-                        "Time filter: 'latest', 'yesterday', 'last week', or 'YYYY-MM-DD'"
-                    ),
-                },
-            },
-            "required": [],
-        }
-
-    @property
-    def requires_auth(self) -> bool:
-        return True
-
-    async def _resolve_agent(
-        self,
-        user_id: str,
-        agent_name: str | None,
-        library_agent_id: str | None,
-        store_slug: str | None,
-    ) -> tuple[LibraryAgent | None, str | None]:
-        """
-        Resolve agent from provided identifiers.
-        Returns (library_agent, error_message).
-        """
-        # Priority 1: Exact library agent ID
-        if library_agent_id:
-            try:
-                agent = await library_db.get_library_agent(library_agent_id, user_id)
-                return agent, None
-            except Exception as e:
-                logger.warning(f"Failed to get library agent by ID: {e}")
-                return None, f"Library agent '{library_agent_id}' not found"
-
-        # Priority 2: Store slug (username/agent-name)
-        if store_slug and "/" in store_slug:
-            username, agent_slug = store_slug.split("/", 1)
-            graph, _ = await fetch_graph_from_store_slug(username, agent_slug)
-            if not graph:
-                return None, f"Agent '{store_slug}' not found in marketplace"
-
-            # Find in user's library by graph_id
-            agent = await library_db.get_library_agent_by_graph_id(user_id, graph.id)
-            if not agent:
-                return (
-                    None,
-                    f"Agent '{store_slug}' is not in your library. "
-                    "Add it first to see outputs.",
-                )
-            return agent, None
-
-        # Priority 3: Fuzzy name search in library
-        if agent_name:
-            try:
-                response = await library_db.list_library_agents(
-                    user_id=user_id,
-                    search_term=agent_name,
-                    page_size=5,
-                )
-                if not response.agents:
-                    return (
-                        None,
-                        f"No agents matching '{agent_name}' found in your library",
-                    )
-
-                # Return best match (first result from search)
-                return response.agents[0], None
-            except Exception as e:
-                logger.error(f"Error searching library agents: {e}")
-                return None, f"Error searching for agent: {e}"
-
-        return (
-            None,
-            "Please specify an agent name, library_agent_id, or store_slug",
-        )
-
-    async def _get_execution(
-        self,
-        user_id: str,
-        graph_id: str,
-        execution_id: str | None,
-        time_start: datetime | None,
-        time_end: datetime | None,
-    ) -> tuple[GraphExecution | None, list[GraphExecutionMeta], str | None]:
-        """
-        Fetch execution(s) based on filters.
-        Returns (single_execution, available_executions_meta, error_message).
-        """
-        # If specific execution_id provided, fetch it directly
-        if execution_id:
-            execution = await execution_db.get_graph_execution(
-                user_id=user_id,
-                execution_id=execution_id,
-                include_node_executions=False,
-            )
-            if not execution:
-                return None, [], f"Execution '{execution_id}' not found"
-            return execution, [], None
-
-        # Get completed executions with time filters
-        executions = await execution_db.get_graph_executions(
-            graph_id=graph_id,
-            user_id=user_id,
-            statuses=[ExecutionStatus.COMPLETED],
-            created_time_gte=time_start,
-            created_time_lte=time_end,
-            limit=10,
-        )
-
-        if not executions:
-            return None, [], None  # No error, just no executions
-
-        # If only one execution, fetch full details
-        if len(executions) == 1:
-            full_execution = await execution_db.get_graph_execution(
-                user_id=user_id,
-                execution_id=executions[0].id,
-                include_node_executions=False,
-            )
-            return full_execution, [], None
-
-        # Multiple executions - return latest with full details, plus list of available
-        full_execution = await execution_db.get_graph_execution(
-            user_id=user_id,
-            execution_id=executions[0].id,
-            include_node_executions=False,
-        )
-        return full_execution, executions, None
-
-    def _build_response(
-        self,
-        agent: LibraryAgent,
-        execution: GraphExecution | None,
-        available_executions: list[GraphExecutionMeta],
-        session_id: str | None,
-    ) -> AgentOutputResponse:
-        """Build the response based on execution data."""
-        library_agent_link = f"/library/agents/{agent.id}"
-
-        if not execution:
-            return AgentOutputResponse(
-                message=f"No completed executions found for agent '{agent.name}'",
-                session_id=session_id,
-                agent_name=agent.name,
-                agent_id=agent.graph_id,
-                library_agent_id=agent.id,
-                library_agent_link=library_agent_link,
-                total_executions=0,
-            )
-
-        execution_info = ExecutionOutputInfo(
-            execution_id=execution.id,
-            status=execution.status.value,
-            started_at=execution.started_at,
-            ended_at=execution.ended_at,
-            outputs=dict(execution.outputs),
-            inputs_summary=execution.inputs if execution.inputs else None,
-        )
-
-        available_list = None
-        if len(available_executions) > 1:
-            available_list = [
-                {
-                    "id": e.id,
-                    "status": e.status.value,
-                    "started_at": e.started_at.isoformat() if e.started_at else None,
-                }
-                for e in available_executions[:5]
-            ]
-
-        message = f"Found execution outputs for agent '{agent.name}'"
-        if len(available_executions) > 1:
-            message += (
-                f". Showing latest of {len(available_executions)} matching executions."
-            )
-
-        return AgentOutputResponse(
-            message=message,
-            session_id=session_id,
-            agent_name=agent.name,
-            agent_id=agent.graph_id,
-            library_agent_id=agent.id,
-            library_agent_link=library_agent_link,
-            execution=execution_info,
-            available_executions=available_list,
-            total_executions=len(available_executions) if available_executions else 1,
-        )
-
-    async def _execute(
-        self,
-        user_id: str | None,
-        session: ChatSession,
-        **kwargs,
-    ) -> ToolResponseBase:
-        """Execute the agent_output tool."""
-        session_id = session.session_id
-
-        # Parse and validate input
-        try:
-            input_data = AgentOutputInput(**kwargs)
-        except Exception as e:
-            logger.error(f"Invalid input: {e}")
-            return ErrorResponse(
-                message="Invalid input parameters",
-                error=str(e),
-                session_id=session_id,
-            )
-
-        # Ensure user_id is present (should be guaranteed by requires_auth)
-        if not user_id:
-            return ErrorResponse(
-                message="User authentication required",
-                session_id=session_id,
-            )
-
-        # Check if at least one identifier is provided
-        if not any(
-            [
-                input_data.agent_name,
-                input_data.library_agent_id,
-                input_data.store_slug,
-                input_data.execution_id,
-            ]
-        ):
-            return ErrorResponse(
-                message=(
-                    "Please specify at least one of: agent_name, "
-                    "library_agent_id, store_slug, or execution_id"
-                ),
-                session_id=session_id,
-            )
-
-        # If only execution_id provided, we need to find the agent differently
-        if (
-            input_data.execution_id
-            and not input_data.agent_name
-            and not input_data.library_agent_id
-            and not input_data.store_slug
-        ):
-            # Fetch execution directly to get graph_id
-            execution = await execution_db.get_graph_execution(
-                user_id=user_id,
-                execution_id=input_data.execution_id,
-                include_node_executions=False,
-            )
-            if not execution:
-                return ErrorResponse(
-                    message=f"Execution '{input_data.execution_id}' not found",
-                    session_id=session_id,
-                )
-
-            # Find library agent by graph_id
-            agent = await library_db.get_library_agent_by_graph_id(
-                user_id, execution.graph_id
-            )
-            if not agent:
-                return NoResultsResponse(
-                    message=(
-                        f"Execution found but agent not in your library. "
-                        f"Graph ID: {execution.graph_id}"
-                    ),
-                    session_id=session_id,
-                    suggestions=["Add the agent to your library to see more details"],
-                )
-
-            return self._build_response(agent, execution, [], session_id)
-
-        # Resolve agent from identifiers
-        agent, error = await self._resolve_agent(
-            user_id=user_id,
-            agent_name=input_data.agent_name or None,
-            library_agent_id=input_data.library_agent_id or None,
-            store_slug=input_data.store_slug or None,
-        )
-
-        if error or not agent:
-            return NoResultsResponse(
-                message=error or "Agent not found",
-                session_id=session_id,
-                suggestions=[
-                    "Check the agent name or ID",
-                    "Make sure the agent is in your library",
-                ],
-            )
-
-        # Parse time expression
-        time_start, time_end = parse_time_expression(input_data.run_time)
-
-        # Fetch execution(s)
-        execution, available_executions, exec_error = await self._get_execution(
-            user_id=user_id,
-            graph_id=agent.graph_id,
-            execution_id=input_data.execution_id or None,
-            time_start=time_start,
-            time_end=time_end,
-        )
-
-        if exec_error:
-            return ErrorResponse(
-                message=exec_error,
-                session_id=session_id,
-            )
-
-        return self._build_response(agent, execution, available_executions, session_id)

autogpt_platform/backend/backend/api/features/chat/tools/agent_search.py

@@ -1,151 +0,0 @@
-"""Shared agent search functionality for find_agent and find_library_agent tools."""
-
-import logging
-from typing import Literal
-
-from backend.api.features.library import db as library_db
-from backend.api.features.store import db as store_db
-from backend.util.exceptions import DatabaseError, NotFoundError
-
-from .models import (
-    AgentInfo,
-    AgentsFoundResponse,
-    ErrorResponse,
-    NoResultsResponse,
-    ToolResponseBase,
-)
-
-logger = logging.getLogger(__name__)
-
-SearchSource = Literal["marketplace", "library"]
-
-
-async def search_agents(
-    query: str,
-    source: SearchSource,
-    session_id: str | None,
-    user_id: str | None = None,
-) -> ToolResponseBase:
-    """
-    Search for agents in marketplace or user library.
-
-    Args:
-        query: Search query string
-        source: "marketplace" or "library"
-        session_id: Chat session ID
-        user_id: User ID (required for library search)
-
-    Returns:
-        AgentsFoundResponse, NoResultsResponse, or ErrorResponse
-    """
-    if not query:
-        return ErrorResponse(
-            message="Please provide a search query", session_id=session_id
-        )
-
-    if source == "library" and not user_id:
-        return ErrorResponse(
-            message="User authentication required to search library",
-            session_id=session_id,
-        )
-
-    agents: list[AgentInfo] = []
-    try:
-        if source == "marketplace":
-            logger.info(f"Searching marketplace for: {query}")
-            results = await store_db.get_store_agents(search_query=query, page_size=5)
-            for agent in results.agents:
-                agents.append(
-                    AgentInfo(
-                        id=f"{agent.creator}/{agent.slug}",
-                        name=agent.agent_name,
-                        description=agent.description or "",
-                        source="marketplace",
-                        in_library=False,
-                        creator=agent.creator,
-                        category="general",
-                        rating=agent.rating,
-                        runs=agent.runs,
-                        is_featured=False,
-                    )
-                )
-        else:  # library
-            logger.info(f"Searching user library for: {query}")
-            results = await library_db.list_library_agents(
-                user_id=user_id,  # type: ignore[arg-type]
-                search_term=query,
-                page_size=10,
-            )
-            for agent in results.agents:
-                agents.append(
-                    AgentInfo(
-                        id=agent.id,
-                        name=agent.name,
-                        description=agent.description or "",
-                        source="library",
-                        in_library=True,
-                        creator=agent.creator_name,
-                        status=agent.status.value,
-                        can_access_graph=agent.can_access_graph,
-                        has_external_trigger=agent.has_external_trigger,
-                        new_output=agent.new_output,
-                        graph_id=agent.graph_id,
-                    )
-                )
-        logger.info(f"Found {len(agents)} agents in {source}")
-    except NotFoundError:
-        pass
-    except DatabaseError as e:
-        logger.error(f"Error searching {source}: {e}", exc_info=True)
-        return ErrorResponse(
-            message=f"Failed to search {source}. Please try again.",
-            error=str(e),
-            session_id=session_id,
-        )
-
-    if not agents:
-        suggestions = (
-            [
-                "Try more general terms",
-                "Browse categories in the marketplace",
-                "Check spelling",
-            ]
-            if source == "marketplace"
-            else [
-                "Try different keywords",
-                "Use find_agent to search the marketplace",
-                "Check your library at /library",
-            ]
-        )
-        no_results_msg = (
-            f"No agents found matching '{query}'. Try different keywords or browse the marketplace."
-            if source == "marketplace"
-            else f"No agents matching '{query}' found in your library."
-        )
-        return NoResultsResponse(
-            message=no_results_msg, session_id=session_id, suggestions=suggestions
-        )
-
-    title = f"Found {len(agents)} agent{'s' if len(agents) != 1 else ''} "
-    title += (
-        f"for '{query}'"
-        if source == "marketplace"
-        else f"in your library for '{query}'"
-    )
-
-    message = (
-        "Now you have found some options for the user to choose from. "
-        "You can add a link to a recommended agent at: /marketplace/agent/agent_id "
-        "Please ask the user if they would like to use any of these agents."
-        if source == "marketplace"
-        else "Found agents in the user's library. You can provide a link to view an agent at: "
-        "/library/agents/{agent_id}. Use agent_output to get execution results, or run_agent to execute."
-    )
-
-    return AgentsFoundResponse(
-        message=message,
-        title=title,
-        agents=agents,
-        count=len(agents),
-        session_id=session_id,
-    )

autogpt_platform/backend/backend/api/features/chat/tools/find_agent.py

@@ -1,46 +0,0 @@
-"""Tool for discovering agents from marketplace."""
-
-from typing import Any
-
-from backend.api.features.chat.model import ChatSession
-
-from .agent_search import search_agents
-from .base import BaseTool
-from .models import ToolResponseBase
-
-
-class FindAgentTool(BaseTool):
-    """Tool for discovering agents from the marketplace."""
-
-    @property
-    def name(self) -> str:
-        return "find_agent"
-
-    @property
-    def description(self) -> str:
-        return (
-            "Discover agents from the marketplace based on capabilities and user needs."
-        )
-
-    @property
-    def parameters(self) -> dict[str, Any]:
-        return {
-            "type": "object",
-            "properties": {
-                "query": {
-                    "type": "string",
-                    "description": "Search query describing what the user wants to accomplish. Use single keywords for best results.",
-                },
-            },
-            "required": ["query"],
-        }
-
-    async def _execute(
-        self, user_id: str | None, session: ChatSession, **kwargs
-    ) -> ToolResponseBase:
-        return await search_agents(
-            query=kwargs.get("query", "").strip(),
-            source="marketplace",
-            session_id=session.session_id,
-            user_id=user_id,
-        )

autogpt_platform/backend/backend/api/features/chat/tools/find_block.py

@@ -1,192 +0,0 @@
-import logging
-from typing import Any
-
-from prisma.enums import ContentType
-
-from backend.api.features.chat.model import ChatSession
-from backend.api.features.chat.tools.base import BaseTool, ToolResponseBase
-from backend.api.features.chat.tools.models import (
-    BlockInfoSummary,
-    BlockInputFieldInfo,
-    BlockListResponse,
-    ErrorResponse,
-    NoResultsResponse,
-)
-from backend.api.features.store.hybrid_search import unified_hybrid_search
-from backend.data.block import get_block
-
-logger = logging.getLogger(__name__)
-
-
-class FindBlockTool(BaseTool):
-    """Tool for searching available blocks."""
-
-    @property
-    def name(self) -> str:
-        return "find_block"
-
-    @property
-    def description(self) -> str:
-        return (
-            "Search for available blocks by name or description. "
-            "Blocks are reusable components that perform specific tasks like "
-            "sending emails, making API calls, processing text, etc. "
-            "IMPORTANT: Use this tool FIRST to get the block's 'id' before calling run_block. "
-            "The response includes each block's id, required_inputs, and input_schema."
-        )
-
-    @property
-    def parameters(self) -> dict[str, Any]:
-        return {
-            "type": "object",
-            "properties": {
-                "query": {
-                    "type": "string",
-                    "description": (
-                        "Search query to find blocks by name or description. "
-                        "Use keywords like 'email', 'http', 'text', 'ai', etc."
-                    ),
-                },
-            },
-            "required": ["query"],
-        }
-
-    @property
-    def requires_auth(self) -> bool:
-        return True
-
-    async def _execute(
-        self,
-        user_id: str | None,
-        session: ChatSession,
-        **kwargs,
-    ) -> ToolResponseBase:
-        """Search for blocks matching the query.
-
-        Args:
-            user_id: User ID (required)
-            session: Chat session
-            query: Search query
-
-        Returns:
-            BlockListResponse: List of matching blocks
-            NoResultsResponse: No blocks found
-            ErrorResponse: Error message
-        """
-        query = kwargs.get("query", "").strip()
-        session_id = session.session_id
-
-        if not query:
-            return ErrorResponse(
-                message="Please provide a search query",
-                session_id=session_id,
-            )
-
-        try:
-            # Search for blocks using hybrid search
-            results, total = await unified_hybrid_search(
-                query=query,
-                content_types=[ContentType.BLOCK],
-                page=1,
-                page_size=10,
-            )
-
-            if not results:
-                return NoResultsResponse(
-                    message=f"No blocks found for '{query}'",
-                    suggestions=[
-                        "Try broader keywords like 'email', 'http', 'text', 'ai'",
-                        "Check spelling of technical terms",
-                    ],
-                    session_id=session_id,
-                )
-
-            # Enrich results with full block information
-            blocks: list[BlockInfoSummary] = []
-            for result in results:
-                block_id = result["content_id"]
-                block = get_block(block_id)
-
-                if block:
-                    # Get input/output schemas
-                    input_schema = {}
-                    output_schema = {}
-                    try:
-                        input_schema = block.input_schema.jsonschema()
-                    except Exception:
-                        pass
-                    try:
-                        output_schema = block.output_schema.jsonschema()
-                    except Exception:
-                        pass
-
-                    # Get categories from block instance
-                    categories = []
-                    if hasattr(block, "categories") and block.categories:
-                        categories = [cat.value for cat in block.categories]
-
-                    # Extract required inputs for easier use
-                    required_inputs: list[BlockInputFieldInfo] = []
-                    if input_schema:
-                        properties = input_schema.get("properties", {})
-                        required_fields = set(input_schema.get("required", []))
-                        # Get credential field names to exclude from required inputs
-                        credentials_fields = set(
-                            block.input_schema.get_credentials_fields().keys()
-                        )
-
-                        for field_name, field_schema in properties.items():
-                            # Skip credential fields - they're handled separately
-                            if field_name in credentials_fields:
-                                continue
-
-                            required_inputs.append(
-                                BlockInputFieldInfo(
-                                    name=field_name,
-                                    type=field_schema.get("type", "string"),
-                                    description=field_schema.get("description", ""),
-                                    required=field_name in required_fields,
-                                    default=field_schema.get("default"),
-                                )
-                            )
-
-                    blocks.append(
-                        BlockInfoSummary(
-                            id=block_id,
-                            name=block.name,
-                            description=block.description or "",
-                            categories=categories,
-                            input_schema=input_schema,
-                            output_schema=output_schema,
-                            required_inputs=required_inputs,
-                        )
-                    )
-
-            if not blocks:
-                return NoResultsResponse(
-                    message=f"No blocks found for '{query}'",
-                    suggestions=[
-                        "Try broader keywords like 'email', 'http', 'text', 'ai'",
-                    ],
-                    session_id=session_id,
-                )
-
-            return BlockListResponse(
-                message=(
-                    f"Found {len(blocks)} block(s) matching '{query}'. "
-                    "To execute a block, use run_block with the block's 'id' field "
-                    "and provide 'input_data' matching the block's input_schema."
-                ),
-                blocks=blocks,
-                count=len(blocks),
-                query=query,
-                session_id=session_id,
-            )
-
-        except Exception as e:
-            logger.error(f"Error searching blocks: {e}", exc_info=True)
-            return ErrorResponse(
-                message="Failed to search blocks",
-                error=str(e),
-                session_id=session_id,
-            )

autogpt_platform/backend/backend/api/features/chat/tools/find_library_agent.py

@@ -1,52 +0,0 @@
-"""Tool for searching agents in the user's library."""
-
-from typing import Any
-
-from backend.api.features.chat.model import ChatSession
-
-from .agent_search import search_agents
-from .base import BaseTool
-from .models import ToolResponseBase
-
-
-class FindLibraryAgentTool(BaseTool):
-    """Tool for searching agents in the user's library."""
-
-    @property
-    def name(self) -> str:
-        return "find_library_agent"
-
-    @property
-    def description(self) -> str:
-        return (
-            "Search for agents in the user's library. Use this to find agents "
-            "the user has already added to their library, including agents they "
-            "created or added from the marketplace."
-        )
-
-    @property
-    def parameters(self) -> dict[str, Any]:
-        return {
-            "type": "object",
-            "properties": {
-                "query": {
-                    "type": "string",
-                    "description": "Search query to find agents by name or description.",
-                },
-            },
-            "required": ["query"],
-        }
-
-    @property
-    def requires_auth(self) -> bool:
-        return True
-
-    async def _execute(
-        self, user_id: str | None, session: ChatSession, **kwargs
-    ) -> ToolResponseBase:
-        return await search_agents(
-            query=kwargs.get("query", "").strip(),
-            source="library",
-            session_id=session.session_id,
-            user_id=user_id,
-        )

autogpt_platform/backend/backend/api/features/chat/tools/get_doc_page.py

@@ -1,148 +0,0 @@
-"""GetDocPageTool - Fetch full content of a documentation page."""
-
-import logging
-from pathlib import Path
-from typing import Any
-
-from backend.api.features.chat.model import ChatSession
-from backend.api.features.chat.tools.base import BaseTool
-from backend.api.features.chat.tools.models import (
-    DocPageResponse,
-    ErrorResponse,
-    ToolResponseBase,
-)
-
-logger = logging.getLogger(__name__)
-
-# Base URL for documentation (can be configured)
-DOCS_BASE_URL = "https://docs.agpt.co"
-
-
-class GetDocPageTool(BaseTool):
-    """Tool for fetching full content of a documentation page."""
-
-    @property
-    def name(self) -> str:
-        return "get_doc_page"
-
-    @property
-    def description(self) -> str:
-        return (
-            "Get the full content of a documentation page by its path. "
-            "Use this after search_docs to read the complete content of a relevant page."
-        )
-
-    @property
-    def parameters(self) -> dict[str, Any]:
-        return {
-            "type": "object",
-            "properties": {
-                "path": {
-                    "type": "string",
-                    "description": (
-                        "The path to the documentation file, as returned by search_docs. "
-                        "Example: 'platform/block-sdk-guide.md'"
-                    ),
-                },
-            },
-            "required": ["path"],
-        }
-
-    @property
-    def requires_auth(self) -> bool:
-        return False  # Documentation is public
-
-    def _get_docs_root(self) -> Path:
-        """Get the documentation root directory."""
-        this_file = Path(__file__)
-        project_root = this_file.parent.parent.parent.parent.parent.parent.parent.parent
-        return project_root / "docs"
-
-    def _extract_title(self, content: str, fallback: str) -> str:
-        """Extract title from markdown content."""
-        lines = content.split("\n")
-        for line in lines:
-            if line.startswith("# "):
-                return line[2:].strip()
-        return fallback
-
-    def _make_doc_url(self, path: str) -> str:
-        """Create a URL for a documentation page."""
-        url_path = path.rsplit(".", 1)[0] if "." in path else path
-        return f"{DOCS_BASE_URL}/{url_path}"
-
-    async def _execute(
-        self,
-        user_id: str | None,
-        session: ChatSession,
-        **kwargs,
-    ) -> ToolResponseBase:
-        """Fetch full content of a documentation page.
-
-        Args:
-            user_id: User ID (not required for docs)
-            session: Chat session
-            path: Path to the documentation file
-
-        Returns:
-            DocPageResponse: Full document content
-            ErrorResponse: Error message
-        """
-        path = kwargs.get("path", "").strip()
-        session_id = session.session_id if session else None
-
-        if not path:
-            return ErrorResponse(
-                message="Please provide a documentation path.",
-                error="Missing path parameter",
-                session_id=session_id,
-            )
-
-        # Sanitize path to prevent directory traversal
-        if ".." in path or path.startswith("/"):
-            return ErrorResponse(
-                message="Invalid documentation path.",
-                error="invalid_path",
-                session_id=session_id,
-            )
-
-        docs_root = self._get_docs_root()
-        full_path = docs_root / path
-
-        if not full_path.exists():
-            return ErrorResponse(
-                message=f"Documentation page not found: {path}",
-                error="not_found",
-                session_id=session_id,
-            )
-
-        # Ensure the path is within docs root
-        try:
-            full_path.resolve().relative_to(docs_root.resolve())
-        except ValueError:
-            return ErrorResponse(
-                message="Invalid documentation path.",
-                error="invalid_path",
-                session_id=session_id,
-            )
-
-        try:
-            content = full_path.read_text(encoding="utf-8")
-            title = self._extract_title(content, path)
-
-            return DocPageResponse(
-                message=f"Retrieved documentation page: {title}",
-                title=title,
-                path=path,
-                content=content,
-                doc_url=self._make_doc_url(path),
-                session_id=session_id,
-            )
-
-        except Exception as e:
-            logger.error(f"Failed to read documentation page {path}: {e}")
-            return ErrorResponse(
-                message=f"Failed to read documentation page: {str(e)}",
-                error="read_failed",
-                session_id=session_id,
-            )

autogpt_platform/backend/backend/api/features/chat/tools/run_block.py

@@ -1,297 +0,0 @@
-"""Tool for executing blocks directly."""
-
-import logging
-from collections import defaultdict
-from typing import Any
-
-from backend.api.features.chat.model import ChatSession
-from backend.data.block import get_block
-from backend.data.execution import ExecutionContext
-from backend.data.model import CredentialsMetaInput
-from backend.integrations.creds_manager import IntegrationCredentialsManager
-from backend.util.exceptions import BlockError
-
-from .base import BaseTool
-from .models import (
-    BlockOutputResponse,
-    ErrorResponse,
-    SetupInfo,
-    SetupRequirementsResponse,
-    ToolResponseBase,
-    UserReadiness,
-)
-
-logger = logging.getLogger(__name__)
-
-
-class RunBlockTool(BaseTool):
-    """Tool for executing a block and returning its outputs."""
-
-    @property
-    def name(self) -> str:
-        return "run_block"
-
-    @property
-    def description(self) -> str:
-        return (
-            "Execute a specific block with the provided input data. "
-            "IMPORTANT: You MUST call find_block first to get the block's 'id' - "
-            "do NOT guess or make up block IDs. "
-            "Use the 'id' from find_block results and provide input_data "
-            "matching the block's required_inputs."
-        )
-
-    @property
-    def parameters(self) -> dict[str, Any]:
-        return {
-            "type": "object",
-            "properties": {
-                "block_id": {
-                    "type": "string",
-                    "description": (
-                        "The block's 'id' field from find_block results. "
-                        "NEVER guess this - always get it from find_block first."
-                    ),
-                },
-                "input_data": {
-                    "type": "object",
-                    "description": (
-                        "Input values for the block. Use the 'required_inputs' field "
-                        "from find_block to see what fields are needed."
-                    ),
-                },
-            },
-            "required": ["block_id", "input_data"],
-        }
-
-    @property
-    def requires_auth(self) -> bool:
-        return True
-
-    async def _check_block_credentials(
-        self,
-        user_id: str,
-        block: Any,
-    ) -> tuple[dict[str, CredentialsMetaInput], list[CredentialsMetaInput]]:
-        """
-        Check if user has required credentials for a block.
-
-        Returns:
-            tuple[matched_credentials, missing_credentials]
-        """
-        matched_credentials: dict[str, CredentialsMetaInput] = {}
-        missing_credentials: list[CredentialsMetaInput] = []
-
-        # Get credential field info from block's input schema
-        credentials_fields_info = block.input_schema.get_credentials_fields_info()
-
-        if not credentials_fields_info:
-            return matched_credentials, missing_credentials
-
-        # Get user's available credentials
-        creds_manager = IntegrationCredentialsManager()
-        available_creds = await creds_manager.store.get_all_creds(user_id)
-
-        for field_name, field_info in credentials_fields_info.items():
-            # field_info.provider is a frozenset of acceptable providers
-            # field_info.supported_types is a frozenset of acceptable types
-            matching_cred = next(
-                (
-                    cred
-                    for cred in available_creds
-                    if cred.provider in field_info.provider
-                    and cred.type in field_info.supported_types
-                ),
-                None,
-            )
-
-            if matching_cred:
-                matched_credentials[field_name] = CredentialsMetaInput(
-                    id=matching_cred.id,
-                    provider=matching_cred.provider,  # type: ignore
-                    type=matching_cred.type,
-                    title=matching_cred.title,
-                )
-            else:
-                # Create a placeholder for the missing credential
-                provider = next(iter(field_info.provider), "unknown")
-                cred_type = next(iter(field_info.supported_types), "api_key")
-                missing_credentials.append(
-                    CredentialsMetaInput(
-                        id=field_name,
-                        provider=provider,  # type: ignore
-                        type=cred_type,  # type: ignore
-                        title=field_name.replace("_", " ").title(),
-                    )
-                )
-
-        return matched_credentials, missing_credentials
-
-    async def _execute(
-        self,
-        user_id: str | None,
-        session: ChatSession,
-        **kwargs,
-    ) -> ToolResponseBase:
-        """Execute a block with the given input data.
-
-        Args:
-            user_id: User ID (required)
-            session: Chat session
-            block_id: Block UUID to execute
-            input_data: Input values for the block
-
-        Returns:
-            BlockOutputResponse: Block execution outputs
-            SetupRequirementsResponse: Missing credentials
-            ErrorResponse: Error message
-        """
-        block_id = kwargs.get("block_id", "").strip()
-        input_data = kwargs.get("input_data", {})
-        session_id = session.session_id
-
-        if not block_id:
-            return ErrorResponse(
-                message="Please provide a block_id",
-                session_id=session_id,
-            )
-
-        if not isinstance(input_data, dict):
-            return ErrorResponse(
-                message="input_data must be an object",
-                session_id=session_id,
-            )
-
-        if not user_id:
-            return ErrorResponse(
-                message="Authentication required",
-                session_id=session_id,
-            )
-
-        # Get the block
-        block = get_block(block_id)
-        if not block:
-            return ErrorResponse(
-                message=f"Block '{block_id}' not found",
-                session_id=session_id,
-            )
-
-        logger.info(f"Executing block {block.name} ({block_id}) for user {user_id}")
-
-        # Check credentials
-        creds_manager = IntegrationCredentialsManager()
-        matched_credentials, missing_credentials = await self._check_block_credentials(
-            user_id, block
-        )
-
-        if missing_credentials:
-            # Return setup requirements response with missing credentials
-            missing_creds_dict = {c.id: c.model_dump() for c in missing_credentials}
-
-            return SetupRequirementsResponse(
-                message=(
-                    f"Block '{block.name}' requires credentials that are not configured. "
-                    "Please set up the required credentials before running this block."
-                ),
-                session_id=session_id,
-                setup_info=SetupInfo(
-                    agent_id=block_id,
-                    agent_name=block.name,
-                    user_readiness=UserReadiness(
-                        has_all_credentials=False,
-                        missing_credentials=missing_creds_dict,
-                        ready_to_run=False,
-                    ),
-                    requirements={
-                        "credentials": [c.model_dump() for c in missing_credentials],
-                        "inputs": self._get_inputs_list(block),
-                        "execution_modes": ["immediate"],
-                    },
-                ),
-                graph_id=None,
-                graph_version=None,
-            )
-
-        try:
-            # Fetch actual credentials and prepare kwargs for block execution
-            # Create execution context with defaults (blocks may require it)
-            exec_kwargs: dict[str, Any] = {
-                "user_id": user_id,
-                "execution_context": ExecutionContext(),
-            }
-
-            for field_name, cred_meta in matched_credentials.items():
-                # Inject metadata into input_data (for validation)
-                if field_name not in input_data:
-                    input_data[field_name] = cred_meta.model_dump()
-
-                # Fetch actual credentials and pass as kwargs (for execution)
-                actual_credentials = await creds_manager.get(
-                    user_id, cred_meta.id, lock=False
-                )
-                if actual_credentials:
-                    exec_kwargs[field_name] = actual_credentials
-                else:
-                    return ErrorResponse(
-                        message=f"Failed to retrieve credentials for {field_name}",
-                        session_id=session_id,
-                    )
-
-            # Execute the block and collect outputs
-            outputs: dict[str, list[Any]] = defaultdict(list)
-            async for output_name, output_data in block.execute(
-                input_data,
-                **exec_kwargs,
-            ):
-                outputs[output_name].append(output_data)
-
-            return BlockOutputResponse(
-                message=f"Block '{block.name}' executed successfully",
-                block_id=block_id,
-                block_name=block.name,
-                outputs=dict(outputs),
-                success=True,
-                session_id=session_id,
-            )
-
-        except BlockError as e:
-            logger.warning(f"Block execution failed: {e}")
-            return ErrorResponse(
-                message=f"Block execution failed: {e}",
-                error=str(e),
-                session_id=session_id,
-            )
-        except Exception as e:
-            logger.error(f"Unexpected error executing block: {e}", exc_info=True)
-            return ErrorResponse(
-                message=f"Failed to execute block: {str(e)}",
-                error=str(e),
-                session_id=session_id,
-            )
-
-    def _get_inputs_list(self, block: Any) -> list[dict[str, Any]]:
-        """Extract non-credential inputs from block schema."""
-        inputs_list = []
-        schema = block.input_schema.jsonschema()
-        properties = schema.get("properties", {})
-        required_fields = set(schema.get("required", []))
-
-        # Get credential field names to exclude
-        credentials_fields = set(block.input_schema.get_credentials_fields().keys())
-
-        for field_name, field_schema in properties.items():
-            # Skip credential fields
-            if field_name in credentials_fields:
-                continue
-
-            inputs_list.append(
-                {
-                    "name": field_name,
-                    "title": field_schema.get("title", field_name),
-                    "type": field_schema.get("type", "string"),
-                    "description": field_schema.get("description", ""),
-                    "required": field_name in required_fields,
-                }
-            )
-
-        return inputs_list

autogpt_platform/backend/backend/api/features/chat/tools/search_docs.py

@@ -1,208 +0,0 @@
-"""SearchDocsTool - Search documentation using hybrid search."""
-
-import logging
-from typing import Any
-
-from prisma.enums import ContentType
-
-from backend.api.features.chat.model import ChatSession
-from backend.api.features.chat.tools.base import BaseTool
-from backend.api.features.chat.tools.models import (
-    DocSearchResult,
-    DocSearchResultsResponse,
-    ErrorResponse,
-    NoResultsResponse,
-    ToolResponseBase,
-)
-from backend.api.features.store.hybrid_search import unified_hybrid_search
-
-logger = logging.getLogger(__name__)
-
-# Base URL for documentation (can be configured)
-DOCS_BASE_URL = "https://docs.agpt.co"
-
-# Maximum number of results to return
-MAX_RESULTS = 5
-
-# Snippet length for preview
-SNIPPET_LENGTH = 200
-
-
-class SearchDocsTool(BaseTool):
-    """Tool for searching AutoGPT platform documentation."""
-
-    @property
-    def name(self) -> str:
-        return "search_docs"
-
-    @property
-    def description(self) -> str:
-        return (
-            "Search the AutoGPT platform documentation for information about "
-            "how to use the platform, build agents, configure blocks, and more. "
-            "Returns relevant documentation sections. Use get_doc_page to read full content."
-        )
-
-    @property
-    def parameters(self) -> dict[str, Any]:
-        return {
-            "type": "object",
-            "properties": {
-                "query": {
-                    "type": "string",
-                    "description": (
-                        "Search query to find relevant documentation. "
-                        "Use natural language to describe what you're looking for."
-                    ),
-                },
-            },
-            "required": ["query"],
-        }
-
-    @property
-    def requires_auth(self) -> bool:
-        return False  # Documentation is public
-
-    def _create_snippet(self, content: str, max_length: int = SNIPPET_LENGTH) -> str:
-        """Create a short snippet from content for preview."""
-        # Remove markdown formatting for cleaner snippet
-        clean_content = content.replace("#", "").replace("*", "").replace("`", "")
-        # Remove extra whitespace
-        clean_content = " ".join(clean_content.split())
-
-        if len(clean_content) <= max_length:
-            return clean_content
-
-        # Truncate at word boundary
-        truncated = clean_content[:max_length]
-        last_space = truncated.rfind(" ")
-        if last_space > max_length // 2:
-            truncated = truncated[:last_space]
-
-        return truncated + "..."
-
-    def _make_doc_url(self, path: str) -> str:
-        """Create a URL for a documentation page."""
-        # Remove file extension for URL
-        url_path = path.rsplit(".", 1)[0] if "." in path else path
-        return f"{DOCS_BASE_URL}/{url_path}"
-
-    async def _execute(
-        self,
-        user_id: str | None,
-        session: ChatSession,
-        **kwargs,
-    ) -> ToolResponseBase:
-        """Search documentation and return relevant sections.
-
-        Args:
-            user_id: User ID (not required for docs)
-            session: Chat session
-            query: Search query
-
-        Returns:
-            DocSearchResultsResponse: List of matching documentation sections
-            NoResultsResponse: No results found
-            ErrorResponse: Error message
-        """
-        query = kwargs.get("query", "").strip()
-        session_id = session.session_id if session else None
-
-        if not query:
-            return ErrorResponse(
-                message="Please provide a search query.",
-                error="Missing query parameter",
-                session_id=session_id,
-            )
-
-        try:
-            # Search using hybrid search for DOCUMENTATION content type only
-            results, total = await unified_hybrid_search(
-                query=query,
-                content_types=[ContentType.DOCUMENTATION],
-                page=1,
-                page_size=MAX_RESULTS * 2,  # Fetch extra for deduplication
-                min_score=0.1,  # Lower threshold for docs
-            )
-
-            if not results:
-                return NoResultsResponse(
-                    message=f"No documentation found for '{query}'.",
-                    suggestions=[
-                        "Try different keywords",
-                        "Use more general terms",
-                        "Check for typos in your query",
-                    ],
-                    session_id=session_id,
-                )
-
-            # Deduplicate by document path (keep highest scoring section per doc)
-            seen_docs: dict[str, dict[str, Any]] = {}
-            for result in results:
-                metadata = result.get("metadata", {})
-                doc_path = metadata.get("path", "")
-
-                if not doc_path:
-                    continue
-
-                # Keep the highest scoring result for each document
-                if doc_path not in seen_docs:
-                    seen_docs[doc_path] = result
-                elif result.get("combined_score", 0) > seen_docs[doc_path].get(
-                    "combined_score", 0
-                ):
-                    seen_docs[doc_path] = result
-
-            # Sort by score and take top MAX_RESULTS
-            deduplicated = sorted(
-                seen_docs.values(),
-                key=lambda x: x.get("combined_score", 0),
-                reverse=True,
-            )[:MAX_RESULTS]
-
-            if not deduplicated:
-                return NoResultsResponse(
-                    message=f"No documentation found for '{query}'.",
-                    suggestions=[
-                        "Try different keywords",
-                        "Use more general terms",
-                    ],
-                    session_id=session_id,
-                )
-
-            # Build response
-            doc_results: list[DocSearchResult] = []
-            for result in deduplicated:
-                metadata = result.get("metadata", {})
-                doc_path = metadata.get("path", "")
-                doc_title = metadata.get("doc_title", "")
-                section_title = metadata.get("section_title", "")
-                searchable_text = result.get("searchable_text", "")
-                score = result.get("combined_score", 0)
-
-                doc_results.append(
-                    DocSearchResult(
-                        title=doc_title or section_title or doc_path,
-                        path=doc_path,
-                        section=section_title,
-                        snippet=self._create_snippet(searchable_text),
-                        score=round(score, 3),
-                        doc_url=self._make_doc_url(doc_path),
-                    )
-                )
-
-            return DocSearchResultsResponse(
-                message=f"Found {len(doc_results)} relevant documentation sections.",
-                results=doc_results,
-                count=len(doc_results),
-                query=query,
-                session_id=session_id,
-            )
-
-        except Exception as e:
-            logger.error(f"Documentation search failed: {e}")
-            return ErrorResponse(
-                message=f"Failed to search documentation: {str(e)}",
-                error="search_failed",
-                session_id=session_id,
-            )

autogpt_platform/backend/backend/api/features/oauth.py

@@ -1,833 +0,0 @@
-"""
-OAuth 2.0 Provider Endpoints
-
-Implements OAuth 2.0 Authorization Code flow with PKCE support.
-
-Flow:
-1. User clicks "Login with AutoGPT" in 3rd party app
-2. App redirects user to /auth/authorize with client_id, redirect_uri, scope, state
-3. User sees consent screen (if not already logged in, redirects to login first)
-4. User approves → backend creates authorization code
-5. User redirected back to app with code
-6. App exchanges code for access/refresh tokens at /api/oauth/token
-7. App uses access token to call external API endpoints
-"""
-
-import io
-import logging
-import os
-import uuid
-from datetime import datetime
-from typing import Literal, Optional
-from urllib.parse import urlencode
-
-from autogpt_libs.auth import get_user_id
-from fastapi import APIRouter, Body, HTTPException, Security, UploadFile, status
-from gcloud.aio import storage as async_storage
-from PIL import Image
-from prisma.enums import APIKeyPermission
-from pydantic import BaseModel, Field
-
-from backend.data.auth.oauth import (
-    InvalidClientError,
-    InvalidGrantError,
-    OAuthApplicationInfo,
-    TokenIntrospectionResult,
-    consume_authorization_code,
-    create_access_token,
-    create_authorization_code,
-    create_refresh_token,
-    get_oauth_application,
-    get_oauth_application_by_id,
-    introspect_token,
-    list_user_oauth_applications,
-    refresh_tokens,
-    revoke_access_token,
-    revoke_refresh_token,
-    update_oauth_application,
-    validate_client_credentials,
-    validate_redirect_uri,
-    validate_scopes,
-)
-from backend.util.settings import Settings
-from backend.util.virus_scanner import scan_content_safe
-
-settings = Settings()
-logger = logging.getLogger(__name__)
-
-router = APIRouter()
-
-
-# ============================================================================
-# Request/Response Models
-# ============================================================================
-
-
-class TokenResponse(BaseModel):
-    """OAuth 2.0 token response"""
-
-    token_type: Literal["Bearer"] = "Bearer"
-    access_token: str
-    access_token_expires_at: datetime
-    refresh_token: str
-    refresh_token_expires_at: datetime
-    scopes: list[str]
-
-
-class ErrorResponse(BaseModel):
-    """OAuth 2.0 error response"""
-
-    error: str
-    error_description: Optional[str] = None
-
-
-class OAuthApplicationPublicInfo(BaseModel):
-    """Public information about an OAuth application (for consent screen)"""
-
-    name: str
-    description: Optional[str] = None
-    logo_url: Optional[str] = None
-    scopes: list[str]
-
-
-# ============================================================================
-# Application Info Endpoint
-# ============================================================================
-
-
-@router.get(
-    "/app/{client_id}",
-    responses={
-        404: {"description": "Application not found or disabled"},
-    },
-)
-async def get_oauth_app_info(
-    client_id: str, user_id: str = Security(get_user_id)
-) -> OAuthApplicationPublicInfo:
-    """
-    Get public information about an OAuth application.
-
-    This endpoint is used by the consent screen to display application details
-    to the user before they authorize access.
-
-    Returns:
-    - name: Application name
-    - description: Application description (if provided)
-    - scopes: List of scopes the application is allowed to request
-    """
-    app = await get_oauth_application(client_id)
-    if not app or not app.is_active:
-        raise HTTPException(
-            status_code=status.HTTP_404_NOT_FOUND,
-            detail="Application not found",
-        )
-
-    return OAuthApplicationPublicInfo(
-        name=app.name,
-        description=app.description,
-        logo_url=app.logo_url,
-        scopes=[s.value for s in app.scopes],
-    )
-
-
-# ============================================================================
-# Authorization Endpoint
-# ============================================================================
-
-
-class AuthorizeRequest(BaseModel):
-    """OAuth 2.0 authorization request"""
-
-    client_id: str = Field(description="Client identifier")
-    redirect_uri: str = Field(description="Redirect URI")
-    scopes: list[str] = Field(description="List of scopes")
-    state: str = Field(description="Anti-CSRF token from client")
-    response_type: str = Field(
-        default="code", description="Must be 'code' for authorization code flow"
-    )
-    code_challenge: str = Field(description="PKCE code challenge (required)")
-    code_challenge_method: Literal["S256", "plain"] = Field(
-        default="S256", description="PKCE code challenge method (S256 recommended)"
-    )
-
-
-class AuthorizeResponse(BaseModel):
-    """OAuth 2.0 authorization response with redirect URL"""
-
-    redirect_url: str = Field(description="URL to redirect the user to")
-
-
-@router.post("/authorize")
-async def authorize(
-    request: AuthorizeRequest = Body(),
-    user_id: str = Security(get_user_id),
-) -> AuthorizeResponse:
-    """
-    OAuth 2.0 Authorization Endpoint
-
-    User must be logged in (authenticated with Supabase JWT).
-    This endpoint creates an authorization code and returns a redirect URL.
-
-    PKCE (Proof Key for Code Exchange) is REQUIRED for all authorization requests.
-
-    The frontend consent screen should call this endpoint after the user approves,
-    then redirect the user to the returned `redirect_url`.
-
-    Request Body:
-    - client_id: The OAuth application's client ID
-    - redirect_uri: Where to redirect after authorization (must match registered URI)
-    - scopes: List of permissions (e.g., "EXECUTE_GRAPH READ_GRAPH")
-    - state: Anti-CSRF token provided by client (will be returned in redirect)
-    - response_type: Must be "code" (for authorization code flow)
-    - code_challenge: PKCE code challenge (required)
-    - code_challenge_method: "S256" (recommended) or "plain"
-
-    Returns:
-    - redirect_url: The URL to redirect the user to (includes authorization code)
-
-    Error cases return a redirect_url with error parameters, or raise HTTPException
-    for critical errors (like invalid redirect_uri).
-    """
-    try:
-        # Validate response_type
-        if request.response_type != "code":
-            return _error_redirect_url(
-                request.redirect_uri,
-                request.state,
-                "unsupported_response_type",
-                "Only 'code' response type is supported",
-            )
-
-        # Get application
-        app = await get_oauth_application(request.client_id)
-        if not app:
-            return _error_redirect_url(
-                request.redirect_uri,
-                request.state,
-                "invalid_client",
-                "Unknown client_id",
-            )
-
-        if not app.is_active:
-            return _error_redirect_url(
-                request.redirect_uri,
-                request.state,
-                "invalid_client",
-                "Application is not active",
-            )
-
-        # Validate redirect URI
-        if not validate_redirect_uri(app, request.redirect_uri):
-            # For invalid redirect_uri, we can't redirect safely
-            # Must return error instead
-            raise HTTPException(
-                status_code=status.HTTP_400_BAD_REQUEST,
-                detail=(
-                    "Invalid redirect_uri. "
-                    f"Must be one of: {', '.join(app.redirect_uris)}"
-                ),
-            )
-
-        # Parse and validate scopes
-        try:
-            requested_scopes = [APIKeyPermission(s.strip()) for s in request.scopes]
-        except ValueError as e:
-            return _error_redirect_url(
-                request.redirect_uri,
-                request.state,
-                "invalid_scope",
-                f"Invalid scope: {e}",
-            )
-
-        if not requested_scopes:
-            return _error_redirect_url(
-                request.redirect_uri,
-                request.state,
-                "invalid_scope",
-                "At least one scope is required",
-            )
-
-        if not validate_scopes(app, requested_scopes):
-            return _error_redirect_url(
-                request.redirect_uri,
-                request.state,
-                "invalid_scope",
-                "Application is not authorized for all requested scopes. "
-                f"Allowed: {', '.join(s.value for s in app.scopes)}",
-            )
-
-        # Create authorization code
-        auth_code = await create_authorization_code(
-            application_id=app.id,
-            user_id=user_id,
-            scopes=requested_scopes,
-            redirect_uri=request.redirect_uri,
-            code_challenge=request.code_challenge,
-            code_challenge_method=request.code_challenge_method,
-        )
-
-        # Build redirect URL with authorization code
-        params = {
-            "code": auth_code.code,
-            "state": request.state,
-        }
-        redirect_url = f"{request.redirect_uri}?{urlencode(params)}"
-
-        logger.info(
-            f"Authorization code issued for user #{user_id} "
-            f"and app {app.name} (#{app.id})"
-        )
-
-        return AuthorizeResponse(redirect_url=redirect_url)
-
-    except HTTPException:
-        raise
-    except Exception as e:
-        logger.error(f"Error in authorization endpoint: {e}", exc_info=True)
-        return _error_redirect_url(
-            request.redirect_uri,
-            request.state,
-            "server_error",
-            "An unexpected error occurred",
-        )
-
-
-def _error_redirect_url(
-    redirect_uri: str,
-    state: str,
-    error: str,
-    error_description: Optional[str] = None,
-) -> AuthorizeResponse:
-    """Helper to build redirect URL with OAuth error parameters"""
-    params = {
-        "error": error,
-        "state": state,
-    }
-    if error_description:
-        params["error_description"] = error_description
-
-    redirect_url = f"{redirect_uri}?{urlencode(params)}"
-    return AuthorizeResponse(redirect_url=redirect_url)
-
-
-# ============================================================================
-# Token Endpoint
-# ============================================================================
-
-
-class TokenRequestByCode(BaseModel):
-    grant_type: Literal["authorization_code"]
-    code: str = Field(description="Authorization code")
-    redirect_uri: str = Field(
-        description="Redirect URI (must match authorization request)"
-    )
-    client_id: str
-    client_secret: str
-    code_verifier: str = Field(description="PKCE code verifier")
-
-
-class TokenRequestByRefreshToken(BaseModel):
-    grant_type: Literal["refresh_token"]
-    refresh_token: str
-    client_id: str
-    client_secret: str
-
-
-@router.post("/token")
-async def token(
-    request: TokenRequestByCode | TokenRequestByRefreshToken = Body(),
-) -> TokenResponse:
-    """
-    OAuth 2.0 Token Endpoint
-
-    Exchanges authorization code or refresh token for access token.
-
-    Grant Types:
-    1. authorization_code: Exchange authorization code for tokens
-       - Required: grant_type, code, redirect_uri, client_id, client_secret
-       - Optional: code_verifier (required if PKCE was used)
-
-    2. refresh_token: Exchange refresh token for new access token
-       - Required: grant_type, refresh_token, client_id, client_secret
-
-    Returns:
-    - access_token: Bearer token for API access (1 hour TTL)
-    - token_type: "Bearer"
-    - expires_in: Seconds until access token expires
-    - refresh_token: Token for refreshing access (30 days TTL)
-    - scopes: List of scopes
-    """
-    # Validate client credentials
-    try:
-        app = await validate_client_credentials(
-            request.client_id, request.client_secret
-        )
-    except InvalidClientError as e:
-        raise HTTPException(
-            status_code=status.HTTP_401_UNAUTHORIZED,
-            detail=str(e),
-        )
-
-    # Handle authorization_code grant
-    if request.grant_type == "authorization_code":
-        # Consume authorization code
-        try:
-            user_id, scopes = await consume_authorization_code(
-                code=request.code,
-                application_id=app.id,
-                redirect_uri=request.redirect_uri,
-                code_verifier=request.code_verifier,
-            )
-        except InvalidGrantError as e:
-            raise HTTPException(
-                status_code=status.HTTP_400_BAD_REQUEST,
-                detail=str(e),
-            )
-
-        # Create access and refresh tokens
-        access_token = await create_access_token(app.id, user_id, scopes)
-        refresh_token = await create_refresh_token(app.id, user_id, scopes)
-
-        logger.info(
-            f"Access token issued for user #{user_id} and app {app.name} (#{app.id})"
-            "via authorization code"
-        )
-
-        if not access_token.token or not refresh_token.token:
-            raise HTTPException(
-                status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
-                detail="Failed to generate tokens",
-            )
-
-        return TokenResponse(
-            token_type="Bearer",
-            access_token=access_token.token.get_secret_value(),
-            access_token_expires_at=access_token.expires_at,
-            refresh_token=refresh_token.token.get_secret_value(),
-            refresh_token_expires_at=refresh_token.expires_at,
-            scopes=list(s.value for s in scopes),
-        )
-
-    # Handle refresh_token grant
-    elif request.grant_type == "refresh_token":
-        # Refresh access token
-        try:
-            new_access_token, new_refresh_token = await refresh_tokens(
-                request.refresh_token, app.id
-            )
-        except InvalidGrantError as e:
-            raise HTTPException(
-                status_code=status.HTTP_400_BAD_REQUEST,
-                detail=str(e),
-            )
-
-        logger.info(
-            f"Tokens refreshed for user #{new_access_token.user_id} "
-            f"by app {app.name} (#{app.id})"
-        )
-
-        if not new_access_token.token or not new_refresh_token.token:
-            raise HTTPException(
-                status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
-                detail="Failed to generate tokens",
-            )
-
-        return TokenResponse(
-            token_type="Bearer",
-            access_token=new_access_token.token.get_secret_value(),
-            access_token_expires_at=new_access_token.expires_at,
-            refresh_token=new_refresh_token.token.get_secret_value(),
-            refresh_token_expires_at=new_refresh_token.expires_at,
-            scopes=list(s.value for s in new_access_token.scopes),
-        )
-
-    else:
-        raise HTTPException(
-            status_code=status.HTTP_400_BAD_REQUEST,
-            detail=f"Unsupported grant_type: {request.grant_type}. "
-            "Must be 'authorization_code' or 'refresh_token'",
-        )
-
-
-# ============================================================================
-# Token Introspection Endpoint
-# ============================================================================
-
-
-@router.post("/introspect")
-async def introspect(
-    token: str = Body(description="Token to introspect"),
-    token_type_hint: Optional[Literal["access_token", "refresh_token"]] = Body(
-        None, description="Hint about token type ('access_token' or 'refresh_token')"
-    ),
-    client_id: str = Body(description="Client identifier"),
-    client_secret: str = Body(description="Client secret"),
-) -> TokenIntrospectionResult:
-    """
-    OAuth 2.0 Token Introspection Endpoint (RFC 7662)
-
-    Allows clients to check if a token is valid and get its metadata.
-
-    Returns:
-    - active: Whether the token is currently active
-    - scopes: List of authorized scopes (if active)
-    - client_id: The client the token was issued to (if active)
-    - user_id: The user the token represents (if active)
-    - exp: Expiration timestamp (if active)
-    - token_type: "access_token" or "refresh_token" (if active)
-    """
-    # Validate client credentials
-    try:
-        await validate_client_credentials(client_id, client_secret)
-    except InvalidClientError as e:
-        raise HTTPException(
-            status_code=status.HTTP_401_UNAUTHORIZED,
-            detail=str(e),
-        )
-
-    # Introspect the token
-    return await introspect_token(token, token_type_hint)
-
-
-# ============================================================================
-# Token Revocation Endpoint
-# ============================================================================
-
-
-@router.post("/revoke")
-async def revoke(
-    token: str = Body(description="Token to revoke"),
-    token_type_hint: Optional[Literal["access_token", "refresh_token"]] = Body(
-        None, description="Hint about token type ('access_token' or 'refresh_token')"
-    ),
-    client_id: str = Body(description="Client identifier"),
-    client_secret: str = Body(description="Client secret"),
-):
-    """
-    OAuth 2.0 Token Revocation Endpoint (RFC 7009)
-
-    Allows clients to revoke an access or refresh token.
-
-    Note: Revoking a refresh token does NOT revoke associated access tokens.
-    Revoking an access token does NOT revoke the associated refresh token.
-    """
-    # Validate client credentials
-    try:
-        app = await validate_client_credentials(client_id, client_secret)
-    except InvalidClientError as e:
-        raise HTTPException(
-            status_code=status.HTTP_401_UNAUTHORIZED,
-            detail=str(e),
-        )
-
-    # Try to revoke as access token first
-    # Note: We pass app.id to ensure the token belongs to the authenticated app
-    if token_type_hint != "refresh_token":
-        revoked = await revoke_access_token(token, app.id)
-        if revoked:
-            logger.info(
-                f"Access token revoked for app {app.name} (#{app.id}); "
-                f"user #{revoked.user_id}"
-            )
-            return {"status": "ok"}
-
-    # Try to revoke as refresh token
-    revoked = await revoke_refresh_token(token, app.id)
-    if revoked:
-        logger.info(
-            f"Refresh token revoked for app {app.name} (#{app.id}); "
-            f"user #{revoked.user_id}"
-        )
-        return {"status": "ok"}
-
-    # Per RFC 7009, revocation endpoint returns 200 even if token not found
-    # or if token belongs to a different application.
-    # This prevents token scanning attacks.
-    logger.warning(f"Unsuccessful token revocation attempt by app {app.name} #{app.id}")
-    return {"status": "ok"}
-
-
-# ============================================================================
-# Application Management Endpoints (for app owners)
-# ============================================================================
-
-
-@router.get("/apps/mine")
-async def list_my_oauth_apps(
-    user_id: str = Security(get_user_id),
-) -> list[OAuthApplicationInfo]:
-    """
-    List all OAuth applications owned by the current user.
-
-    Returns a list of OAuth applications with their details including:
-    - id, name, description, logo_url
-    - client_id (public identifier)
-    - redirect_uris, grant_types, scopes
-    - is_active status
-    - created_at, updated_at timestamps
-
-    Note: client_secret is never returned for security reasons.
-    """
-    return await list_user_oauth_applications(user_id)
-
-
-@router.patch("/apps/{app_id}/status")
-async def update_app_status(
-    app_id: str,
-    user_id: str = Security(get_user_id),
-    is_active: bool = Body(description="Whether the app should be active", embed=True),
-) -> OAuthApplicationInfo:
-    """
-    Enable or disable an OAuth application.
-
-    Only the application owner can update the status.
-    When disabled, the application cannot be used for new authorizations
-    and existing access tokens will fail validation.
-
-    Returns the updated application info.
-    """
-    updated_app = await update_oauth_application(
-        app_id=app_id,
-        owner_id=user_id,
-        is_active=is_active,
-    )
-
-    if not updated_app:
-        raise HTTPException(
-            status_code=status.HTTP_404_NOT_FOUND,
-            detail="Application not found or you don't have permission to update it",
-        )
-
-    action = "enabled" if is_active else "disabled"
-    logger.info(f"OAuth app {updated_app.name} (#{app_id}) {action} by user #{user_id}")
-
-    return updated_app
-
-
-class UpdateAppLogoRequest(BaseModel):
-    logo_url: str = Field(description="URL of the uploaded logo image")
-
-
-@router.patch("/apps/{app_id}/logo")
-async def update_app_logo(
-    app_id: str,
-    request: UpdateAppLogoRequest = Body(),
-    user_id: str = Security(get_user_id),
-) -> OAuthApplicationInfo:
-    """
-    Update the logo URL for an OAuth application.
-
-    Only the application owner can update the logo.
-    The logo should be uploaded first using the media upload endpoint,
-    then this endpoint is called with the resulting URL.
-
-    Logo requirements:
-    - Must be square (1:1 aspect ratio)
-    - Minimum 512x512 pixels
-    - Maximum 2048x2048 pixels
-
-    Returns the updated application info.
-    """
-    if (
-        not (app := await get_oauth_application_by_id(app_id))
-        or app.owner_id != user_id
-    ):
-        raise HTTPException(
-            status_code=status.HTTP_404_NOT_FOUND,
-            detail="OAuth App not found",
-        )
-
-    # Delete the current app logo file (if any and it's in our cloud storage)
-    await _delete_app_current_logo_file(app)
-
-    updated_app = await update_oauth_application(
-        app_id=app_id,
-        owner_id=user_id,
-        logo_url=request.logo_url,
-    )
-
-    if not updated_app:
-        raise HTTPException(
-            status_code=status.HTTP_404_NOT_FOUND,
-            detail="Application not found or you don't have permission to update it",
-        )
-
-    logger.info(
-        f"OAuth app {updated_app.name} (#{app_id}) logo updated by user #{user_id}"
-    )
-
-    return updated_app
-
-
-# Logo upload constraints
-LOGO_MIN_SIZE = 512
-LOGO_MAX_SIZE = 2048
-LOGO_ALLOWED_TYPES = {"image/jpeg", "image/png", "image/webp"}
-LOGO_MAX_FILE_SIZE = 3 * 1024 * 1024  # 3MB
-
-
-@router.post("/apps/{app_id}/logo/upload")
-async def upload_app_logo(
-    app_id: str,
-    file: UploadFile,
-    user_id: str = Security(get_user_id),
-) -> OAuthApplicationInfo:
-    """
-    Upload a logo image for an OAuth application.
-
-    Requirements:
-    - Image must be square (1:1 aspect ratio)
-    - Minimum 512x512 pixels
-    - Maximum 2048x2048 pixels
-    - Allowed formats: JPEG, PNG, WebP
-    - Maximum file size: 3MB
-
-    The image is uploaded to cloud storage and the app's logoUrl is updated.
-    Returns the updated application info.
-    """
-    # Verify ownership to reduce vulnerability to DoS(torage) or DoM(oney) attacks
-    if (
-        not (app := await get_oauth_application_by_id(app_id))
-        or app.owner_id != user_id
-    ):
-        raise HTTPException(
-            status_code=status.HTTP_404_NOT_FOUND,
-            detail="OAuth App not found",
-        )
-
-    # Check GCS configuration
-    if not settings.config.media_gcs_bucket_name:
-        raise HTTPException(
-            status_code=status.HTTP_503_SERVICE_UNAVAILABLE,
-            detail="Media storage is not configured",
-        )
-
-    # Validate content type
-    content_type = file.content_type
-    if content_type not in LOGO_ALLOWED_TYPES:
-        raise HTTPException(
-            status_code=status.HTTP_400_BAD_REQUEST,
-            detail=f"Invalid file type. Allowed: JPEG, PNG, WebP. Got: {content_type}",
-        )
-
-    # Read file content
-    try:
-        file_bytes = await file.read()
-    except Exception as e:
-        logger.error(f"Error reading logo file: {e}")
-        raise HTTPException(
-            status_code=status.HTTP_400_BAD_REQUEST,
-            detail="Failed to read uploaded file",
-        )
-
-    # Check file size
-    if len(file_bytes) > LOGO_MAX_FILE_SIZE:
-        raise HTTPException(
-            status_code=status.HTTP_400_BAD_REQUEST,
-            detail=(
-                "File too large. "
-                f"Maximum size is {LOGO_MAX_FILE_SIZE // 1024 // 1024}MB"
-            ),
-        )
-
-    # Validate image dimensions
-    try:
-        image = Image.open(io.BytesIO(file_bytes))
-        width, height = image.size
-
-        if width != height:
-            raise HTTPException(
-                status_code=status.HTTP_400_BAD_REQUEST,
-                detail=f"Logo must be square. Got {width}x{height}",
-            )
-
-        if width < LOGO_MIN_SIZE:
-            raise HTTPException(
-                status_code=status.HTTP_400_BAD_REQUEST,
-                detail=f"Logo too small. Minimum {LOGO_MIN_SIZE}x{LOGO_MIN_SIZE}. "
-                f"Got {width}x{height}",
-            )
-
-        if width > LOGO_MAX_SIZE:
-            raise HTTPException(
-                status_code=status.HTTP_400_BAD_REQUEST,
-                detail=f"Logo too large. Maximum {LOGO_MAX_SIZE}x{LOGO_MAX_SIZE}. "
-                f"Got {width}x{height}",
-            )
-    except HTTPException:
-        raise
-    except Exception as e:
-        logger.error(f"Error validating logo image: {e}")
-        raise HTTPException(
-            status_code=status.HTTP_400_BAD_REQUEST,
-            detail="Invalid image file",
-        )
-
-    # Scan for viruses
-    filename = file.filename or "logo"
-    await scan_content_safe(file_bytes, filename=filename)
-
-    # Generate unique filename
-    file_ext = os.path.splitext(filename)[1].lower() or ".png"
-    unique_filename = f"{uuid.uuid4()}{file_ext}"
-    storage_path = f"oauth-apps/{app_id}/logo/{unique_filename}"
-
-    # Upload to GCS
-    try:
-        async with async_storage.Storage() as async_client:
-            bucket_name = settings.config.media_gcs_bucket_name
-
-            await async_client.upload(
-                bucket_name, storage_path, file_bytes, content_type=content_type
-            )
-
-            logo_url = f"https://storage.googleapis.com/{bucket_name}/{storage_path}"
-    except Exception as e:
-        logger.error(f"Error uploading logo to GCS: {e}")
-        raise HTTPException(
-            status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
-            detail="Failed to upload logo",
-        )
-
-    # Delete the current app logo file (if any and it's in our cloud storage)
-    await _delete_app_current_logo_file(app)
-
-    # Update the app with the new logo URL
-    updated_app = await update_oauth_application(
-        app_id=app_id,
-        owner_id=user_id,
-        logo_url=logo_url,
-    )
-
-    if not updated_app:
-        raise HTTPException(
-            status_code=status.HTTP_404_NOT_FOUND,
-            detail="Application not found or you don't have permission to update it",
-        )
-
-    logger.info(
-        f"OAuth app {updated_app.name} (#{app_id}) logo uploaded by user #{user_id}"
-    )
-
-    return updated_app
-
-
-async def _delete_app_current_logo_file(app: OAuthApplicationInfo):
-    """
-    Delete the current logo file for the given app, if there is one in our cloud storage
-    """
-    bucket_name = settings.config.media_gcs_bucket_name
-    storage_base_url = f"https://storage.googleapis.com/{bucket_name}/"
-
-    if app.logo_url and app.logo_url.startswith(storage_base_url):
-        # Parse blob path from URL: https://storage.googleapis.com/{bucket}/{path}
-        old_path = app.logo_url.replace(storage_base_url, "")
-        try:
-            async with async_storage.Storage() as async_client:
-                await async_client.delete(bucket_name, old_path)
-            logger.info(f"Deleted old logo for OAuth app #{app.id}: {old_path}")
-        except Exception as e:
-            # Log but don't fail - the new logo was uploaded successfully
-            logger.warning(
-                f"Failed to delete old logo for OAuth app #{app.id}: {e}", exc_info=e
-            )

autogpt_platform/backend/backend/api/features/store/content_handlers.py

@@ -1,610 +0,0 @@
-"""
-Content Type Handlers for Unified Embeddings
-
-Pluggable system for different content sources (store agents, blocks, docs).
-Each handler knows how to fetch and process its content type for embedding.
-"""
-
-import logging
-from abc import ABC, abstractmethod
-from dataclasses import dataclass
-from pathlib import Path
-from typing import Any
-
-from prisma.enums import ContentType
-
-from backend.data.db import query_raw_with_schema
-
-logger = logging.getLogger(__name__)
-
-
-@dataclass
-class ContentItem:
-    """Represents a piece of content to be embedded."""
-
-    content_id: str  # Unique identifier (DB ID or file path)
-    content_type: ContentType
-    searchable_text: str  # Combined text for embedding
-    metadata: dict[str, Any]  # Content-specific metadata
-    user_id: str | None = None  # For user-scoped content
-
-
-class ContentHandler(ABC):
-    """Base handler for fetching and processing content for embeddings."""
-
-    @property
-    @abstractmethod
-    def content_type(self) -> ContentType:
-        """The ContentType this handler manages."""
-        pass
-
-    @abstractmethod
-    async def get_missing_items(self, batch_size: int) -> list[ContentItem]:
-        """
-        Fetch items that don't have embeddings yet.
-
-        Args:
-            batch_size: Maximum number of items to return
-
-        Returns:
-            List of ContentItem objects ready for embedding
-        """
-        pass
-
-    @abstractmethod
-    async def get_stats(self) -> dict[str, int]:
-        """
-        Get statistics about embedding coverage.
-
-        Returns:
-            Dict with keys: total, with_embeddings, without_embeddings
-        """
-        pass
-
-
-class StoreAgentHandler(ContentHandler):
-    """Handler for marketplace store agent listings."""
-
-    @property
-    def content_type(self) -> ContentType:
-        return ContentType.STORE_AGENT
-
-    async def get_missing_items(self, batch_size: int) -> list[ContentItem]:
-        """Fetch approved store listings without embeddings."""
-        from backend.api.features.store.embeddings import build_searchable_text
-
-        missing = await query_raw_with_schema(
-            """
-            SELECT
-                slv.id,
-                slv.name,
-                slv.description,
-                slv."subHeading",
-                slv.categories
-            FROM {schema_prefix}"StoreListingVersion" slv
-            LEFT JOIN {schema_prefix}"UnifiedContentEmbedding" uce
-                ON slv.id = uce."contentId" AND uce."contentType" = 'STORE_AGENT'::{schema_prefix}"ContentType"
-            WHERE slv."submissionStatus" = 'APPROVED'
-            AND slv."isDeleted" = false
-            AND uce."contentId" IS NULL
-            LIMIT $1
-            """,
-            batch_size,
-        )
-
-        return [
-            ContentItem(
-                content_id=row["id"],
-                content_type=ContentType.STORE_AGENT,
-                searchable_text=build_searchable_text(
-                    name=row["name"],
-                    description=row["description"],
-                    sub_heading=row["subHeading"],
-                    categories=row["categories"] or [],
-                ),
-                metadata={
-                    "name": row["name"],
-                    "categories": row["categories"] or [],
-                },
-                user_id=None,  # Store agents are public
-            )
-            for row in missing
-        ]
-
-    async def get_stats(self) -> dict[str, int]:
-        """Get statistics about store agent embedding coverage."""
-        # Count approved versions
-        approved_result = await query_raw_with_schema(
-            """
-            SELECT COUNT(*) as count
-            FROM {schema_prefix}"StoreListingVersion"
-            WHERE "submissionStatus" = 'APPROVED'
-            AND "isDeleted" = false
-            """
-        )
-        total_approved = approved_result[0]["count"] if approved_result else 0
-
-        # Count versions with embeddings
-        embedded_result = await query_raw_with_schema(
-            """
-            SELECT COUNT(*) as count
-            FROM {schema_prefix}"StoreListingVersion" slv
-            JOIN {schema_prefix}"UnifiedContentEmbedding" uce ON slv.id = uce."contentId" AND uce."contentType" = 'STORE_AGENT'::{schema_prefix}"ContentType"
-            WHERE slv."submissionStatus" = 'APPROVED'
-            AND slv."isDeleted" = false
-            """
-        )
-        with_embeddings = embedded_result[0]["count"] if embedded_result else 0
-
-        return {
-            "total": total_approved,
-            "with_embeddings": with_embeddings,
-            "without_embeddings": total_approved - with_embeddings,
-        }
-
-
-class BlockHandler(ContentHandler):
-    """Handler for block definitions (Python classes)."""
-
-    @property
-    def content_type(self) -> ContentType:
-        return ContentType.BLOCK
-
-    async def get_missing_items(self, batch_size: int) -> list[ContentItem]:
-        """Fetch blocks without embeddings."""
-        from backend.data.block import get_blocks
-
-        # Get all available blocks
-        all_blocks = get_blocks()
-
-        # Check which ones have embeddings
-        if not all_blocks:
-            return []
-
-        block_ids = list(all_blocks.keys())
-
-        # Query for existing embeddings
-        placeholders = ",".join([f"${i+1}" for i in range(len(block_ids))])
-        existing_result = await query_raw_with_schema(
-            f"""
-            SELECT "contentId"
-            FROM {{schema_prefix}}"UnifiedContentEmbedding"
-            WHERE "contentType" = 'BLOCK'::{{schema_prefix}}"ContentType"
-            AND "contentId" = ANY(ARRAY[{placeholders}])
-            """,
-            *block_ids,
-        )
-
-        existing_ids = {row["contentId"] for row in existing_result}
-        missing_blocks = [
-            (block_id, block_cls)
-            for block_id, block_cls in all_blocks.items()
-            if block_id not in existing_ids
-        ]
-
-        # Convert to ContentItem
-        items = []
-        for block_id, block_cls in missing_blocks[:batch_size]:
-            try:
-                block_instance = block_cls()
-
-                # Build searchable text from block metadata
-                parts = []
-                if hasattr(block_instance, "name") and block_instance.name:
-                    parts.append(block_instance.name)
-                if (
-                    hasattr(block_instance, "description")
-                    and block_instance.description
-                ):
-                    parts.append(block_instance.description)
-                if hasattr(block_instance, "categories") and block_instance.categories:
-                    # Convert BlockCategory enum to strings
-                    parts.append(
-                        " ".join(str(cat.value) for cat in block_instance.categories)
-                    )
-
-                # Add input/output schema info
-                if hasattr(block_instance, "input_schema"):
-                    schema = block_instance.input_schema
-                    if hasattr(schema, "model_json_schema"):
-                        schema_dict = schema.model_json_schema()
-                        if "properties" in schema_dict:
-                            for prop_name, prop_info in schema_dict[
-                                "properties"
-                            ].items():
-                                if "description" in prop_info:
-                                    parts.append(
-                                        f"{prop_name}: {prop_info['description']}"
-                                    )
-
-                searchable_text = " ".join(parts)
-
-                # Convert categories set of enums to list of strings for JSON serialization
-                categories = getattr(block_instance, "categories", set())
-                categories_list = (
-                    [cat.value for cat in categories] if categories else []
-                )
-
-                items.append(
-                    ContentItem(
-                        content_id=block_id,
-                        content_type=ContentType.BLOCK,
-                        searchable_text=searchable_text,
-                        metadata={
-                            "name": getattr(block_instance, "name", ""),
-                            "categories": categories_list,
-                        },
-                        user_id=None,  # Blocks are public
-                    )
-                )
-            except Exception as e:
-                logger.warning(f"Failed to process block {block_id}: {e}")
-                continue
-
-        return items
-
-    async def get_stats(self) -> dict[str, int]:
-        """Get statistics about block embedding coverage."""
-        from backend.data.block import get_blocks
-
-        all_blocks = get_blocks()
-        total_blocks = len(all_blocks)
-
-        if total_blocks == 0:
-            return {"total": 0, "with_embeddings": 0, "without_embeddings": 0}
-
-        block_ids = list(all_blocks.keys())
-        placeholders = ",".join([f"${i+1}" for i in range(len(block_ids))])
-
-        embedded_result = await query_raw_with_schema(
-            f"""
-            SELECT COUNT(*) as count
-            FROM {{schema_prefix}}"UnifiedContentEmbedding"
-            WHERE "contentType" = 'BLOCK'::{{schema_prefix}}"ContentType"
-            AND "contentId" = ANY(ARRAY[{placeholders}])
-            """,
-            *block_ids,
-        )
-
-        with_embeddings = embedded_result[0]["count"] if embedded_result else 0
-
-        return {
-            "total": total_blocks,
-            "with_embeddings": with_embeddings,
-            "without_embeddings": total_blocks - with_embeddings,
-        }
-
-
-@dataclass
-class MarkdownSection:
-    """Represents a section of a markdown document."""
-
-    title: str  # Section heading text
-    content: str  # Section content (including the heading line)
-    level: int  # Heading level (1 for #, 2 for ##, etc.)
-    index: int  # Section index within the document
-
-
-class DocumentationHandler(ContentHandler):
-    """Handler for documentation files (.md/.mdx).
-
-    Chunks documents by markdown headings to create multiple embeddings per file.
-    Each section (## heading) becomes a separate embedding for better retrieval.
-    """
-
-    @property
-    def content_type(self) -> ContentType:
-        return ContentType.DOCUMENTATION
-
-    def _get_docs_root(self) -> Path:
-        """Get the documentation root directory."""
-        # content_handlers.py is at: backend/backend/api/features/store/content_handlers.py
-        # Need to go up to project root then into docs/
-        # In container: /app/autogpt_platform/backend/backend/api/features/store -> /app/docs
-        # In development: /repo/autogpt_platform/backend/backend/api/features/store -> /repo/docs
-        this_file = Path(
-            __file__
-        )  # .../backend/backend/api/features/store/content_handlers.py
-        project_root = (
-            this_file.parent.parent.parent.parent.parent.parent.parent
-        )  # -> /app or /repo
-        docs_root = project_root / "docs"
-        return docs_root
-
-    def _extract_doc_title(self, file_path: Path) -> str:
-        """Extract the document title from a markdown file."""
-        try:
-            content = file_path.read_text(encoding="utf-8")
-            lines = content.split("\n")
-
-            # Try to extract title from first # heading
-            for line in lines:
-                if line.startswith("# "):
-                    return line[2:].strip()
-
-            # If no title found, use filename
-            return file_path.stem.replace("-", " ").replace("_", " ").title()
-        except Exception as e:
-            logger.warning(f"Failed to read title from {file_path}: {e}")
-            return file_path.stem.replace("-", " ").replace("_", " ").title()
-
-    def _chunk_markdown_by_headings(
-        self, file_path: Path, min_heading_level: int = 2
-    ) -> list[MarkdownSection]:
-        """
-        Split a markdown file into sections based on headings.
-
-        Args:
-            file_path: Path to the markdown file
-            min_heading_level: Minimum heading level to split on (default: 2 for ##)
-
-        Returns:
-            List of MarkdownSection objects, one per section.
-            If no headings found, returns a single section with all content.
-        """
-        try:
-            content = file_path.read_text(encoding="utf-8")
-        except Exception as e:
-            logger.warning(f"Failed to read {file_path}: {e}")
-            return []
-
-        lines = content.split("\n")
-        sections: list[MarkdownSection] = []
-        current_section_lines: list[str] = []
-        current_title = ""
-        current_level = 0
-        section_index = 0
-        doc_title = ""
-
-        for line in lines:
-            # Check if line is a heading
-            if line.startswith("#"):
-                # Count heading level
-                level = 0
-                for char in line:
-                    if char == "#":
-                        level += 1
-                    else:
-                        break
-
-                heading_text = line[level:].strip()
-
-                # Track document title (level 1 heading)
-                if level == 1 and not doc_title:
-                    doc_title = heading_text
-                    # Don't create a section for just the title - add it to first section
-                    current_section_lines.append(line)
-                    continue
-
-                # Check if this heading should start a new section
-                if level >= min_heading_level:
-                    # Save previous section if it has content
-                    if current_section_lines:
-                        section_content = "\n".join(current_section_lines).strip()
-                        if section_content:
-                            # Use doc title for first section if no specific title
-                            title = current_title if current_title else doc_title
-                            if not title:
-                                title = file_path.stem.replace("-", " ").replace(
-                                    "_", " "
-                                )
-                            sections.append(
-                                MarkdownSection(
-                                    title=title,
-                                    content=section_content,
-                                    level=current_level if current_level else 1,
-                                    index=section_index,
-                                )
-                            )
-                            section_index += 1
-
-                    # Start new section
-                    current_section_lines = [line]
-                    current_title = heading_text
-                    current_level = level
-                else:
-                    # Lower level heading (e.g., # when splitting on ##)
-                    current_section_lines.append(line)
-            else:
-                current_section_lines.append(line)
-
-        # Don't forget the last section
-        if current_section_lines:
-            section_content = "\n".join(current_section_lines).strip()
-            if section_content:
-                title = current_title if current_title else doc_title
-                if not title:
-                    title = file_path.stem.replace("-", " ").replace("_", " ")
-                sections.append(
-                    MarkdownSection(
-                        title=title,
-                        content=section_content,
-                        level=current_level if current_level else 1,
-                        index=section_index,
-                    )
-                )
-
-        # If no sections were created (no headings found), create one section with all content
-        if not sections and content.strip():
-            title = (
-                doc_title
-                if doc_title
-                else file_path.stem.replace("-", " ").replace("_", " ")
-            )
-            sections.append(
-                MarkdownSection(
-                    title=title,
-                    content=content.strip(),
-                    level=1,
-                    index=0,
-                )
-            )
-
-        return sections
-
-    def _make_section_content_id(self, doc_path: str, section_index: int) -> str:
-        """Create a unique content ID for a document section.
-
-        Format: doc_path::section_index
-        Example: 'platform/getting-started.md::0'
-        """
-        return f"{doc_path}::{section_index}"
-
-    def _parse_section_content_id(self, content_id: str) -> tuple[str, int]:
-        """Parse a section content ID back into doc_path and section_index.
-
-        Returns: (doc_path, section_index)
-        """
-        if "::" in content_id:
-            parts = content_id.rsplit("::", 1)
-            return parts[0], int(parts[1])
-        # Legacy format (whole document)
-        return content_id, 0
-
-    async def get_missing_items(self, batch_size: int) -> list[ContentItem]:
-        """Fetch documentation sections without embeddings.
-
-        Chunks each document by markdown headings and creates embeddings for each section.
-        Content IDs use the format: 'path/to/doc.md::section_index'
-        """
-        docs_root = self._get_docs_root()
-
-        if not docs_root.exists():
-            logger.warning(f"Documentation root not found: {docs_root}")
-            return []
-
-        # Find all .md and .mdx files
-        all_docs = list(docs_root.rglob("*.md")) + list(docs_root.rglob("*.mdx"))
-
-        if not all_docs:
-            return []
-
-        # Build list of all sections from all documents
-        all_sections: list[tuple[str, Path, MarkdownSection]] = []
-        for doc_file in all_docs:
-            doc_path = str(doc_file.relative_to(docs_root))
-            sections = self._chunk_markdown_by_headings(doc_file)
-            for section in sections:
-                all_sections.append((doc_path, doc_file, section))
-
-        if not all_sections:
-            return []
-
-        # Generate content IDs for all sections
-        section_content_ids = [
-            self._make_section_content_id(doc_path, section.index)
-            for doc_path, _, section in all_sections
-        ]
-
-        # Check which ones have embeddings
-        placeholders = ",".join([f"${i+1}" for i in range(len(section_content_ids))])
-        existing_result = await query_raw_with_schema(
-            f"""
-            SELECT "contentId"
-            FROM {{schema_prefix}}"UnifiedContentEmbedding"
-            WHERE "contentType" = 'DOCUMENTATION'::{{schema_prefix}}"ContentType"
-            AND "contentId" = ANY(ARRAY[{placeholders}])
-            """,
-            *section_content_ids,
-        )
-
-        existing_ids = {row["contentId"] for row in existing_result}
-
-        # Filter to missing sections
-        missing_sections = [
-            (doc_path, doc_file, section, content_id)
-            for (doc_path, doc_file, section), content_id in zip(
-                all_sections, section_content_ids
-            )
-            if content_id not in existing_ids
-        ]
-
-        # Convert to ContentItem (up to batch_size)
-        items = []
-        for doc_path, doc_file, section, content_id in missing_sections[:batch_size]:
-            try:
-                # Get document title for context
-                doc_title = self._extract_doc_title(doc_file)
-
-                # Build searchable text with context
-                # Include doc title and section title for better search relevance
-                searchable_text = f"{doc_title} - {section.title}\n\n{section.content}"
-
-                items.append(
-                    ContentItem(
-                        content_id=content_id,
-                        content_type=ContentType.DOCUMENTATION,
-                        searchable_text=searchable_text,
-                        metadata={
-                            "doc_title": doc_title,
-                            "section_title": section.title,
-                            "section_index": section.index,
-                            "heading_level": section.level,
-                            "path": doc_path,
-                        },
-                        user_id=None,  # Documentation is public
-                    )
-                )
-            except Exception as e:
-                logger.warning(f"Failed to process section {content_id}: {e}")
-                continue
-
-        return items
-
-    def _get_all_section_content_ids(self, docs_root: Path) -> set[str]:
-        """Get all current section content IDs from the docs directory.
-
-        Used for stats and cleanup to know what sections should exist.
-        """
-        all_docs = list(docs_root.rglob("*.md")) + list(docs_root.rglob("*.mdx"))
-        content_ids = set()
-
-        for doc_file in all_docs:
-            doc_path = str(doc_file.relative_to(docs_root))
-            sections = self._chunk_markdown_by_headings(doc_file)
-            for section in sections:
-                content_ids.add(self._make_section_content_id(doc_path, section.index))
-
-        return content_ids
-
-    async def get_stats(self) -> dict[str, int]:
-        """Get statistics about documentation embedding coverage.
-
-        Counts sections (not documents) since each section gets its own embedding.
-        """
-        docs_root = self._get_docs_root()
-
-        if not docs_root.exists():
-            return {"total": 0, "with_embeddings": 0, "without_embeddings": 0}
-
-        # Get all section content IDs
-        all_section_ids = self._get_all_section_content_ids(docs_root)
-        total_sections = len(all_section_ids)
-
-        if total_sections == 0:
-            return {"total": 0, "with_embeddings": 0, "without_embeddings": 0}
-
-        # Count embeddings in database for DOCUMENTATION type
-        embedded_result = await query_raw_with_schema(
-            """
-            SELECT COUNT(*) as count
-            FROM {schema_prefix}"UnifiedContentEmbedding"
-            WHERE "contentType" = 'DOCUMENTATION'::{schema_prefix}"ContentType"
-            """
-        )
-
-        with_embeddings = embedded_result[0]["count"] if embedded_result else 0
-
-        return {
-            "total": total_sections,
-            "with_embeddings": with_embeddings,
-            "without_embeddings": total_sections - with_embeddings,
-        }
-
-
-# Content handler registry
-CONTENT_HANDLERS: dict[ContentType, ContentHandler] = {
-    ContentType.STORE_AGENT: StoreAgentHandler(),
-    ContentType.BLOCK: BlockHandler(),
-    ContentType.DOCUMENTATION: DocumentationHandler(),
-}

autogpt_platform/backend/backend/api/features/store/content_handlers_integration_test.py

@@ -1,215 +0,0 @@
-"""
-Integration tests for content handlers using real DB.
-
-Run with: poetry run pytest backend/api/features/store/content_handlers_integration_test.py -xvs
-
-These tests use the real database but mock OpenAI calls.
-"""
-
-from unittest.mock import patch
-
-import pytest
-
-from backend.api.features.store.content_handlers import (
-    CONTENT_HANDLERS,
-    BlockHandler,
-    DocumentationHandler,
-    StoreAgentHandler,
-)
-from backend.api.features.store.embeddings import (
-    EMBEDDING_DIM,
-    backfill_all_content_types,
-    ensure_content_embedding,
-    get_embedding_stats,
-)
-
-
-@pytest.mark.asyncio(loop_scope="session")
-async def test_store_agent_handler_real_db():
-    """Test StoreAgentHandler with real database queries."""
-    handler = StoreAgentHandler()
-
-    # Get stats from real DB
-    stats = await handler.get_stats()
-
-    # Stats should have correct structure
-    assert "total" in stats
-    assert "with_embeddings" in stats
-    assert "without_embeddings" in stats
-    assert stats["total"] >= 0
-    assert stats["with_embeddings"] >= 0
-    assert stats["without_embeddings"] >= 0
-
-    # Get missing items (max 1 to keep test fast)
-    items = await handler.get_missing_items(batch_size=1)
-
-    # Items should be list (may be empty if all have embeddings)
-    assert isinstance(items, list)
-
-    if items:
-        item = items[0]
-        assert item.content_id is not None
-        assert item.content_type.value == "STORE_AGENT"
-        assert item.searchable_text != ""
-        assert item.user_id is None
-
-
-@pytest.mark.asyncio(loop_scope="session")
-async def test_block_handler_real_db():
-    """Test BlockHandler with real database queries."""
-    handler = BlockHandler()
-
-    # Get stats from real DB
-    stats = await handler.get_stats()
-
-    # Stats should have correct structure
-    assert "total" in stats
-    assert "with_embeddings" in stats
-    assert "without_embeddings" in stats
-    assert stats["total"] >= 0  # Should have at least some blocks
-    assert stats["with_embeddings"] >= 0
-    assert stats["without_embeddings"] >= 0
-
-    # Get missing items (max 1 to keep test fast)
-    items = await handler.get_missing_items(batch_size=1)
-
-    # Items should be list
-    assert isinstance(items, list)
-
-    if items:
-        item = items[0]
-        assert item.content_id is not None  # Should be block UUID
-        assert item.content_type.value == "BLOCK"
-        assert item.searchable_text != ""
-        assert item.user_id is None
-
-
-@pytest.mark.asyncio(loop_scope="session")
-async def test_documentation_handler_real_fs():
-    """Test DocumentationHandler with real filesystem."""
-    handler = DocumentationHandler()
-
-    # Get stats from real filesystem
-    stats = await handler.get_stats()
-
-    # Stats should have correct structure
-    assert "total" in stats
-    assert "with_embeddings" in stats
-    assert "without_embeddings" in stats
-    assert stats["total"] >= 0
-    assert stats["with_embeddings"] >= 0
-    assert stats["without_embeddings"] >= 0
-
-    # Get missing items (max 1 to keep test fast)
-    items = await handler.get_missing_items(batch_size=1)
-
-    # Items should be list
-    assert isinstance(items, list)
-
-    if items:
-        item = items[0]
-        assert item.content_id is not None  # Should be relative path
-        assert item.content_type.value == "DOCUMENTATION"
-        assert item.searchable_text != ""
-        assert item.user_id is None
-
-
-@pytest.mark.asyncio(loop_scope="session")
-async def test_get_embedding_stats_all_types():
-    """Test get_embedding_stats aggregates all content types."""
-    stats = await get_embedding_stats()
-
-    # Should have structure with by_type and totals
-    assert "by_type" in stats
-    assert "totals" in stats
-
-    # Check each content type is present
-    by_type = stats["by_type"]
-    assert "STORE_AGENT" in by_type
-    assert "BLOCK" in by_type
-    assert "DOCUMENTATION" in by_type
-
-    # Check totals are aggregated
-    totals = stats["totals"]
-    assert totals["total"] >= 0
-    assert totals["with_embeddings"] >= 0
-    assert totals["without_embeddings"] >= 0
-    assert "coverage_percent" in totals
-
-
-@pytest.mark.asyncio(loop_scope="session")
-@patch("backend.api.features.store.embeddings.generate_embedding")
-async def test_ensure_content_embedding_blocks(mock_generate):
-    """Test creating embeddings for blocks (mocked OpenAI)."""
-    # Mock OpenAI to return fake embedding
-    mock_generate.return_value = [0.1] * EMBEDDING_DIM
-
-    # Get one block without embedding
-    handler = BlockHandler()
-    items = await handler.get_missing_items(batch_size=1)
-
-    if not items:
-        pytest.skip("No blocks without embeddings")
-
-    item = items[0]
-
-    # Try to create embedding (OpenAI mocked)
-    result = await ensure_content_embedding(
-        content_type=item.content_type,
-        content_id=item.content_id,
-        searchable_text=item.searchable_text,
-        metadata=item.metadata,
-        user_id=item.user_id,
-    )
-
-    # Should succeed with mocked OpenAI
-    assert result is True
-    mock_generate.assert_called_once()
-
-
-@pytest.mark.asyncio(loop_scope="session")
-@patch("backend.api.features.store.embeddings.generate_embedding")
-async def test_backfill_all_content_types_dry_run(mock_generate):
-    """Test backfill_all_content_types processes all handlers in order."""
-    # Mock OpenAI to return fake embedding
-    mock_generate.return_value = [0.1] * EMBEDDING_DIM
-
-    # Run backfill with batch_size=1 to process max 1 per type
-    result = await backfill_all_content_types(batch_size=1)
-
-    # Should have results for all content types
-    assert "by_type" in result
-    assert "totals" in result
-
-    by_type = result["by_type"]
-    assert "BLOCK" in by_type
-    assert "STORE_AGENT" in by_type
-    assert "DOCUMENTATION" in by_type
-
-    # Each type should have correct structure
-    for content_type, type_result in by_type.items():
-        assert "processed" in type_result
-        assert "success" in type_result
-        assert "failed" in type_result
-
-    # Totals should aggregate
-    totals = result["totals"]
-    assert totals["processed"] >= 0
-    assert totals["success"] >= 0
-    assert totals["failed"] >= 0
-
-
-@pytest.mark.asyncio(loop_scope="session")
-async def test_content_handler_registry():
-    """Test all handlers are registered in correct order."""
-    from prisma.enums import ContentType
-
-    # All three types should be registered
-    assert ContentType.STORE_AGENT in CONTENT_HANDLERS
-    assert ContentType.BLOCK in CONTENT_HANDLERS
-    assert ContentType.DOCUMENTATION in CONTENT_HANDLERS
-
-    # Check handler types
-    assert isinstance(CONTENT_HANDLERS[ContentType.STORE_AGENT], StoreAgentHandler)
-    assert isinstance(CONTENT_HANDLERS[ContentType.BLOCK], BlockHandler)
-    assert isinstance(CONTENT_HANDLERS[ContentType.DOCUMENTATION], DocumentationHandler)

autogpt_platform/backend/backend/api/features/store/content_handlers_test.py

@@ -1,381 +0,0 @@
-"""
-E2E tests for content handlers (blocks, store agents, documentation).
-
-Tests the full flow: discovering content → generating embeddings → storing.
-"""
-
-from pathlib import Path
-from unittest.mock import MagicMock, patch
-
-import pytest
-from prisma.enums import ContentType
-
-from backend.api.features.store.content_handlers import (
-    CONTENT_HANDLERS,
-    BlockHandler,
-    DocumentationHandler,
-    StoreAgentHandler,
-)
-
-
-@pytest.mark.asyncio(loop_scope="session")
-async def test_store_agent_handler_get_missing_items(mocker):
-    """Test StoreAgentHandler fetches approved agents without embeddings."""
-    handler = StoreAgentHandler()
-
-    # Mock database query
-    mock_missing = [
-        {
-            "id": "agent-1",
-            "name": "Test Agent",
-            "description": "A test agent",
-            "subHeading": "Test heading",
-            "categories": ["AI", "Testing"],
-        }
-    ]
-
-    with patch(
-        "backend.api.features.store.content_handlers.query_raw_with_schema",
-        return_value=mock_missing,
-    ):
-        items = await handler.get_missing_items(batch_size=10)
-
-        assert len(items) == 1
-        assert items[0].content_id == "agent-1"
-        assert items[0].content_type == ContentType.STORE_AGENT
-        assert "Test Agent" in items[0].searchable_text
-        assert "A test agent" in items[0].searchable_text
-        assert items[0].metadata["name"] == "Test Agent"
-        assert items[0].user_id is None
-
-
-@pytest.mark.asyncio(loop_scope="session")
-async def test_store_agent_handler_get_stats(mocker):
-    """Test StoreAgentHandler returns correct stats."""
-    handler = StoreAgentHandler()
-
-    # Mock approved count query
-    mock_approved = [{"count": 50}]
-    # Mock embedded count query
-    mock_embedded = [{"count": 30}]
-
-    with patch(
-        "backend.api.features.store.content_handlers.query_raw_with_schema",
-        side_effect=[mock_approved, mock_embedded],
-    ):
-        stats = await handler.get_stats()
-
-        assert stats["total"] == 50
-        assert stats["with_embeddings"] == 30
-        assert stats["without_embeddings"] == 20
-
-
-@pytest.mark.asyncio(loop_scope="session")
-async def test_block_handler_get_missing_items(mocker):
-    """Test BlockHandler discovers blocks without embeddings."""
-    handler = BlockHandler()
-
-    # Mock get_blocks to return test blocks
-    mock_block_class = MagicMock()
-    mock_block_instance = MagicMock()
-    mock_block_instance.name = "Calculator Block"
-    mock_block_instance.description = "Performs calculations"
-    mock_block_instance.categories = [MagicMock(value="MATH")]
-    mock_block_instance.input_schema.model_json_schema.return_value = {
-        "properties": {"expression": {"description": "Math expression to evaluate"}}
-    }
-    mock_block_class.return_value = mock_block_instance
-
-    mock_blocks = {"block-uuid-1": mock_block_class}
-
-    # Mock existing embeddings query (no embeddings exist)
-    mock_existing = []
-
-    with patch(
-        "backend.data.block.get_blocks",
-        return_value=mock_blocks,
-    ):
-        with patch(
-            "backend.api.features.store.content_handlers.query_raw_with_schema",
-            return_value=mock_existing,
-        ):
-            items = await handler.get_missing_items(batch_size=10)
-
-            assert len(items) == 1
-            assert items[0].content_id == "block-uuid-1"
-            assert items[0].content_type == ContentType.BLOCK
-            assert "Calculator Block" in items[0].searchable_text
-            assert "Performs calculations" in items[0].searchable_text
-            assert "MATH" in items[0].searchable_text
-            assert "expression: Math expression" in items[0].searchable_text
-            assert items[0].user_id is None
-
-
-@pytest.mark.asyncio(loop_scope="session")
-async def test_block_handler_get_stats(mocker):
-    """Test BlockHandler returns correct stats."""
-    handler = BlockHandler()
-
-    # Mock get_blocks
-    mock_blocks = {
-        "block-1": MagicMock(),
-        "block-2": MagicMock(),
-        "block-3": MagicMock(),
-    }
-
-    # Mock embedded count query (2 blocks have embeddings)
-    mock_embedded = [{"count": 2}]
-
-    with patch(
-        "backend.data.block.get_blocks",
-        return_value=mock_blocks,
-    ):
-        with patch(
-            "backend.api.features.store.content_handlers.query_raw_with_schema",
-            return_value=mock_embedded,
-        ):
-            stats = await handler.get_stats()
-
-            assert stats["total"] == 3
-            assert stats["with_embeddings"] == 2
-            assert stats["without_embeddings"] == 1
-
-
-@pytest.mark.asyncio(loop_scope="session")
-async def test_documentation_handler_get_missing_items(tmp_path, mocker):
-    """Test DocumentationHandler discovers docs without embeddings."""
-    handler = DocumentationHandler()
-
-    # Create temporary docs directory with test files
-    docs_root = tmp_path / "docs"
-    docs_root.mkdir()
-
-    (docs_root / "guide.md").write_text("# Getting Started\n\nThis is a guide.")
-    (docs_root / "api.mdx").write_text("# API Reference\n\nAPI documentation.")
-
-    # Mock _get_docs_root to return temp dir
-    with patch.object(handler, "_get_docs_root", return_value=docs_root):
-        # Mock existing embeddings query (no embeddings exist)
-        with patch(
-            "backend.api.features.store.content_handlers.query_raw_with_schema",
-            return_value=[],
-        ):
-            items = await handler.get_missing_items(batch_size=10)
-
-            assert len(items) == 2
-
-            # Check guide.md (content_id format: doc_path::section_index)
-            guide_item = next(
-                (item for item in items if item.content_id == "guide.md::0"), None
-            )
-            assert guide_item is not None
-            assert guide_item.content_type == ContentType.DOCUMENTATION
-            assert "Getting Started" in guide_item.searchable_text
-            assert "This is a guide" in guide_item.searchable_text
-            assert guide_item.metadata["doc_title"] == "Getting Started"
-            assert guide_item.user_id is None
-
-            # Check api.mdx (content_id format: doc_path::section_index)
-            api_item = next(
-                (item for item in items if item.content_id == "api.mdx::0"), None
-            )
-            assert api_item is not None
-            assert "API Reference" in api_item.searchable_text
-
-
-@pytest.mark.asyncio(loop_scope="session")
-async def test_documentation_handler_get_stats(tmp_path, mocker):
-    """Test DocumentationHandler returns correct stats."""
-    handler = DocumentationHandler()
-
-    # Create temporary docs directory
-    docs_root = tmp_path / "docs"
-    docs_root.mkdir()
-    (docs_root / "doc1.md").write_text("# Doc 1")
-    (docs_root / "doc2.md").write_text("# Doc 2")
-    (docs_root / "doc3.mdx").write_text("# Doc 3")
-
-    # Mock embedded count query (1 doc has embedding)
-    mock_embedded = [{"count": 1}]
-
-    with patch.object(handler, "_get_docs_root", return_value=docs_root):
-        with patch(
-            "backend.api.features.store.content_handlers.query_raw_with_schema",
-            return_value=mock_embedded,
-        ):
-            stats = await handler.get_stats()
-
-            assert stats["total"] == 3
-            assert stats["with_embeddings"] == 1
-            assert stats["without_embeddings"] == 2
-
-
-@pytest.mark.asyncio(loop_scope="session")
-async def test_documentation_handler_title_extraction(tmp_path):
-    """Test DocumentationHandler extracts title from markdown heading."""
-    handler = DocumentationHandler()
-
-    # Test with heading
-    doc_with_heading = tmp_path / "with_heading.md"
-    doc_with_heading.write_text("# My Title\n\nContent here")
-    title = handler._extract_doc_title(doc_with_heading)
-    assert title == "My Title"
-
-    # Test without heading
-    doc_without_heading = tmp_path / "no-heading.md"
-    doc_without_heading.write_text("Just content, no heading")
-    title = handler._extract_doc_title(doc_without_heading)
-    assert title == "No Heading"  # Uses filename
-
-
-@pytest.mark.asyncio(loop_scope="session")
-async def test_documentation_handler_markdown_chunking(tmp_path):
-    """Test DocumentationHandler chunks markdown by headings."""
-    handler = DocumentationHandler()
-
-    # Test document with multiple sections
-    doc_with_sections = tmp_path / "sections.md"
-    doc_with_sections.write_text(
-        "# Document Title\n\n"
-        "Intro paragraph.\n\n"
-        "## Section One\n\n"
-        "Content for section one.\n\n"
-        "## Section Two\n\n"
-        "Content for section two.\n"
-    )
-    sections = handler._chunk_markdown_by_headings(doc_with_sections)
-
-    # Should have 3 sections: intro (with doc title), section one, section two
-    assert len(sections) == 3
-    assert sections[0].title == "Document Title"
-    assert sections[0].index == 0
-    assert "Intro paragraph" in sections[0].content
-
-    assert sections[1].title == "Section One"
-    assert sections[1].index == 1
-    assert "Content for section one" in sections[1].content
-
-    assert sections[2].title == "Section Two"
-    assert sections[2].index == 2
-    assert "Content for section two" in sections[2].content
-
-    # Test document without headings
-    doc_no_sections = tmp_path / "no-sections.md"
-    doc_no_sections.write_text("Just plain content without any headings.")
-    sections = handler._chunk_markdown_by_headings(doc_no_sections)
-    assert len(sections) == 1
-    assert sections[0].index == 0
-    assert "Just plain content" in sections[0].content
-
-
-@pytest.mark.asyncio(loop_scope="session")
-async def test_documentation_handler_section_content_ids():
-    """Test DocumentationHandler creates and parses section content IDs."""
-    handler = DocumentationHandler()
-
-    # Test making content ID
-    content_id = handler._make_section_content_id("docs/guide.md", 2)
-    assert content_id == "docs/guide.md::2"
-
-    # Test parsing content ID
-    doc_path, section_index = handler._parse_section_content_id("docs/guide.md::2")
-    assert doc_path == "docs/guide.md"
-    assert section_index == 2
-
-    # Test parsing legacy format (no section index)
-    doc_path, section_index = handler._parse_section_content_id("docs/old-format.md")
-    assert doc_path == "docs/old-format.md"
-    assert section_index == 0
-
-
-@pytest.mark.asyncio(loop_scope="session")
-async def test_content_handlers_registry():
-    """Test all content types are registered."""
-    assert ContentType.STORE_AGENT in CONTENT_HANDLERS
-    assert ContentType.BLOCK in CONTENT_HANDLERS
-    assert ContentType.DOCUMENTATION in CONTENT_HANDLERS
-
-    assert isinstance(CONTENT_HANDLERS[ContentType.STORE_AGENT], StoreAgentHandler)
-    assert isinstance(CONTENT_HANDLERS[ContentType.BLOCK], BlockHandler)
-    assert isinstance(CONTENT_HANDLERS[ContentType.DOCUMENTATION], DocumentationHandler)
-
-
-@pytest.mark.asyncio(loop_scope="session")
-async def test_block_handler_handles_missing_attributes():
-    """Test BlockHandler gracefully handles blocks with missing attributes."""
-    handler = BlockHandler()
-
-    # Mock block with minimal attributes
-    mock_block_class = MagicMock()
-    mock_block_instance = MagicMock()
-    mock_block_instance.name = "Minimal Block"
-    # No description, categories, or schema
-    del mock_block_instance.description
-    del mock_block_instance.categories
-    del mock_block_instance.input_schema
-    mock_block_class.return_value = mock_block_instance
-
-    mock_blocks = {"block-minimal": mock_block_class}
-
-    with patch(
-        "backend.data.block.get_blocks",
-        return_value=mock_blocks,
-    ):
-        with patch(
-            "backend.api.features.store.content_handlers.query_raw_with_schema",
-            return_value=[],
-        ):
-            items = await handler.get_missing_items(batch_size=10)
-
-            assert len(items) == 1
-            assert items[0].searchable_text == "Minimal Block"
-
-
-@pytest.mark.asyncio(loop_scope="session")
-async def test_block_handler_skips_failed_blocks():
-    """Test BlockHandler skips blocks that fail to instantiate."""
-    handler = BlockHandler()
-
-    # Mock one good block and one bad block
-    good_block = MagicMock()
-    good_instance = MagicMock()
-    good_instance.name = "Good Block"
-    good_instance.description = "Works fine"
-    good_instance.categories = []
-    good_block.return_value = good_instance
-
-    bad_block = MagicMock()
-    bad_block.side_effect = Exception("Instantiation failed")
-
-    mock_blocks = {"good-block": good_block, "bad-block": bad_block}
-
-    with patch(
-        "backend.data.block.get_blocks",
-        return_value=mock_blocks,
-    ):
-        with patch(
-            "backend.api.features.store.content_handlers.query_raw_with_schema",
-            return_value=[],
-        ):
-            items = await handler.get_missing_items(batch_size=10)
-
-            # Should only get the good block
-            assert len(items) == 1
-            assert items[0].content_id == "good-block"
-
-
-@pytest.mark.asyncio(loop_scope="session")
-async def test_documentation_handler_missing_docs_directory():
-    """Test DocumentationHandler handles missing docs directory gracefully."""
-    handler = DocumentationHandler()
-
-    # Mock _get_docs_root to return non-existent path
-    fake_path = Path("/nonexistent/docs")
-    with patch.object(handler, "_get_docs_root", return_value=fake_path):
-        items = await handler.get_missing_items(batch_size=10)
-        assert items == []
-
-        stats = await handler.get_stats()
-        assert stats["total"] == 0
-        assert stats["with_embeddings"] == 0
-        assert stats["without_embeddings"] == 0

autogpt_platform/backend/backend/api/features/store/embeddings.py

@@ -1,962 +0,0 @@
-"""
-Unified Content Embeddings Service
-
-Handles generation and storage of OpenAI embeddings for all content types
-(store listings, blocks, documentation, library agents) to enable semantic/hybrid search.
-"""
-
-import asyncio
-import logging
-import time
-from typing import Any
-
-import prisma
-from prisma.enums import ContentType
-from tiktoken import encoding_for_model
-
-from backend.api.features.store.content_handlers import CONTENT_HANDLERS
-from backend.data.db import execute_raw_with_schema, query_raw_with_schema
-from backend.util.clients import get_openai_client
-from backend.util.json import dumps
-
-logger = logging.getLogger(__name__)
-
-
-# OpenAI embedding model configuration
-EMBEDDING_MODEL = "text-embedding-3-small"
-# Embedding dimension for the model above
-# text-embedding-3-small: 1536, text-embedding-3-large: 3072
-EMBEDDING_DIM = 1536
-# OpenAI embedding token limit (8,191 with 1 token buffer for safety)
-EMBEDDING_MAX_TOKENS = 8191
-
-
-def build_searchable_text(
-    name: str,
-    description: str,
-    sub_heading: str,
-    categories: list[str],
-) -> str:
-    """
-    Build searchable text from listing version fields.
-
-    Combines relevant fields into a single string for embedding.
-    """
-    parts = []
-
-    # Name is important - include it
-    if name:
-        parts.append(name)
-
-    # Sub-heading provides context
-    if sub_heading:
-        parts.append(sub_heading)
-
-    # Description is the main content
-    if description:
-        parts.append(description)
-
-    # Categories help with semantic matching
-    if categories:
-        parts.append(" ".join(categories))
-
-    return " ".join(parts)
-
-
-async def generate_embedding(text: str) -> list[float] | None:
-    """
-    Generate embedding for text using OpenAI API.
-
-    Returns None if embedding generation fails.
-    Fail-fast: no retries to maintain consistency with approval flow.
-    """
-    try:
-        client = get_openai_client()
-        if not client:
-            logger.error("openai_internal_api_key not set, cannot generate embedding")
-            return None
-
-        # Truncate text to token limit using tiktoken
-        # Character-based truncation is insufficient because token ratios vary by content type
-        enc = encoding_for_model(EMBEDDING_MODEL)
-        tokens = enc.encode(text)
-        if len(tokens) > EMBEDDING_MAX_TOKENS:
-            tokens = tokens[:EMBEDDING_MAX_TOKENS]
-            truncated_text = enc.decode(tokens)
-            logger.info(
-                f"Truncated text from {len(enc.encode(text))} to {len(tokens)} tokens"
-            )
-        else:
-            truncated_text = text
-
-        start_time = time.time()
-        response = await client.embeddings.create(
-            model=EMBEDDING_MODEL,
-            input=truncated_text,
-        )
-        latency_ms = (time.time() - start_time) * 1000
-
-        embedding = response.data[0].embedding
-        logger.info(
-            f"Generated embedding: {len(embedding)} dims, "
-            f"{len(tokens)} tokens, {latency_ms:.0f}ms"
-        )
-        return embedding
-
-    except Exception as e:
-        logger.error(f"Failed to generate embedding: {e}")
-        return None
-
-
-async def store_embedding(
-    version_id: str,
-    embedding: list[float],
-    tx: prisma.Prisma | None = None,
-) -> bool:
-    """
-    Store embedding in the database.
-
-    BACKWARD COMPATIBILITY: Maintained for existing store listing usage.
-    DEPRECATED: Use ensure_embedding() instead (includes searchable_text).
-    """
-    return await store_content_embedding(
-        content_type=ContentType.STORE_AGENT,
-        content_id=version_id,
-        embedding=embedding,
-        searchable_text="",  # Empty for backward compat; ensure_embedding() populates this
-        metadata=None,
-        user_id=None,  # Store agents are public
-        tx=tx,
-    )
-
-
-async def store_content_embedding(
-    content_type: ContentType,
-    content_id: str,
-    embedding: list[float],
-    searchable_text: str,
-    metadata: dict | None = None,
-    user_id: str | None = None,
-    tx: prisma.Prisma | None = None,
-) -> bool:
-    """
-    Store embedding in the unified content embeddings table.
-
-    New function for unified content embedding storage.
-    Uses raw SQL since Prisma doesn't natively support pgvector.
-    """
-    try:
-        client = tx if tx else prisma.get_client()
-
-        # Convert embedding to PostgreSQL vector format
-        embedding_str = embedding_to_vector_string(embedding)
-        metadata_json = dumps(metadata or {})
-
-        # Upsert the embedding
-        # WHERE clause in DO UPDATE prevents PostgreSQL 15 bug with NULLS NOT DISTINCT
-        await execute_raw_with_schema(
-            """
-            INSERT INTO {schema_prefix}"UnifiedContentEmbedding" (
-                "id", "contentType", "contentId", "userId", "embedding", "searchableText", "metadata", "createdAt", "updatedAt"
-            )
-            VALUES (gen_random_uuid()::text, $1::{schema_prefix}"ContentType", $2, $3, $4::vector, $5, $6::jsonb, NOW(), NOW())
-            ON CONFLICT ("contentType", "contentId", "userId")
-            DO UPDATE SET
-                "embedding" = $4::vector,
-                "searchableText" = $5,
-                "metadata" = $6::jsonb,
-                "updatedAt" = NOW()
-            WHERE {schema_prefix}"UnifiedContentEmbedding"."contentType" = $1::{schema_prefix}"ContentType"
-                AND {schema_prefix}"UnifiedContentEmbedding"."contentId" = $2
-                AND ({schema_prefix}"UnifiedContentEmbedding"."userId" = $3 OR ($3 IS NULL AND {schema_prefix}"UnifiedContentEmbedding"."userId" IS NULL))
-            """,
-            content_type,
-            content_id,
-            user_id,
-            embedding_str,
-            searchable_text,
-            metadata_json,
-            client=client,
-            set_public_search_path=True,
-        )
-
-        logger.info(f"Stored embedding for {content_type}:{content_id}")
-        return True
-
-    except Exception as e:
-        logger.error(f"Failed to store embedding for {content_type}:{content_id}: {e}")
-        return False
-
-
-async def get_embedding(version_id: str) -> dict[str, Any] | None:
-    """
-    Retrieve embedding record for a listing version.
-
-    BACKWARD COMPATIBILITY: Maintained for existing store listing usage.
-    Returns dict with storeListingVersionId, embedding, timestamps or None if not found.
-    """
-    result = await get_content_embedding(
-        ContentType.STORE_AGENT, version_id, user_id=None
-    )
-    if result:
-        # Transform to old format for backward compatibility
-        return {
-            "storeListingVersionId": result["contentId"],
-            "embedding": result["embedding"],
-            "createdAt": result["createdAt"],
-            "updatedAt": result["updatedAt"],
-        }
-    return None
-
-
-async def get_content_embedding(
-    content_type: ContentType, content_id: str, user_id: str | None = None
-) -> dict[str, Any] | None:
-    """
-    Retrieve embedding record for any content type.
-
-    New function for unified content embedding retrieval.
-    Returns dict with contentType, contentId, embedding, timestamps or None if not found.
-    """
-    try:
-        result = await query_raw_with_schema(
-            """
-            SELECT
-                "contentType",
-                "contentId",
-                "userId",
-                "embedding"::text as "embedding",
-                "searchableText",
-                "metadata",
-                "createdAt",
-                "updatedAt"
-            FROM {schema_prefix}"UnifiedContentEmbedding"
-            WHERE "contentType" = $1::{schema_prefix}"ContentType" AND "contentId" = $2 AND ("userId" = $3 OR ($3 IS NULL AND "userId" IS NULL))
-            """,
-            content_type,
-            content_id,
-            user_id,
-            set_public_search_path=True,
-        )
-
-        if result and len(result) > 0:
-            return result[0]
-        return None
-
-    except Exception as e:
-        logger.error(f"Failed to get embedding for {content_type}:{content_id}: {e}")
-        return None
-
-
-async def ensure_embedding(
-    version_id: str,
-    name: str,
-    description: str,
-    sub_heading: str,
-    categories: list[str],
-    force: bool = False,
-    tx: prisma.Prisma | None = None,
-) -> bool:
-    """
-    Ensure an embedding exists for the listing version.
-
-    Creates embedding if missing. Use force=True to regenerate.
-    Backward-compatible wrapper for store listings.
-
-    Args:
-        version_id: The StoreListingVersion ID
-        name: Agent name
-        description: Agent description
-        sub_heading: Agent sub-heading
-        categories: Agent categories
-        force: Force regeneration even if embedding exists
-        tx: Optional transaction client
-
-    Returns:
-        True if embedding exists/was created, False on failure
-    """
-    try:
-        # Check if embedding already exists
-        if not force:
-            existing = await get_embedding(version_id)
-            if existing and existing.get("embedding"):
-                logger.debug(f"Embedding for version {version_id} already exists")
-                return True
-
-        # Build searchable text for embedding
-        searchable_text = build_searchable_text(
-            name, description, sub_heading, categories
-        )
-
-        # Generate new embedding
-        embedding = await generate_embedding(searchable_text)
-        if embedding is None:
-            logger.warning(f"Could not generate embedding for version {version_id}")
-            return False
-
-        # Store the embedding with metadata using new function
-        metadata = {
-            "name": name,
-            "subHeading": sub_heading,
-            "categories": categories,
-        }
-        return await store_content_embedding(
-            content_type=ContentType.STORE_AGENT,
-            content_id=version_id,
-            embedding=embedding,
-            searchable_text=searchable_text,
-            metadata=metadata,
-            user_id=None,  # Store agents are public
-            tx=tx,
-        )
-
-    except Exception as e:
-        logger.error(f"Failed to ensure embedding for version {version_id}: {e}")
-        return False
-
-
-async def delete_embedding(version_id: str) -> bool:
-    """
-    Delete embedding for a listing version.
-
-    BACKWARD COMPATIBILITY: Maintained for existing store listing usage.
-    Note: This is usually handled automatically by CASCADE delete,
-    but provided for manual cleanup if needed.
-    """
-    return await delete_content_embedding(ContentType.STORE_AGENT, version_id)
-
-
-async def delete_content_embedding(
-    content_type: ContentType, content_id: str, user_id: str | None = None
-) -> bool:
-    """
-    Delete embedding for any content type.
-
-    New function for unified content embedding deletion.
-    Note: This is usually handled automatically by CASCADE delete,
-    but provided for manual cleanup if needed.
-
-    Args:
-        content_type: The type of content (STORE_AGENT, LIBRARY_AGENT, etc.)
-        content_id: The unique identifier for the content
-        user_id: Optional user ID. For public content (STORE_AGENT, BLOCK), pass None.
-                 For user-scoped content (LIBRARY_AGENT), pass the user's ID to avoid
-                 deleting embeddings belonging to other users.
-
-    Returns:
-        True if deletion succeeded, False otherwise
-    """
-    try:
-        client = prisma.get_client()
-
-        await execute_raw_with_schema(
-            """
-            DELETE FROM {schema_prefix}"UnifiedContentEmbedding"
-            WHERE "contentType" = $1::{schema_prefix}"ContentType"
-              AND "contentId" = $2
-              AND ("userId" = $3 OR ($3 IS NULL AND "userId" IS NULL))
-            """,
-            content_type,
-            content_id,
-            user_id,
-            client=client,
-        )
-
-        user_str = f" (user: {user_id})" if user_id else ""
-        logger.info(f"Deleted embedding for {content_type}:{content_id}{user_str}")
-        return True
-
-    except Exception as e:
-        logger.error(f"Failed to delete embedding for {content_type}:{content_id}: {e}")
-        return False
-
-
-async def get_embedding_stats() -> dict[str, Any]:
-    """
-    Get statistics about embedding coverage for all content types.
-
-    Returns stats per content type and overall totals.
-    """
-    try:
-        stats_by_type = {}
-        total_items = 0
-        total_with_embeddings = 0
-        total_without_embeddings = 0
-
-        # Aggregate stats from all handlers
-        for content_type, handler in CONTENT_HANDLERS.items():
-            try:
-                stats = await handler.get_stats()
-                stats_by_type[content_type.value] = {
-                    "total": stats["total"],
-                    "with_embeddings": stats["with_embeddings"],
-                    "without_embeddings": stats["without_embeddings"],
-                    "coverage_percent": (
-                        round(stats["with_embeddings"] / stats["total"] * 100, 1)
-                        if stats["total"] > 0
-                        else 0
-                    ),
-                }
-
-                total_items += stats["total"]
-                total_with_embeddings += stats["with_embeddings"]
-                total_without_embeddings += stats["without_embeddings"]
-
-            except Exception as e:
-                logger.error(f"Failed to get stats for {content_type.value}: {e}")
-                stats_by_type[content_type.value] = {
-                    "total": 0,
-                    "with_embeddings": 0,
-                    "without_embeddings": 0,
-                    "coverage_percent": 0,
-                    "error": str(e),
-                }
-
-        return {
-            "by_type": stats_by_type,
-            "totals": {
-                "total": total_items,
-                "with_embeddings": total_with_embeddings,
-                "without_embeddings": total_without_embeddings,
-                "coverage_percent": (
-                    round(total_with_embeddings / total_items * 100, 1)
-                    if total_items > 0
-                    else 0
-                ),
-            },
-        }
-
-    except Exception as e:
-        logger.error(f"Failed to get embedding stats: {e}")
-        return {
-            "by_type": {},
-            "totals": {
-                "total": 0,
-                "with_embeddings": 0,
-                "without_embeddings": 0,
-                "coverage_percent": 0,
-            },
-            "error": str(e),
-        }
-
-
-async def backfill_missing_embeddings(batch_size: int = 10) -> dict[str, Any]:
-    """
-    Generate embeddings for approved listings that don't have them.
-
-    BACKWARD COMPATIBILITY: Maintained for existing usage.
-    This now delegates to backfill_all_content_types() to process all content types.
-
-    Args:
-        batch_size: Number of embeddings to generate per content type
-
-    Returns:
-        Dict with success/failure counts aggregated across all content types
-    """
-    # Delegate to the new generic backfill system
-    result = await backfill_all_content_types(batch_size)
-
-    # Return in the old format for backward compatibility
-    return result["totals"]
-
-
-async def backfill_all_content_types(batch_size: int = 10) -> dict[str, Any]:
-    """
-    Generate embeddings for all content types using registered handlers.
-
-    Processes content types in order: BLOCK → STORE_AGENT → DOCUMENTATION.
-    This ensures foundational content (blocks) are searchable first.
-
-    Args:
-        batch_size: Number of embeddings to generate per content type
-
-    Returns:
-        Dict with stats per content type and overall totals
-    """
-    results_by_type = {}
-    total_processed = 0
-    total_success = 0
-    total_failed = 0
-
-    # Process content types in explicit order
-    processing_order = [
-        ContentType.BLOCK,
-        ContentType.STORE_AGENT,
-        ContentType.DOCUMENTATION,
-    ]
-
-    for content_type in processing_order:
-        handler = CONTENT_HANDLERS.get(content_type)
-        if not handler:
-            logger.warning(f"No handler registered for {content_type.value}")
-            continue
-        try:
-            logger.info(f"Processing {content_type.value} content type...")
-
-            # Get missing items from handler
-            missing_items = await handler.get_missing_items(batch_size)
-
-            if not missing_items:
-                results_by_type[content_type.value] = {
-                    "processed": 0,
-                    "success": 0,
-                    "failed": 0,
-                    "message": "No missing embeddings",
-                }
-                continue
-
-            # Process embeddings concurrently for better performance
-            embedding_tasks = [
-                ensure_content_embedding(
-                    content_type=item.content_type,
-                    content_id=item.content_id,
-                    searchable_text=item.searchable_text,
-                    metadata=item.metadata,
-                    user_id=item.user_id,
-                )
-                for item in missing_items
-            ]
-
-            results = await asyncio.gather(*embedding_tasks, return_exceptions=True)
-
-            success = sum(1 for result in results if result is True)
-            failed = len(results) - success
-
-            results_by_type[content_type.value] = {
-                "processed": len(missing_items),
-                "success": success,
-                "failed": failed,
-                "message": f"Backfilled {success} embeddings, {failed} failed",
-            }
-
-            total_processed += len(missing_items)
-            total_success += success
-            total_failed += failed
-
-            logger.info(
-                f"{content_type.value}: processed {len(missing_items)}, "
-                f"success {success}, failed {failed}"
-            )
-
-        except Exception as e:
-            logger.error(f"Failed to process {content_type.value}: {e}")
-            results_by_type[content_type.value] = {
-                "processed": 0,
-                "success": 0,
-                "failed": 0,
-                "error": str(e),
-            }
-
-    return {
-        "by_type": results_by_type,
-        "totals": {
-            "processed": total_processed,
-            "success": total_success,
-            "failed": total_failed,
-            "message": f"Overall: {total_success} succeeded, {total_failed} failed",
-        },
-    }
-
-
-async def embed_query(query: str) -> list[float] | None:
-    """
-    Generate embedding for a search query.
-
-    Same as generate_embedding but with clearer intent.
-    """
-    return await generate_embedding(query)
-
-
-def embedding_to_vector_string(embedding: list[float]) -> str:
-    """Convert embedding list to PostgreSQL vector string format."""
-    return "[" + ",".join(str(x) for x in embedding) + "]"
-
-
-async def ensure_content_embedding(
-    content_type: ContentType,
-    content_id: str,
-    searchable_text: str,
-    metadata: dict | None = None,
-    user_id: str | None = None,
-    force: bool = False,
-    tx: prisma.Prisma | None = None,
-) -> bool:
-    """
-    Ensure an embedding exists for any content type.
-
-    Generic function for creating embeddings for store agents, blocks, docs, etc.
-
-    Args:
-        content_type: ContentType enum value (STORE_AGENT, BLOCK, etc.)
-        content_id: Unique identifier for the content
-        searchable_text: Combined text for embedding generation
-        metadata: Optional metadata to store with embedding
-        force: Force regeneration even if embedding exists
-        tx: Optional transaction client
-
-    Returns:
-        True if embedding exists/was created, False on failure
-    """
-    try:
-        # Check if embedding already exists
-        if not force:
-            existing = await get_content_embedding(content_type, content_id, user_id)
-            if existing and existing.get("embedding"):
-                logger.debug(
-                    f"Embedding for {content_type}:{content_id} already exists"
-                )
-                return True
-
-        # Generate new embedding
-        embedding = await generate_embedding(searchable_text)
-        if embedding is None:
-            logger.warning(
-                f"Could not generate embedding for {content_type}:{content_id}"
-            )
-            return False
-
-        # Store the embedding
-        return await store_content_embedding(
-            content_type=content_type,
-            content_id=content_id,
-            embedding=embedding,
-            searchable_text=searchable_text,
-            metadata=metadata or {},
-            user_id=user_id,
-            tx=tx,
-        )
-
-    except Exception as e:
-        logger.error(f"Failed to ensure embedding for {content_type}:{content_id}: {e}")
-        return False
-
-
-async def cleanup_orphaned_embeddings() -> dict[str, Any]:
-    """
-    Clean up embeddings for content that no longer exists or is no longer valid.
-
-    Compares current content with embeddings in database and removes orphaned records:
-    - STORE_AGENT: Removes embeddings for rejected/deleted store listings
-    - BLOCK: Removes embeddings for blocks no longer registered
-    - DOCUMENTATION: Removes embeddings for deleted doc files
-
-    Returns:
-        Dict with cleanup statistics per content type
-    """
-    results_by_type = {}
-    total_deleted = 0
-
-    # Cleanup orphaned embeddings for all content types
-    cleanup_types = [
-        ContentType.STORE_AGENT,
-        ContentType.BLOCK,
-        ContentType.DOCUMENTATION,
-    ]
-
-    for content_type in cleanup_types:
-        try:
-            handler = CONTENT_HANDLERS.get(content_type)
-            if not handler:
-                logger.warning(f"No handler registered for {content_type}")
-                results_by_type[content_type.value] = {
-                    "deleted": 0,
-                    "error": "No handler registered",
-                }
-                continue
-
-            # Get all current content IDs from handler
-            if content_type == ContentType.STORE_AGENT:
-                # Get IDs of approved store listing versions from non-deleted listings
-                valid_agents = await query_raw_with_schema(
-                    """
-                    SELECT slv.id
-                    FROM {schema_prefix}"StoreListingVersion" slv
-                    JOIN {schema_prefix}"StoreListing" sl ON slv."storeListingId" = sl.id
-                    WHERE slv."submissionStatus" = 'APPROVED'
-                      AND slv."isDeleted" = false
-                      AND sl."isDeleted" = false
-                    """,
-                )
-                current_ids = {row["id"] for row in valid_agents}
-            elif content_type == ContentType.BLOCK:
-                from backend.data.block import get_blocks
-
-                current_ids = set(get_blocks().keys())
-            elif content_type == ContentType.DOCUMENTATION:
-                # Use DocumentationHandler to get section-based content IDs
-                from backend.api.features.store.content_handlers import (
-                    DocumentationHandler,
-                )
-
-                doc_handler = CONTENT_HANDLERS.get(ContentType.DOCUMENTATION)
-                if isinstance(doc_handler, DocumentationHandler):
-                    docs_root = doc_handler._get_docs_root()
-                    if docs_root.exists():
-                        current_ids = doc_handler._get_all_section_content_ids(
-                            docs_root
-                        )
-                    else:
-                        current_ids = set()
-                else:
-                    current_ids = set()
-            else:
-                # Skip unknown content types to avoid accidental deletion
-                logger.warning(
-                    f"Skipping cleanup for unknown content type: {content_type}"
-                )
-                results_by_type[content_type.value] = {
-                    "deleted": 0,
-                    "error": "Unknown content type - skipped for safety",
-                }
-                continue
-
-            # Get all embedding IDs from database
-            db_embeddings = await query_raw_with_schema(
-                """
-                SELECT "contentId"
-                FROM {schema_prefix}"UnifiedContentEmbedding"
-                WHERE "contentType" = $1::{schema_prefix}"ContentType"
-                """,
-                content_type,
-            )
-
-            db_ids = {row["contentId"] for row in db_embeddings}
-
-            # Find orphaned embeddings (in DB but not in current content)
-            orphaned_ids = db_ids - current_ids
-
-            if not orphaned_ids:
-                logger.info(f"{content_type.value}: No orphaned embeddings found")
-                results_by_type[content_type.value] = {
-                    "deleted": 0,
-                    "message": "No orphaned embeddings",
-                }
-                continue
-
-            # Delete orphaned embeddings in batch for better performance
-            orphaned_list = list(orphaned_ids)
-            try:
-                await execute_raw_with_schema(
-                    """
-                    DELETE FROM {schema_prefix}"UnifiedContentEmbedding"
-                    WHERE "contentType" = $1::{schema_prefix}"ContentType"
-                      AND "contentId" = ANY($2::text[])
-                    """,
-                    content_type,
-                    orphaned_list,
-                )
-                deleted = len(orphaned_list)
-            except Exception as e:
-                logger.error(f"Failed to batch delete orphaned embeddings: {e}")
-                deleted = 0
-
-            logger.info(
-                f"{content_type.value}: Deleted {deleted}/{len(orphaned_ids)} orphaned embeddings"
-            )
-            results_by_type[content_type.value] = {
-                "deleted": deleted,
-                "orphaned": len(orphaned_ids),
-                "message": f"Deleted {deleted} orphaned embeddings",
-            }
-
-            total_deleted += deleted
-
-        except Exception as e:
-            logger.error(f"Failed to cleanup {content_type.value}: {e}")
-            results_by_type[content_type.value] = {
-                "deleted": 0,
-                "error": str(e),
-            }
-
-    return {
-        "by_type": results_by_type,
-        "totals": {
-            "deleted": total_deleted,
-            "message": f"Deleted {total_deleted} orphaned embeddings",
-        },
-    }
-
-
-async def semantic_search(
-    query: str,
-    content_types: list[ContentType] | None = None,
-    user_id: str | None = None,
-    limit: int = 20,
-    min_similarity: float = 0.5,
-) -> list[dict[str, Any]]:
-    """
-    Semantic search across content types using embeddings.
-
-    Performs vector similarity search on UnifiedContentEmbedding table.
-    Used directly for blocks/docs/library agents, or as the semantic component
-    within hybrid_search for store agents.
-
-    If embedding generation fails, falls back to lexical search on searchableText.
-
-    Args:
-        query: Search query string
-        content_types: List of ContentType to search. Defaults to [BLOCK, STORE_AGENT, DOCUMENTATION]
-        user_id: Optional user ID for searching private content (library agents)
-        limit: Maximum number of results to return (default: 20)
-        min_similarity: Minimum cosine similarity threshold (0-1, default: 0.5)
-
-    Returns:
-        List of search results with the following structure:
-        [
-            {
-                "content_id": str,
-                "content_type": str,  # "BLOCK", "STORE_AGENT", "DOCUMENTATION", or "LIBRARY_AGENT"
-                "searchable_text": str,
-                "metadata": dict,
-                "similarity": float,  # Cosine similarity score (0-1)
-            },
-            ...
-        ]
-
-    Examples:
-        # Search blocks only
-        results = await semantic_search("calculate", content_types=[ContentType.BLOCK])
-
-        # Search blocks and documentation
-        results = await semantic_search(
-            "how to use API",
-            content_types=[ContentType.BLOCK, ContentType.DOCUMENTATION]
-        )
-
-        # Search all public content (default)
-        results = await semantic_search("AI agent")
-
-        # Search user's library agents
-        results = await semantic_search(
-            "my custom agent",
-            content_types=[ContentType.LIBRARY_AGENT],
-            user_id="user123"
-        )
-    """
-    # Default to searching all public content types
-    if content_types is None:
-        content_types = [
-            ContentType.BLOCK,
-            ContentType.STORE_AGENT,
-            ContentType.DOCUMENTATION,
-        ]
-
-    # Validate inputs
-    if not content_types:
-        return []  # Empty content_types would cause invalid SQL (IN ())
-
-    query = query.strip()
-    if not query:
-        return []
-
-    if limit < 1:
-        limit = 1
-    if limit > 100:
-        limit = 100
-
-    # Generate query embedding
-    query_embedding = await embed_query(query)
-
-    if query_embedding is not None:
-        # Semantic search with embeddings
-        embedding_str = embedding_to_vector_string(query_embedding)
-
-        # Build params in order: limit, then user_id (if provided), then content types
-        params: list[Any] = [limit]
-        user_filter = ""
-        if user_id is not None:
-            user_filter = 'AND "userId" = ${}'.format(len(params) + 1)
-            params.append(user_id)
-
-        # Add content type parameters and build placeholders dynamically
-        content_type_start_idx = len(params) + 1
-        content_type_placeholders = ", ".join(
-            f'${content_type_start_idx + i}::{{{{schema_prefix}}}}"ContentType"'
-            for i in range(len(content_types))
-        )
-        params.extend([ct.value for ct in content_types])
-
-        sql = f"""
-            SELECT
-                "contentId" as content_id,
-                "contentType" as content_type,
-                "searchableText" as searchable_text,
-                metadata,
-                1 - (embedding <=> '{embedding_str}'::vector) as similarity
-            FROM {{{{schema_prefix}}}}"UnifiedContentEmbedding"
-            WHERE "contentType" IN ({content_type_placeholders})
-            {user_filter}
-            AND 1 - (embedding <=> '{embedding_str}'::vector) >= ${len(params) + 1}
-            ORDER BY similarity DESC
-            LIMIT $1
-        """
-        params.append(min_similarity)
-
-        try:
-            results = await query_raw_with_schema(
-                sql, *params, set_public_search_path=True
-            )
-            return [
-                {
-                    "content_id": row["content_id"],
-                    "content_type": row["content_type"],
-                    "searchable_text": row["searchable_text"],
-                    "metadata": row["metadata"],
-                    "similarity": float(row["similarity"]),
-                }
-                for row in results
-            ]
-        except Exception as e:
-            logger.error(f"Semantic search failed: {e}")
-            # Fall through to lexical search below
-
-    # Fallback to lexical search if embeddings unavailable
-    logger.warning("Falling back to lexical search (embeddings unavailable)")
-
-    params_lexical: list[Any] = [limit]
-    user_filter = ""
-    if user_id is not None:
-        user_filter = 'AND "userId" = ${}'.format(len(params_lexical) + 1)
-        params_lexical.append(user_id)
-
-    # Add content type parameters and build placeholders dynamically
-    content_type_start_idx = len(params_lexical) + 1
-    content_type_placeholders_lexical = ", ".join(
-        f'${content_type_start_idx + i}::{{{{schema_prefix}}}}"ContentType"'
-        for i in range(len(content_types))
-    )
-    params_lexical.extend([ct.value for ct in content_types])
-
-    sql_lexical = f"""
-        SELECT
-            "contentId" as content_id,
-            "contentType" as content_type,
-            "searchableText" as searchable_text,
-            metadata,
-            0.0 as similarity
-        FROM {{{{schema_prefix}}}}"UnifiedContentEmbedding"
-        WHERE "contentType" IN ({content_type_placeholders_lexical})
-        {user_filter}
-        AND "searchableText" ILIKE ${len(params_lexical) + 1}
-        ORDER BY "updatedAt" DESC
-        LIMIT $1
-    """
-    params_lexical.append(f"%{query}%")
-
-    try:
-        results = await query_raw_with_schema(
-            sql_lexical, *params_lexical, set_public_search_path=True
-        )
-        return [
-            {
-                "content_id": row["content_id"],
-                "content_type": row["content_type"],
-                "searchable_text": row["searchable_text"],
-                "metadata": row["metadata"],
-                "similarity": 0.0,  # Lexical search doesn't provide similarity
-            }
-            for row in results
-        ]
-    except Exception as e:
-        logger.error(f"Lexical search failed: {e}")
-        return []

autogpt_platform/backend/backend/api/features/store/embeddings_e2e_test.py

@@ -1,666 +0,0 @@
-"""
-End-to-end database tests for embeddings and hybrid search.
-
-These tests hit the actual database to verify SQL queries work correctly.
-Tests cover:
-1. Embedding storage (store_content_embedding)
-2. Embedding retrieval (get_content_embedding)
-3. Embedding deletion (delete_content_embedding)
-4. Unified hybrid search across content types
-5. Store agent hybrid search
-"""
-
-import uuid
-from typing import AsyncGenerator
-
-import pytest
-from prisma.enums import ContentType
-
-from backend.api.features.store import embeddings
-from backend.api.features.store.embeddings import EMBEDDING_DIM
-from backend.api.features.store.hybrid_search import (
-    hybrid_search,
-    unified_hybrid_search,
-)
-
-# ============================================================================
-# Test Fixtures
-# ============================================================================
-
-
-@pytest.fixture
-def test_content_id() -> str:
-    """Generate unique content ID for test isolation."""
-    return f"test-content-{uuid.uuid4()}"
-
-
-@pytest.fixture
-def test_user_id() -> str:
-    """Generate unique user ID for test isolation."""
-    return f"test-user-{uuid.uuid4()}"
-
-
-@pytest.fixture
-def mock_embedding() -> list[float]:
-    """Generate a mock embedding vector."""
-    # Create a normalized embedding vector
-    import math
-
-    raw = [float(i % 10) / 10.0 for i in range(EMBEDDING_DIM)]
-    # Normalize to unit length (required for cosine similarity)
-    magnitude = math.sqrt(sum(x * x for x in raw))
-    return [x / magnitude for x in raw]
-
-
-@pytest.fixture
-def similar_embedding() -> list[float]:
-    """Generate an embedding similar to mock_embedding."""
-    import math
-
-    # Similar but slightly different values
-    raw = [float(i % 10) / 10.0 + 0.01 for i in range(EMBEDDING_DIM)]
-    magnitude = math.sqrt(sum(x * x for x in raw))
-    return [x / magnitude for x in raw]
-
-
-@pytest.fixture
-def different_embedding() -> list[float]:
-    """Generate an embedding very different from mock_embedding."""
-    import math
-
-    # Reversed pattern to be maximally different
-    raw = [float((EMBEDDING_DIM - i) % 10) / 10.0 for i in range(EMBEDDING_DIM)]
-    magnitude = math.sqrt(sum(x * x for x in raw))
-    return [x / magnitude for x in raw]
-
-
-@pytest.fixture
-async def cleanup_embeddings(
-    server,
-) -> AsyncGenerator[list[tuple[ContentType, str, str | None]], None]:
-    """
-    Fixture that tracks created embeddings and cleans them up after tests.
-
-    Yields a list to which tests can append (content_type, content_id, user_id) tuples.
-    """
-    created_embeddings: list[tuple[ContentType, str, str | None]] = []
-    yield created_embeddings
-
-    # Cleanup all created embeddings
-    for content_type, content_id, user_id in created_embeddings:
-        try:
-            await embeddings.delete_content_embedding(content_type, content_id, user_id)
-        except Exception:
-            pass  # Ignore cleanup errors
-
-
-# ============================================================================
-# store_content_embedding Tests
-# ============================================================================
-
-
-@pytest.mark.asyncio(loop_scope="session")
-async def test_store_content_embedding_store_agent(
-    server,
-    test_content_id: str,
-    mock_embedding: list[float],
-    cleanup_embeddings: list,
-):
-    """Test storing embedding for STORE_AGENT content type."""
-    # Track for cleanup
-    cleanup_embeddings.append((ContentType.STORE_AGENT, test_content_id, None))
-
-    result = await embeddings.store_content_embedding(
-        content_type=ContentType.STORE_AGENT,
-        content_id=test_content_id,
-        embedding=mock_embedding,
-        searchable_text="AI assistant for productivity tasks",
-        metadata={"name": "Test Agent", "categories": ["productivity"]},
-        user_id=None,  # Store agents are public
-    )
-
-    assert result is True
-
-    # Verify it was stored
-    stored = await embeddings.get_content_embedding(
-        ContentType.STORE_AGENT, test_content_id, user_id=None
-    )
-    assert stored is not None
-    assert stored["contentId"] == test_content_id
-    assert stored["contentType"] == "STORE_AGENT"
-    assert stored["searchableText"] == "AI assistant for productivity tasks"
-
-
-@pytest.mark.asyncio(loop_scope="session")
-async def test_store_content_embedding_block(
-    server,
-    test_content_id: str,
-    mock_embedding: list[float],
-    cleanup_embeddings: list,
-):
-    """Test storing embedding for BLOCK content type."""
-    cleanup_embeddings.append((ContentType.BLOCK, test_content_id, None))
-
-    result = await embeddings.store_content_embedding(
-        content_type=ContentType.BLOCK,
-        content_id=test_content_id,
-        embedding=mock_embedding,
-        searchable_text="HTTP request block for API calls",
-        metadata={"name": "HTTP Request Block"},
-        user_id=None,  # Blocks are public
-    )
-
-    assert result is True
-
-    stored = await embeddings.get_content_embedding(
-        ContentType.BLOCK, test_content_id, user_id=None
-    )
-    assert stored is not None
-    assert stored["contentType"] == "BLOCK"
-
-
-@pytest.mark.asyncio(loop_scope="session")
-async def test_store_content_embedding_documentation(
-    server,
-    test_content_id: str,
-    mock_embedding: list[float],
-    cleanup_embeddings: list,
-):
-    """Test storing embedding for DOCUMENTATION content type."""
-    cleanup_embeddings.append((ContentType.DOCUMENTATION, test_content_id, None))
-
-    result = await embeddings.store_content_embedding(
-        content_type=ContentType.DOCUMENTATION,
-        content_id=test_content_id,
-        embedding=mock_embedding,
-        searchable_text="Getting started guide for AutoGPT platform",
-        metadata={"title": "Getting Started", "url": "/docs/getting-started"},
-        user_id=None,  # Docs are public
-    )
-
-    assert result is True
-
-    stored = await embeddings.get_content_embedding(
-        ContentType.DOCUMENTATION, test_content_id, user_id=None
-    )
-    assert stored is not None
-    assert stored["contentType"] == "DOCUMENTATION"
-
-
-@pytest.mark.asyncio(loop_scope="session")
-async def test_store_content_embedding_upsert(
-    server,
-    test_content_id: str,
-    mock_embedding: list[float],
-    cleanup_embeddings: list,
-):
-    """Test that storing embedding twice updates instead of duplicates."""
-    cleanup_embeddings.append((ContentType.BLOCK, test_content_id, None))
-
-    # Store first time
-    result1 = await embeddings.store_content_embedding(
-        content_type=ContentType.BLOCK,
-        content_id=test_content_id,
-        embedding=mock_embedding,
-        searchable_text="Original text",
-        metadata={"version": 1},
-        user_id=None,
-    )
-    assert result1 is True
-
-    # Store again with different text (upsert)
-    result2 = await embeddings.store_content_embedding(
-        content_type=ContentType.BLOCK,
-        content_id=test_content_id,
-        embedding=mock_embedding,
-        searchable_text="Updated text",
-        metadata={"version": 2},
-        user_id=None,
-    )
-    assert result2 is True
-
-    # Verify only one record with updated text
-    stored = await embeddings.get_content_embedding(
-        ContentType.BLOCK, test_content_id, user_id=None
-    )
-    assert stored is not None
-    assert stored["searchableText"] == "Updated text"
-
-
-# ============================================================================
-# get_content_embedding Tests
-# ============================================================================
-
-
-@pytest.mark.asyncio(loop_scope="session")
-async def test_get_content_embedding_not_found(server):
-    """Test retrieving non-existent embedding returns None."""
-    result = await embeddings.get_content_embedding(
-        ContentType.STORE_AGENT, "non-existent-id", user_id=None
-    )
-    assert result is None
-
-
-@pytest.mark.asyncio(loop_scope="session")
-async def test_get_content_embedding_with_metadata(
-    server,
-    test_content_id: str,
-    mock_embedding: list[float],
-    cleanup_embeddings: list,
-):
-    """Test that metadata is correctly stored and retrieved."""
-    cleanup_embeddings.append((ContentType.STORE_AGENT, test_content_id, None))
-
-    metadata = {
-        "name": "Test Agent",
-        "subHeading": "A test agent",
-        "categories": ["ai", "productivity"],
-        "customField": 123,
-    }
-
-    await embeddings.store_content_embedding(
-        content_type=ContentType.STORE_AGENT,
-        content_id=test_content_id,
-        embedding=mock_embedding,
-        searchable_text="test",
-        metadata=metadata,
-        user_id=None,
-    )
-
-    stored = await embeddings.get_content_embedding(
-        ContentType.STORE_AGENT, test_content_id, user_id=None
-    )
-
-    assert stored is not None
-    assert stored["metadata"]["name"] == "Test Agent"
-    assert stored["metadata"]["categories"] == ["ai", "productivity"]
-    assert stored["metadata"]["customField"] == 123
-
-
-# ============================================================================
-# delete_content_embedding Tests
-# ============================================================================
-
-
-@pytest.mark.asyncio(loop_scope="session")
-async def test_delete_content_embedding(
-    server,
-    test_content_id: str,
-    mock_embedding: list[float],
-):
-    """Test deleting embedding removes it from database."""
-    # Store embedding
-    await embeddings.store_content_embedding(
-        content_type=ContentType.BLOCK,
-        content_id=test_content_id,
-        embedding=mock_embedding,
-        searchable_text="To be deleted",
-        metadata=None,
-        user_id=None,
-    )
-
-    # Verify it exists
-    stored = await embeddings.get_content_embedding(
-        ContentType.BLOCK, test_content_id, user_id=None
-    )
-    assert stored is not None
-
-    # Delete it
-    result = await embeddings.delete_content_embedding(
-        ContentType.BLOCK, test_content_id, user_id=None
-    )
-    assert result is True
-
-    # Verify it's gone
-    stored = await embeddings.get_content_embedding(
-        ContentType.BLOCK, test_content_id, user_id=None
-    )
-    assert stored is None
-
-
-@pytest.mark.asyncio(loop_scope="session")
-async def test_delete_content_embedding_not_found(server):
-    """Test deleting non-existent embedding doesn't error."""
-    result = await embeddings.delete_content_embedding(
-        ContentType.BLOCK, "non-existent-id", user_id=None
-    )
-    # Should succeed even if nothing to delete
-    assert result is True
-
-
-# ============================================================================
-# unified_hybrid_search Tests
-# ============================================================================
-
-
-@pytest.mark.asyncio(loop_scope="session")
-async def test_unified_hybrid_search_finds_matching_content(
-    server,
-    mock_embedding: list[float],
-    cleanup_embeddings: list,
-):
-    """Test unified search finds content matching the query."""
-    # Create unique content IDs
-    agent_id = f"test-agent-{uuid.uuid4()}"
-    block_id = f"test-block-{uuid.uuid4()}"
-    doc_id = f"test-doc-{uuid.uuid4()}"
-
-    cleanup_embeddings.append((ContentType.STORE_AGENT, agent_id, None))
-    cleanup_embeddings.append((ContentType.BLOCK, block_id, None))
-    cleanup_embeddings.append((ContentType.DOCUMENTATION, doc_id, None))
-
-    # Store embeddings for different content types
-    await embeddings.store_content_embedding(
-        content_type=ContentType.STORE_AGENT,
-        content_id=agent_id,
-        embedding=mock_embedding,
-        searchable_text="AI writing assistant for blog posts",
-        metadata={"name": "Writing Assistant"},
-        user_id=None,
-    )
-
-    await embeddings.store_content_embedding(
-        content_type=ContentType.BLOCK,
-        content_id=block_id,
-        embedding=mock_embedding,
-        searchable_text="Text generation block for creative writing",
-        metadata={"name": "Text Generator"},
-        user_id=None,
-    )
-
-    await embeddings.store_content_embedding(
-        content_type=ContentType.DOCUMENTATION,
-        content_id=doc_id,
-        embedding=mock_embedding,
-        searchable_text="How to use writing blocks in AutoGPT",
-        metadata={"title": "Writing Guide"},
-        user_id=None,
-    )
-
-    # Search for "writing" - should find all three
-    results, total = await unified_hybrid_search(
-        query="writing",
-        page=1,
-        page_size=20,
-    )
-
-    # Should find at least our test content (may find others too)
-    content_ids = [r["content_id"] for r in results]
-    assert agent_id in content_ids or total >= 1  # Lexical search should find it
-
-
-@pytest.mark.asyncio(loop_scope="session")
-async def test_unified_hybrid_search_filter_by_content_type(
-    server,
-    mock_embedding: list[float],
-    cleanup_embeddings: list,
-):
-    """Test unified search can filter by content type."""
-    agent_id = f"test-agent-{uuid.uuid4()}"
-    block_id = f"test-block-{uuid.uuid4()}"
-
-    cleanup_embeddings.append((ContentType.STORE_AGENT, agent_id, None))
-    cleanup_embeddings.append((ContentType.BLOCK, block_id, None))
-
-    # Store both types with same searchable text
-    await embeddings.store_content_embedding(
-        content_type=ContentType.STORE_AGENT,
-        content_id=agent_id,
-        embedding=mock_embedding,
-        searchable_text="unique_search_term_xyz123",
-        metadata={},
-        user_id=None,
-    )
-
-    await embeddings.store_content_embedding(
-        content_type=ContentType.BLOCK,
-        content_id=block_id,
-        embedding=mock_embedding,
-        searchable_text="unique_search_term_xyz123",
-        metadata={},
-        user_id=None,
-    )
-
-    # Search only for BLOCK type
-    results, total = await unified_hybrid_search(
-        query="unique_search_term_xyz123",
-        content_types=[ContentType.BLOCK],
-        page=1,
-        page_size=20,
-    )
-
-    # All results should be BLOCK type
-    for r in results:
-        assert r["content_type"] == "BLOCK"
-
-
-@pytest.mark.asyncio(loop_scope="session")
-async def test_unified_hybrid_search_empty_query(server):
-    """Test unified search with empty query returns empty results."""
-    results, total = await unified_hybrid_search(
-        query="",
-        page=1,
-        page_size=20,
-    )
-
-    assert results == []
-    assert total == 0
-
-
-@pytest.mark.asyncio(loop_scope="session")
-async def test_unified_hybrid_search_pagination(
-    server,
-    mock_embedding: list[float],
-    cleanup_embeddings: list,
-):
-    """Test unified search pagination works correctly."""
-    # Create multiple items
-    content_ids = []
-    for i in range(5):
-        content_id = f"test-pagination-{uuid.uuid4()}"
-        content_ids.append(content_id)
-        cleanup_embeddings.append((ContentType.BLOCK, content_id, None))
-
-        await embeddings.store_content_embedding(
-            content_type=ContentType.BLOCK,
-            content_id=content_id,
-            embedding=mock_embedding,
-            searchable_text=f"pagination test item number {i}",
-            metadata={"index": i},
-            user_id=None,
-        )
-
-    # Get first page
-    page1_results, total1 = await unified_hybrid_search(
-        query="pagination test",
-        content_types=[ContentType.BLOCK],
-        page=1,
-        page_size=2,
-    )
-
-    # Get second page
-    page2_results, total2 = await unified_hybrid_search(
-        query="pagination test",
-        content_types=[ContentType.BLOCK],
-        page=2,
-        page_size=2,
-    )
-
-    # Total should be consistent
-    assert total1 == total2
-
-    # Pages should have different content (if we have enough results)
-    if len(page1_results) > 0 and len(page2_results) > 0:
-        page1_ids = {r["content_id"] for r in page1_results}
-        page2_ids = {r["content_id"] for r in page2_results}
-        # No overlap between pages
-        assert page1_ids.isdisjoint(page2_ids)
-
-
-@pytest.mark.asyncio(loop_scope="session")
-async def test_unified_hybrid_search_min_score_filtering(
-    server,
-    mock_embedding: list[float],
-    cleanup_embeddings: list,
-):
-    """Test unified search respects min_score threshold."""
-    content_id = f"test-minscore-{uuid.uuid4()}"
-    cleanup_embeddings.append((ContentType.BLOCK, content_id, None))
-
-    await embeddings.store_content_embedding(
-        content_type=ContentType.BLOCK,
-        content_id=content_id,
-        embedding=mock_embedding,
-        searchable_text="completely unrelated content about bananas",
-        metadata={},
-        user_id=None,
-    )
-
-    # Search with very high min_score - should filter out low relevance
-    results_high, _ = await unified_hybrid_search(
-        query="quantum computing algorithms",
-        content_types=[ContentType.BLOCK],
-        min_score=0.9,  # Very high threshold
-        page=1,
-        page_size=20,
-    )
-
-    # Search with low min_score
-    results_low, _ = await unified_hybrid_search(
-        query="quantum computing algorithms",
-        content_types=[ContentType.BLOCK],
-        min_score=0.01,  # Very low threshold
-        page=1,
-        page_size=20,
-    )
-
-    # High threshold should have fewer or equal results
-    assert len(results_high) <= len(results_low)
-
-
-# ============================================================================
-# hybrid_search (Store Agents) Tests
-# ============================================================================
-
-
-@pytest.mark.asyncio(loop_scope="session")
-async def test_hybrid_search_store_agents_sql_valid(server):
-    """Test that hybrid_search SQL executes without errors."""
-    # This test verifies the SQL is syntactically correct
-    # even if no results are found
-    results, total = await hybrid_search(
-        query="test agent",
-        page=1,
-        page_size=20,
-    )
-
-    # Should not raise - verifies SQL is valid
-    assert isinstance(results, list)
-    assert isinstance(total, int)
-    assert total >= 0
-
-
-@pytest.mark.asyncio(loop_scope="session")
-async def test_hybrid_search_with_filters(server):
-    """Test hybrid_search with various filter options."""
-    # Test with all filter types
-    results, total = await hybrid_search(
-        query="productivity",
-        featured=True,
-        creators=["test-creator"],
-        category="productivity",
-        page=1,
-        page_size=10,
-    )
-
-    # Should not raise - verifies filter SQL is valid
-    assert isinstance(results, list)
-    assert isinstance(total, int)
-
-
-@pytest.mark.asyncio(loop_scope="session")
-async def test_hybrid_search_pagination(server):
-    """Test hybrid_search pagination."""
-    # Page 1
-    results1, total1 = await hybrid_search(
-        query="agent",
-        page=1,
-        page_size=5,
-    )
-
-    # Page 2
-    results2, total2 = await hybrid_search(
-        query="agent",
-        page=2,
-        page_size=5,
-    )
-
-    # Verify SQL executes without error
-    assert isinstance(results1, list)
-    assert isinstance(results2, list)
-    assert isinstance(total1, int)
-    assert isinstance(total2, int)
-
-    # If page 1 has results, total should be > 0
-    # Note: total from page 2 may be 0 if no results on that page (COUNT(*) OVER limitation)
-    if results1:
-        assert total1 > 0
-
-
-# ============================================================================
-# SQL Validity Tests (verify queries don't break)
-# ============================================================================
-
-
-@pytest.mark.asyncio(loop_scope="session")
-async def test_all_content_types_searchable(server):
-    """Test that all content types can be searched without SQL errors."""
-    for content_type in [
-        ContentType.STORE_AGENT,
-        ContentType.BLOCK,
-        ContentType.DOCUMENTATION,
-    ]:
-        results, total = await unified_hybrid_search(
-            query="test",
-            content_types=[content_type],
-            page=1,
-            page_size=10,
-        )
-
-        # Should not raise
-        assert isinstance(results, list)
-        assert isinstance(total, int)
-
-
-@pytest.mark.asyncio(loop_scope="session")
-async def test_multiple_content_types_searchable(server):
-    """Test searching multiple content types at once."""
-    results, total = await unified_hybrid_search(
-        query="test",
-        content_types=[ContentType.BLOCK, ContentType.DOCUMENTATION],
-        page=1,
-        page_size=20,
-    )
-
-    # Should not raise
-    assert isinstance(results, list)
-    assert isinstance(total, int)
-
-
-@pytest.mark.asyncio(loop_scope="session")
-async def test_search_all_content_types_default(server):
-    """Test searching all content types (default behavior)."""
-    results, total = await unified_hybrid_search(
-        query="test",
-        content_types=None,  # Should search all
-        page=1,
-        page_size=20,
-    )
-
-    # Should not raise
-    assert isinstance(results, list)
-    assert isinstance(total, int)
-
-
-if __name__ == "__main__":
-    pytest.main([__file__, "-v", "-s"])

autogpt_platform/backend/backend/api/features/store/embeddings_schema_test.py

@@ -1,315 +0,0 @@
-"""
-Integration tests for embeddings with schema handling.
-
-These tests verify that embeddings operations work correctly across different database schemas.
-"""
-
-from unittest.mock import AsyncMock, MagicMock, patch
-
-import pytest
-from prisma.enums import ContentType
-
-from backend.api.features.store import embeddings
-from backend.api.features.store.embeddings import EMBEDDING_DIM
-
-# Schema prefix tests removed - functionality moved to db.raw_with_schema() helper
-
-
-@pytest.mark.asyncio(loop_scope="session")
-@pytest.mark.integration
-async def test_store_content_embedding_with_schema():
-    """Test storing embeddings with proper schema handling."""
-    with patch("backend.data.db.get_database_schema") as mock_schema:
-        mock_schema.return_value = "platform"
-
-        with patch("prisma.get_client") as mock_get_client:
-            mock_client = AsyncMock()
-            mock_get_client.return_value = mock_client
-
-            result = await embeddings.store_content_embedding(
-                content_type=ContentType.STORE_AGENT,
-                content_id="test-id",
-                embedding=[0.1] * EMBEDDING_DIM,
-                searchable_text="test text",
-                metadata={"test": "data"},
-                user_id=None,
-            )
-
-            # Verify the query was called
-            assert mock_client.execute_raw.called
-
-            # Get the SQL query that was executed
-            call_args = mock_client.execute_raw.call_args
-            sql_query = call_args[0][0]
-
-            # Verify schema prefix is in the query
-            assert '"platform"."UnifiedContentEmbedding"' in sql_query
-
-            # Verify result
-            assert result is True
-
-
-@pytest.mark.asyncio(loop_scope="session")
-@pytest.mark.integration
-async def test_get_content_embedding_with_schema():
-    """Test retrieving embeddings with proper schema handling."""
-    with patch("backend.data.db.get_database_schema") as mock_schema:
-        mock_schema.return_value = "platform"
-
-        with patch("prisma.get_client") as mock_get_client:
-            mock_client = AsyncMock()
-            mock_client.query_raw.return_value = [
-                {
-                    "contentType": "STORE_AGENT",
-                    "contentId": "test-id",
-                    "userId": None,
-                    "embedding": "[0.1, 0.2]",
-                    "searchableText": "test",
-                    "metadata": {},
-                    "createdAt": "2024-01-01",
-                    "updatedAt": "2024-01-01",
-                }
-            ]
-            mock_get_client.return_value = mock_client
-
-            result = await embeddings.get_content_embedding(
-                ContentType.STORE_AGENT,
-                "test-id",
-                user_id=None,
-            )
-
-            # Verify the query was called
-            assert mock_client.query_raw.called
-
-            # Get the SQL query that was executed
-            call_args = mock_client.query_raw.call_args
-            sql_query = call_args[0][0]
-
-            # Verify schema prefix is in the query
-            assert '"platform"."UnifiedContentEmbedding"' in sql_query
-
-            # Verify result
-            assert result is not None
-            assert result["contentId"] == "test-id"
-
-
-@pytest.mark.asyncio(loop_scope="session")
-@pytest.mark.integration
-async def test_delete_content_embedding_with_schema():
-    """Test deleting embeddings with proper schema handling."""
-    with patch("backend.data.db.get_database_schema") as mock_schema:
-        mock_schema.return_value = "platform"
-
-        with patch("prisma.get_client") as mock_get_client:
-            mock_client = AsyncMock()
-            mock_get_client.return_value = mock_client
-
-            result = await embeddings.delete_content_embedding(
-                ContentType.STORE_AGENT,
-                "test-id",
-            )
-
-            # Verify the query was called
-            assert mock_client.execute_raw.called
-
-            # Get the SQL query that was executed
-            call_args = mock_client.execute_raw.call_args
-            sql_query = call_args[0][0]
-
-            # Verify schema prefix is in the query
-            assert '"platform"."UnifiedContentEmbedding"' in sql_query
-
-            # Verify result
-            assert result is True
-
-
-@pytest.mark.asyncio(loop_scope="session")
-@pytest.mark.integration
-async def test_get_embedding_stats_with_schema():
-    """Test embedding statistics with proper schema handling via content handlers."""
-    # Mock handler to return stats
-    mock_handler = MagicMock()
-    mock_handler.get_stats = AsyncMock(
-        return_value={
-            "total": 100,
-            "with_embeddings": 80,
-            "without_embeddings": 20,
-        }
-    )
-
-    with patch(
-        "backend.api.features.store.embeddings.CONTENT_HANDLERS",
-        {ContentType.STORE_AGENT: mock_handler},
-    ):
-        result = await embeddings.get_embedding_stats()
-
-        # Verify handler was called
-        mock_handler.get_stats.assert_called_once()
-
-        # Verify new result structure
-        assert "by_type" in result
-        assert "totals" in result
-        assert result["totals"]["total"] == 100
-        assert result["totals"]["with_embeddings"] == 80
-        assert result["totals"]["without_embeddings"] == 20
-        assert result["totals"]["coverage_percent"] == 80.0
-
-
-@pytest.mark.asyncio(loop_scope="session")
-@pytest.mark.integration
-async def test_backfill_missing_embeddings_with_schema():
-    """Test backfilling embeddings via content handlers."""
-    from backend.api.features.store.content_handlers import ContentItem
-
-    # Create mock content item
-    mock_item = ContentItem(
-        content_id="version-1",
-        content_type=ContentType.STORE_AGENT,
-        searchable_text="Test Agent Test description",
-        metadata={"name": "Test Agent"},
-    )
-
-    # Mock handler
-    mock_handler = MagicMock()
-    mock_handler.get_missing_items = AsyncMock(return_value=[mock_item])
-
-    with patch(
-        "backend.api.features.store.embeddings.CONTENT_HANDLERS",
-        {ContentType.STORE_AGENT: mock_handler},
-    ):
-        with patch(
-            "backend.api.features.store.embeddings.generate_embedding",
-            return_value=[0.1] * EMBEDDING_DIM,
-        ):
-            with patch(
-                "backend.api.features.store.embeddings.store_content_embedding",
-                return_value=True,
-            ):
-                result = await embeddings.backfill_missing_embeddings(batch_size=10)
-
-                # Verify handler was called
-                mock_handler.get_missing_items.assert_called_once_with(10)
-
-                # Verify results
-                assert result["processed"] == 1
-                assert result["success"] == 1
-                assert result["failed"] == 0
-
-
-@pytest.mark.asyncio(loop_scope="session")
-@pytest.mark.integration
-async def test_ensure_content_embedding_with_schema():
-    """Test ensuring embeddings exist with proper schema handling."""
-    with patch("backend.data.db.get_database_schema") as mock_schema:
-        mock_schema.return_value = "platform"
-
-        with patch(
-            "backend.api.features.store.embeddings.get_content_embedding"
-        ) as mock_get:
-            # Simulate no existing embedding
-            mock_get.return_value = None
-
-            with patch(
-                "backend.api.features.store.embeddings.generate_embedding"
-            ) as mock_generate:
-                mock_generate.return_value = [0.1] * EMBEDDING_DIM
-
-                with patch(
-                    "backend.api.features.store.embeddings.store_content_embedding"
-                ) as mock_store:
-                    mock_store.return_value = True
-
-                    result = await embeddings.ensure_content_embedding(
-                        content_type=ContentType.STORE_AGENT,
-                        content_id="test-id",
-                        searchable_text="test text",
-                        metadata={"test": "data"},
-                        user_id=None,
-                        force=False,
-                    )
-
-                    # Verify the flow
-                    assert mock_get.called
-                    assert mock_generate.called
-                    assert mock_store.called
-                    assert result is True
-
-
-@pytest.mark.asyncio(loop_scope="session")
-@pytest.mark.integration
-async def test_backward_compatibility_store_embedding():
-    """Test backward compatibility wrapper for store_embedding."""
-    with patch(
-        "backend.api.features.store.embeddings.store_content_embedding"
-    ) as mock_store:
-        mock_store.return_value = True
-
-        result = await embeddings.store_embedding(
-            version_id="test-version-id",
-            embedding=[0.1] * EMBEDDING_DIM,
-            tx=None,
-        )
-
-        # Verify it calls the new function with correct parameters
-        assert mock_store.called
-        call_args = mock_store.call_args
-
-        assert call_args[1]["content_type"] == ContentType.STORE_AGENT
-        assert call_args[1]["content_id"] == "test-version-id"
-        assert call_args[1]["user_id"] is None
-        assert result is True
-
-
-@pytest.mark.asyncio(loop_scope="session")
-@pytest.mark.integration
-async def test_backward_compatibility_get_embedding():
-    """Test backward compatibility wrapper for get_embedding."""
-    with patch(
-        "backend.api.features.store.embeddings.get_content_embedding"
-    ) as mock_get:
-        mock_get.return_value = {
-            "contentType": "STORE_AGENT",
-            "contentId": "test-version-id",
-            "embedding": "[0.1, 0.2]",
-            "createdAt": "2024-01-01",
-            "updatedAt": "2024-01-01",
-        }
-
-        result = await embeddings.get_embedding("test-version-id")
-
-        # Verify it calls the new function
-        assert mock_get.called
-
-        # Verify it transforms to old format
-        assert result is not None
-        assert result["storeListingVersionId"] == "test-version-id"
-        assert "embedding" in result
-
-
-@pytest.mark.asyncio(loop_scope="session")
-@pytest.mark.integration
-async def test_schema_handling_error_cases():
-    """Test error handling in schema-aware operations."""
-    with patch("backend.data.db.get_database_schema") as mock_schema:
-        mock_schema.return_value = "platform"
-
-        with patch("prisma.get_client") as mock_get_client:
-            mock_client = AsyncMock()
-            mock_client.execute_raw.side_effect = Exception("Database error")
-            mock_get_client.return_value = mock_client
-
-            result = await embeddings.store_content_embedding(
-                content_type=ContentType.STORE_AGENT,
-                content_id="test-id",
-                embedding=[0.1] * EMBEDDING_DIM,
-                searchable_text="test",
-                metadata=None,
-                user_id=None,
-            )
-
-            # Should return False on error, not raise
-            assert result is False
-
-
-if __name__ == "__main__":
-    pytest.main([__file__, "-v", "-s"])

autogpt_platform/backend/backend/api/features/store/embeddings_test.py

@@ -1,407 +0,0 @@
-from unittest.mock import AsyncMock, MagicMock, patch
-
-import prisma
-import pytest
-from prisma import Prisma
-from prisma.enums import ContentType
-
-from backend.api.features.store import embeddings
-
-
-@pytest.fixture(autouse=True)
-async def setup_prisma():
-    """Setup Prisma client for tests."""
-    try:
-        Prisma()
-    except prisma.errors.ClientAlreadyRegisteredError:
-        pass
-    yield
-
-
-@pytest.mark.asyncio(loop_scope="session")
-async def test_build_searchable_text():
-    """Test searchable text building from listing fields."""
-    result = embeddings.build_searchable_text(
-        name="AI Assistant",
-        description="A helpful AI assistant for productivity",
-        sub_heading="Boost your productivity",
-        categories=["AI", "Productivity"],
-    )
-
-    expected = "AI Assistant Boost your productivity A helpful AI assistant for productivity AI Productivity"
-    assert result == expected
-
-
-@pytest.mark.asyncio(loop_scope="session")
-async def test_build_searchable_text_empty_fields():
-    """Test searchable text building with empty fields."""
-    result = embeddings.build_searchable_text(
-        name="", description="Test description", sub_heading="", categories=[]
-    )
-
-    assert result == "Test description"
-
-
-@pytest.mark.asyncio(loop_scope="session")
-async def test_generate_embedding_success():
-    """Test successful embedding generation."""
-    # Mock OpenAI response
-    mock_client = MagicMock()
-    mock_response = MagicMock()
-    mock_response.data = [MagicMock()]
-    mock_response.data[0].embedding = [0.1, 0.2, 0.3] * 512  # 1536 dimensions
-
-    # Use AsyncMock for async embeddings.create method
-    mock_client.embeddings.create = AsyncMock(return_value=mock_response)
-
-    # Patch at the point of use in embeddings.py
-    with patch(
-        "backend.api.features.store.embeddings.get_openai_client"
-    ) as mock_get_client:
-        mock_get_client.return_value = mock_client
-
-        result = await embeddings.generate_embedding("test text")
-
-        assert result is not None
-        assert len(result) == embeddings.EMBEDDING_DIM
-        assert result[0] == 0.1
-
-        mock_client.embeddings.create.assert_called_once_with(
-            model="text-embedding-3-small", input="test text"
-        )
-
-
-@pytest.mark.asyncio(loop_scope="session")
-async def test_generate_embedding_no_api_key():
-    """Test embedding generation without API key."""
-    # Patch at the point of use in embeddings.py
-    with patch(
-        "backend.api.features.store.embeddings.get_openai_client"
-    ) as mock_get_client:
-        mock_get_client.return_value = None
-
-        result = await embeddings.generate_embedding("test text")
-
-        assert result is None
-
-
-@pytest.mark.asyncio(loop_scope="session")
-async def test_generate_embedding_api_error():
-    """Test embedding generation with API error."""
-    mock_client = MagicMock()
-    mock_client.embeddings.create = AsyncMock(side_effect=Exception("API Error"))
-
-    # Patch at the point of use in embeddings.py
-    with patch(
-        "backend.api.features.store.embeddings.get_openai_client"
-    ) as mock_get_client:
-        mock_get_client.return_value = mock_client
-
-        result = await embeddings.generate_embedding("test text")
-
-        assert result is None
-
-
-@pytest.mark.asyncio(loop_scope="session")
-async def test_generate_embedding_text_truncation():
-    """Test that long text is properly truncated using tiktoken."""
-    from tiktoken import encoding_for_model
-
-    mock_client = MagicMock()
-    mock_response = MagicMock()
-    mock_response.data = [MagicMock()]
-    mock_response.data[0].embedding = [0.1] * embeddings.EMBEDDING_DIM
-
-    # Use AsyncMock for async embeddings.create method
-    mock_client.embeddings.create = AsyncMock(return_value=mock_response)
-
-    # Patch at the point of use in embeddings.py
-    with patch(
-        "backend.api.features.store.embeddings.get_openai_client"
-    ) as mock_get_client:
-        mock_get_client.return_value = mock_client
-
-        # Create text that will exceed 8191 tokens
-        # Use varied characters to ensure token-heavy text: each word is ~1 token
-        words = [f"word{i}" for i in range(10000)]
-        long_text = " ".join(words)  # ~10000 tokens
-
-        await embeddings.generate_embedding(long_text)
-
-        # Verify text was truncated to 8191 tokens
-        call_args = mock_client.embeddings.create.call_args
-        truncated_text = call_args.kwargs["input"]
-
-        # Count actual tokens in truncated text
-        enc = encoding_for_model("text-embedding-3-small")
-        actual_tokens = len(enc.encode(truncated_text))
-
-        # Should be at or just under 8191 tokens
-        assert actual_tokens <= 8191
-        # Should be close to the limit (not over-truncated)
-        assert actual_tokens >= 8100
-
-
-@pytest.mark.asyncio(loop_scope="session")
-async def test_store_embedding_success(mocker):
-    """Test successful embedding storage."""
-    mock_client = mocker.AsyncMock()
-    mock_client.execute_raw = mocker.AsyncMock()
-
-    embedding = [0.1, 0.2, 0.3]
-
-    result = await embeddings.store_embedding(
-        version_id="test-version-id", embedding=embedding, tx=mock_client
-    )
-
-    assert result is True
-    # execute_raw is called twice: once for SET search_path, once for INSERT
-    assert mock_client.execute_raw.call_count == 2
-
-    # First call: SET search_path
-    first_call_args = mock_client.execute_raw.call_args_list[0][0]
-    assert "SET search_path" in first_call_args[0]
-
-    # Second call: INSERT query with the actual data
-    second_call_args = mock_client.execute_raw.call_args_list[1][0]
-    assert "test-version-id" in second_call_args
-    assert "[0.1,0.2,0.3]" in second_call_args
-    assert None in second_call_args  # userId should be None for store agents
-
-
-@pytest.mark.asyncio(loop_scope="session")
-async def test_store_embedding_database_error(mocker):
-    """Test embedding storage with database error."""
-    mock_client = mocker.AsyncMock()
-    mock_client.execute_raw.side_effect = Exception("Database error")
-
-    embedding = [0.1, 0.2, 0.3]
-
-    result = await embeddings.store_embedding(
-        version_id="test-version-id", embedding=embedding, tx=mock_client
-    )
-
-    assert result is False
-
-
-@pytest.mark.asyncio(loop_scope="session")
-async def test_get_embedding_success():
-    """Test successful embedding retrieval."""
-    mock_result = [
-        {
-            "contentType": "STORE_AGENT",
-            "contentId": "test-version-id",
-            "userId": None,
-            "embedding": "[0.1,0.2,0.3]",
-            "searchableText": "Test text",
-            "metadata": {},
-            "createdAt": "2024-01-01T00:00:00Z",
-            "updatedAt": "2024-01-01T00:00:00Z",
-        }
-    ]
-
-    with patch(
-        "backend.api.features.store.embeddings.query_raw_with_schema",
-        return_value=mock_result,
-    ):
-        result = await embeddings.get_embedding("test-version-id")
-
-        assert result is not None
-        assert result["storeListingVersionId"] == "test-version-id"
-        assert result["embedding"] == "[0.1,0.2,0.3]"
-
-
-@pytest.mark.asyncio(loop_scope="session")
-async def test_get_embedding_not_found():
-    """Test embedding retrieval when not found."""
-    with patch(
-        "backend.api.features.store.embeddings.query_raw_with_schema",
-        return_value=[],
-    ):
-        result = await embeddings.get_embedding("test-version-id")
-
-        assert result is None
-
-
-@pytest.mark.asyncio(loop_scope="session")
-@patch("backend.api.features.store.embeddings.generate_embedding")
-@patch("backend.api.features.store.embeddings.store_embedding")
-@patch("backend.api.features.store.embeddings.get_embedding")
-async def test_ensure_embedding_already_exists(mock_get, mock_store, mock_generate):
-    """Test ensure_embedding when embedding already exists."""
-    mock_get.return_value = {"embedding": "[0.1,0.2,0.3]"}
-
-    result = await embeddings.ensure_embedding(
-        version_id="test-id",
-        name="Test",
-        description="Test description",
-        sub_heading="Test heading",
-        categories=["test"],
-    )
-
-    assert result is True
-    mock_generate.assert_not_called()
-    mock_store.assert_not_called()
-
-
-@pytest.mark.asyncio(loop_scope="session")
-@patch("backend.api.features.store.embeddings.generate_embedding")
-@patch("backend.api.features.store.embeddings.store_content_embedding")
-@patch("backend.api.features.store.embeddings.get_embedding")
-async def test_ensure_embedding_create_new(mock_get, mock_store, mock_generate):
-    """Test ensure_embedding creating new embedding."""
-    mock_get.return_value = None
-    mock_generate.return_value = [0.1, 0.2, 0.3]
-    mock_store.return_value = True
-
-    result = await embeddings.ensure_embedding(
-        version_id="test-id",
-        name="Test",
-        description="Test description",
-        sub_heading="Test heading",
-        categories=["test"],
-    )
-
-    assert result is True
-    mock_generate.assert_called_once_with("Test Test heading Test description test")
-    mock_store.assert_called_once_with(
-        content_type=ContentType.STORE_AGENT,
-        content_id="test-id",
-        embedding=[0.1, 0.2, 0.3],
-        searchable_text="Test Test heading Test description test",
-        metadata={"name": "Test", "subHeading": "Test heading", "categories": ["test"]},
-        user_id=None,
-        tx=None,
-    )
-
-
-@pytest.mark.asyncio(loop_scope="session")
-@patch("backend.api.features.store.embeddings.generate_embedding")
-@patch("backend.api.features.store.embeddings.get_embedding")
-async def test_ensure_embedding_generation_fails(mock_get, mock_generate):
-    """Test ensure_embedding when generation fails."""
-    mock_get.return_value = None
-    mock_generate.return_value = None
-
-    result = await embeddings.ensure_embedding(
-        version_id="test-id",
-        name="Test",
-        description="Test description",
-        sub_heading="Test heading",
-        categories=["test"],
-    )
-
-    assert result is False
-
-
-@pytest.mark.asyncio(loop_scope="session")
-async def test_get_embedding_stats():
-    """Test embedding statistics retrieval."""
-    # Mock handler stats for each content type
-    mock_handler = MagicMock()
-    mock_handler.get_stats = AsyncMock(
-        return_value={
-            "total": 100,
-            "with_embeddings": 75,
-            "without_embeddings": 25,
-        }
-    )
-
-    # Patch the CONTENT_HANDLERS where it's used (in embeddings module)
-    with patch(
-        "backend.api.features.store.embeddings.CONTENT_HANDLERS",
-        {ContentType.STORE_AGENT: mock_handler},
-    ):
-        result = await embeddings.get_embedding_stats()
-
-        assert "by_type" in result
-        assert "totals" in result
-        assert result["totals"]["total"] == 100
-        assert result["totals"]["with_embeddings"] == 75
-        assert result["totals"]["without_embeddings"] == 25
-        assert result["totals"]["coverage_percent"] == 75.0
-
-
-@pytest.mark.asyncio(loop_scope="session")
-@patch("backend.api.features.store.embeddings.store_content_embedding")
-async def test_backfill_missing_embeddings_success(mock_store):
-    """Test backfill with successful embedding generation."""
-    # Mock ContentItem from handlers
-    from backend.api.features.store.content_handlers import ContentItem
-
-    mock_items = [
-        ContentItem(
-            content_id="version-1",
-            content_type=ContentType.STORE_AGENT,
-            searchable_text="Agent 1 Description 1",
-            metadata={"name": "Agent 1"},
-        ),
-        ContentItem(
-            content_id="version-2",
-            content_type=ContentType.STORE_AGENT,
-            searchable_text="Agent 2 Description 2",
-            metadata={"name": "Agent 2"},
-        ),
-    ]
-
-    # Mock handler to return missing items
-    mock_handler = MagicMock()
-    mock_handler.get_missing_items = AsyncMock(return_value=mock_items)
-
-    # Mock store_content_embedding to succeed for first, fail for second
-    mock_store.side_effect = [True, False]
-
-    with patch(
-        "backend.api.features.store.embeddings.CONTENT_HANDLERS",
-        {ContentType.STORE_AGENT: mock_handler},
-    ):
-        with patch(
-            "backend.api.features.store.embeddings.generate_embedding",
-            return_value=[0.1] * embeddings.EMBEDDING_DIM,
-        ):
-            result = await embeddings.backfill_missing_embeddings(batch_size=5)
-
-            assert result["processed"] == 2
-            assert result["success"] == 1
-            assert result["failed"] == 1
-            assert mock_store.call_count == 2
-
-
-@pytest.mark.asyncio(loop_scope="session")
-async def test_backfill_missing_embeddings_no_missing():
-    """Test backfill when no embeddings are missing."""
-    # Mock handler to return no missing items
-    mock_handler = MagicMock()
-    mock_handler.get_missing_items = AsyncMock(return_value=[])
-
-    with patch(
-        "backend.api.features.store.embeddings.CONTENT_HANDLERS",
-        {ContentType.STORE_AGENT: mock_handler},
-    ):
-        result = await embeddings.backfill_missing_embeddings(batch_size=5)
-
-        assert result["processed"] == 0
-        assert result["success"] == 0
-        assert result["failed"] == 0
-
-
-@pytest.mark.asyncio(loop_scope="session")
-async def test_embedding_to_vector_string():
-    """Test embedding to PostgreSQL vector string conversion."""
-    embedding = [0.1, 0.2, 0.3, -0.4]
-    result = embeddings.embedding_to_vector_string(embedding)
-    assert result == "[0.1,0.2,0.3,-0.4]"
-
-
-@pytest.mark.asyncio(loop_scope="session")
-async def test_embed_query():
-    """Test embed_query function (alias for generate_embedding)."""
-    with patch(
-        "backend.api.features.store.embeddings.generate_embedding"
-    ) as mock_generate:
-        mock_generate.return_value = [0.1, 0.2, 0.3]
-
-        result = await embeddings.embed_query("test query")
-
-        assert result == [0.1, 0.2, 0.3]
-        mock_generate.assert_called_once_with("test query")

autogpt_platform/backend/backend/api/features/store/hybrid_search.py

@@ -1,727 +0,0 @@
-"""
-Unified Hybrid Search
-
-Combines semantic (embedding) search with lexical (tsvector) search
-for improved relevance across all content types (agents, blocks, docs).
-Includes BM25 reranking for improved lexical relevance.
-"""
-
-import logging
-import re
-from dataclasses import dataclass
-from typing import Any, Literal
-
-from prisma.enums import ContentType
-from rank_bm25 import BM25Okapi
-
-from backend.api.features.store.embeddings import (
-    EMBEDDING_DIM,
-    embed_query,
-    embedding_to_vector_string,
-)
-from backend.data.db import query_raw_with_schema
-
-logger = logging.getLogger(__name__)
-
-
-# ============================================================================
-# BM25 Reranking
-# ============================================================================
-
-
-def tokenize(text: str) -> list[str]:
-    """Simple tokenizer for BM25 - lowercase and split on non-alphanumeric."""
-    if not text:
-        return []
-    # Lowercase and split on non-alphanumeric characters
-    tokens = re.findall(r"\b\w+\b", text.lower())
-    return tokens
-
-
-def bm25_rerank(
-    query: str,
-    results: list[dict[str, Any]],
-    text_field: str = "searchable_text",
-    bm25_weight: float = 0.3,
-    original_score_field: str = "combined_score",
-) -> list[dict[str, Any]]:
-    """
-    Rerank search results using BM25.
-
-    Combines the original combined_score with BM25 score for improved
-    lexical relevance, especially for exact term matches.
-
-    Args:
-        query: The search query
-        results: List of result dicts with text_field and original_score_field
-        text_field: Field name containing the text to score
-        bm25_weight: Weight for BM25 score (0-1). Original score gets (1 - bm25_weight)
-        original_score_field: Field name containing the original score
-
-    Returns:
-        Results list sorted by combined score (BM25 + original)
-    """
-    if not results or not query:
-        return results
-
-    # Extract texts and tokenize
-    corpus = [tokenize(r.get(text_field, "") or "") for r in results]
-
-    # Handle edge case where all documents are empty
-    if all(len(doc) == 0 for doc in corpus):
-        return results
-
-    # Build BM25 index
-    bm25 = BM25Okapi(corpus)
-
-    # Score query against corpus
-    query_tokens = tokenize(query)
-    if not query_tokens:
-        return results
-
-    bm25_scores = bm25.get_scores(query_tokens)
-
-    # Normalize BM25 scores to 0-1 range
-    max_bm25 = max(bm25_scores) if max(bm25_scores) > 0 else 1.0
-    normalized_bm25 = [s / max_bm25 for s in bm25_scores]
-
-    # Combine scores
-    original_weight = 1.0 - bm25_weight
-    for i, result in enumerate(results):
-        original_score = result.get(original_score_field, 0) or 0
-        result["bm25_score"] = normalized_bm25[i]
-        final_score = (
-            original_weight * original_score + bm25_weight * normalized_bm25[i]
-        )
-        result["final_score"] = final_score
-        result["relevance"] = final_score
-
-    # Sort by relevance descending
-    results.sort(key=lambda x: x.get("relevance", 0), reverse=True)
-
-    return results
-
-
-@dataclass
-class UnifiedSearchWeights:
-    """Weights for unified search (no popularity signal)."""
-
-    semantic: float = 0.40  # Embedding cosine similarity
-    lexical: float = 0.40  # tsvector ts_rank_cd score
-    category: float = 0.10  # Category match boost (for types that have categories)
-    recency: float = 0.10  # Newer content ranked higher
-
-    def __post_init__(self):
-        """Validate weights are non-negative and sum to approximately 1.0."""
-        total = self.semantic + self.lexical + self.category + self.recency
-
-        if any(
-            w < 0 for w in [self.semantic, self.lexical, self.category, self.recency]
-        ):
-            raise ValueError("All weights must be non-negative")
-
-        if not (0.99 <= total <= 1.01):
-            raise ValueError(f"Weights must sum to ~1.0, got {total:.3f}")
-
-
-# Default weights for unified search
-DEFAULT_UNIFIED_WEIGHTS = UnifiedSearchWeights()
-
-# Minimum relevance score thresholds
-DEFAULT_MIN_SCORE = 0.15  # For unified search (more permissive)
-DEFAULT_STORE_AGENT_MIN_SCORE = 0.20  # For store agent search (original threshold)
-
-
-async def unified_hybrid_search(
-    query: str,
-    content_types: list[ContentType] | None = None,
-    category: str | None = None,
-    page: int = 1,
-    page_size: int = 20,
-    weights: UnifiedSearchWeights | None = None,
-    min_score: float | None = None,
-    user_id: str | None = None,
-) -> tuple[list[dict[str, Any]], int]:
-    """
-    Unified hybrid search across all content types.
-
-    Searches UnifiedContentEmbedding using both semantic (vector) and lexical (tsvector) signals.
-
-    Args:
-        query: Search query string
-        content_types: List of content types to search. Defaults to all public types.
-        category: Filter by category (for content types that support it)
-        page: Page number (1-indexed)
-        page_size: Results per page
-        weights: Custom weights for search signals
-        min_score: Minimum relevance score threshold (0-1)
-        user_id: User ID for searching private content (library agents)
-
-    Returns:
-        Tuple of (results list, total count)
-    """
-    # Validate inputs
-    query = query.strip()
-    if not query:
-        return [], 0
-
-    if page < 1:
-        page = 1
-    if page_size < 1:
-        page_size = 1
-    if page_size > 100:
-        page_size = 100
-
-    if content_types is None:
-        content_types = [
-            ContentType.STORE_AGENT,
-            ContentType.BLOCK,
-            ContentType.DOCUMENTATION,
-        ]
-
-    if weights is None:
-        weights = DEFAULT_UNIFIED_WEIGHTS
-    if min_score is None:
-        min_score = DEFAULT_MIN_SCORE
-
-    offset = (page - 1) * page_size
-
-    # Generate query embedding
-    query_embedding = await embed_query(query)
-
-    # Graceful degradation if embedding unavailable
-    if query_embedding is None or not query_embedding:
-        logger.warning(
-            "Failed to generate query embedding - falling back to lexical-only search. "
-            "Check that openai_internal_api_key is configured and OpenAI API is accessible."
-        )
-        query_embedding = [0.0] * EMBEDDING_DIM
-        # Redistribute semantic weight to lexical
-        total_non_semantic = weights.lexical + weights.category + weights.recency
-        if total_non_semantic > 0:
-            factor = 1.0 / total_non_semantic
-            weights = UnifiedSearchWeights(
-                semantic=0.0,
-                lexical=weights.lexical * factor,
-                category=weights.category * factor,
-                recency=weights.recency * factor,
-            )
-        else:
-            weights = UnifiedSearchWeights(
-                semantic=0.0, lexical=1.0, category=0.0, recency=0.0
-            )
-
-    # Build parameters
-    params: list[Any] = []
-    param_idx = 1
-
-    # Query for lexical search
-    params.append(query)
-    query_param = f"${param_idx}"
-    param_idx += 1
-
-    # Query lowercase for category matching
-    params.append(query.lower())
-    query_lower_param = f"${param_idx}"
-    param_idx += 1
-
-    # Embedding
-    embedding_str = embedding_to_vector_string(query_embedding)
-    params.append(embedding_str)
-    embedding_param = f"${param_idx}"
-    param_idx += 1
-
-    # Content types
-    content_type_values = [ct.value for ct in content_types]
-    params.append(content_type_values)
-    content_types_param = f"${param_idx}"
-    param_idx += 1
-
-    # User ID filter (for private content)
-    user_filter = ""
-    if user_id is not None:
-        params.append(user_id)
-        user_filter = f'AND (uce."userId" = ${param_idx} OR uce."userId" IS NULL)'
-        param_idx += 1
-    else:
-        user_filter = 'AND uce."userId" IS NULL'
-
-    # Weights
-    params.append(weights.semantic)
-    w_semantic = f"${param_idx}"
-    param_idx += 1
-
-    params.append(weights.lexical)
-    w_lexical = f"${param_idx}"
-    param_idx += 1
-
-    params.append(weights.category)
-    w_category = f"${param_idx}"
-    param_idx += 1
-
-    params.append(weights.recency)
-    w_recency = f"${param_idx}"
-    param_idx += 1
-
-    # Min score
-    params.append(min_score)
-    min_score_param = f"${param_idx}"
-    param_idx += 1
-
-    # Pagination
-    params.append(page_size)
-    limit_param = f"${param_idx}"
-    param_idx += 1
-
-    params.append(offset)
-    offset_param = f"${param_idx}"
-    param_idx += 1
-
-    # Unified search query on UnifiedContentEmbedding
-    sql_query = f"""
-        WITH candidates AS (
-            -- Lexical matches (uses GIN index on search column)
-            SELECT uce.id, uce."contentType", uce."contentId"
-            FROM {{schema_prefix}}"UnifiedContentEmbedding" uce
-            WHERE uce."contentType" = ANY({content_types_param}::{{schema_prefix}}"ContentType"[])
-            {user_filter}
-            AND uce.search @@ plainto_tsquery('english', {query_param})
-
-            UNION
-
-            -- Semantic matches (uses HNSW index on embedding)
-            (
-                SELECT uce.id, uce."contentType", uce."contentId"
-                FROM {{schema_prefix}}"UnifiedContentEmbedding" uce
-                WHERE uce."contentType" = ANY({content_types_param}::{{schema_prefix}}"ContentType"[])
-                {user_filter}
-                ORDER BY uce.embedding <=> {embedding_param}::vector
-                LIMIT 200
-            )
-        ),
-        search_scores AS (
-            SELECT
-                uce."contentType" as content_type,
-                uce."contentId" as content_id,
-                uce."searchableText" as searchable_text,
-                uce.metadata,
-                uce."updatedAt" as updated_at,
-                -- Semantic score: cosine similarity (1 - distance)
-                COALESCE(1 - (uce.embedding <=> {embedding_param}::vector), 0) as semantic_score,
-                -- Lexical score: ts_rank_cd
-                COALESCE(ts_rank_cd(uce.search, plainto_tsquery('english', {query_param})), 0) as lexical_raw,
-                -- Category match from metadata
-                CASE
-                    WHEN uce.metadata ? 'categories' AND EXISTS (
-                        SELECT 1 FROM jsonb_array_elements_text(uce.metadata->'categories') cat
-                        WHERE LOWER(cat) LIKE '%' || {query_lower_param} || '%'
-                    )
-                    THEN 1.0
-                    ELSE 0.0
-                END as category_score,
-                -- Recency score: linear decay over 90 days
-                GREATEST(0, 1 - EXTRACT(EPOCH FROM (NOW() - uce."updatedAt")) / (90 * 24 * 3600)) as recency_score
-            FROM candidates c
-            INNER JOIN {{schema_prefix}}"UnifiedContentEmbedding" uce ON c.id = uce.id
-        ),
-        max_lexical AS (
-            SELECT GREATEST(MAX(lexical_raw), 0.001) as max_val FROM search_scores
-        ),
-        normalized AS (
-            SELECT
-                ss.*,
-                ss.lexical_raw / ml.max_val as lexical_score
-            FROM search_scores ss
-            CROSS JOIN max_lexical ml
-        ),
-        scored AS (
-            SELECT
-                content_type,
-                content_id,
-                searchable_text,
-                metadata,
-                updated_at,
-                semantic_score,
-                lexical_score,
-                category_score,
-                recency_score,
-                (
-                    {w_semantic} * semantic_score +
-                    {w_lexical} * lexical_score +
-                    {w_category} * category_score +
-                    {w_recency} * recency_score
-                ) as combined_score
-            FROM normalized
-        ),
-        filtered AS (
-            SELECT *, COUNT(*) OVER () as total_count
-            FROM scored
-            WHERE combined_score >= {min_score_param}
-        )
-        SELECT * FROM filtered
-        ORDER BY combined_score DESC
-        LIMIT {limit_param} OFFSET {offset_param}
-    """
-
-    results = await query_raw_with_schema(
-        sql_query, *params, set_public_search_path=True
-    )
-
-    total = results[0]["total_count"] if results else 0
-    # Apply BM25 reranking
-    if results:
-        results = bm25_rerank(
-            query=query,
-            results=results,
-            text_field="searchable_text",
-            bm25_weight=0.3,
-            original_score_field="combined_score",
-        )
-
-    # Clean up results
-    for result in results:
-        result.pop("total_count", None)
-
-    logger.info(f"Unified hybrid search: {len(results)} results, {total} total")
-
-    return results, total
-
-
-# ============================================================================
-# Store Agent specific search (with full metadata)
-# ============================================================================
-
-
-@dataclass
-class StoreAgentSearchWeights:
-    """Weights for store agent search including popularity."""
-
-    semantic: float = 0.30
-    lexical: float = 0.30
-    category: float = 0.20
-    recency: float = 0.10
-    popularity: float = 0.10
-
-    def __post_init__(self):
-        total = (
-            self.semantic
-            + self.lexical
-            + self.category
-            + self.recency
-            + self.popularity
-        )
-        if any(
-            w < 0
-            for w in [
-                self.semantic,
-                self.lexical,
-                self.category,
-                self.recency,
-                self.popularity,
-            ]
-        ):
-            raise ValueError("All weights must be non-negative")
-        if not (0.99 <= total <= 1.01):
-            raise ValueError(f"Weights must sum to ~1.0, got {total:.3f}")
-
-
-DEFAULT_STORE_AGENT_WEIGHTS = StoreAgentSearchWeights()
-
-
-async def hybrid_search(
-    query: str,
-    featured: bool = False,
-    creators: list[str] | None = None,
-    category: str | None = None,
-    sorted_by: (
-        Literal["relevance", "rating", "runs", "name", "updated_at"] | None
-    ) = None,
-    page: int = 1,
-    page_size: int = 20,
-    weights: StoreAgentSearchWeights | None = None,
-    min_score: float | None = None,
-) -> tuple[list[dict[str, Any]], int]:
-    """
-    Hybrid search for store agents with full metadata.
-
-    Uses UnifiedContentEmbedding for search, joins to StoreAgent for metadata.
-    """
-    query = query.strip()
-    if not query:
-        return [], 0
-
-    if page < 1:
-        page = 1
-    if page_size < 1:
-        page_size = 1
-    if page_size > 100:
-        page_size = 100
-
-    if weights is None:
-        weights = DEFAULT_STORE_AGENT_WEIGHTS
-    if min_score is None:
-        min_score = (
-            DEFAULT_STORE_AGENT_MIN_SCORE  # Use original threshold for store agents
-        )
-
-    offset = (page - 1) * page_size
-
-    # Generate query embedding
-    query_embedding = await embed_query(query)
-
-    # Graceful degradation
-    if query_embedding is None or not query_embedding:
-        logger.warning(
-            "Failed to generate query embedding - falling back to lexical-only search."
-        )
-        query_embedding = [0.0] * EMBEDDING_DIM
-        total_non_semantic = (
-            weights.lexical + weights.category + weights.recency + weights.popularity
-        )
-        if total_non_semantic > 0:
-            factor = 1.0 / total_non_semantic
-            weights = StoreAgentSearchWeights(
-                semantic=0.0,
-                lexical=weights.lexical * factor,
-                category=weights.category * factor,
-                recency=weights.recency * factor,
-                popularity=weights.popularity * factor,
-            )
-        else:
-            weights = StoreAgentSearchWeights(
-                semantic=0.0, lexical=1.0, category=0.0, recency=0.0, popularity=0.0
-            )
-
-    # Build parameters
-    params: list[Any] = []
-    param_idx = 1
-
-    params.append(query)
-    query_param = f"${param_idx}"
-    param_idx += 1
-
-    params.append(query.lower())
-    query_lower_param = f"${param_idx}"
-    param_idx += 1
-
-    embedding_str = embedding_to_vector_string(query_embedding)
-    params.append(embedding_str)
-    embedding_param = f"${param_idx}"
-    param_idx += 1
-
-    # Build WHERE clause for StoreAgent filters
-    where_parts = ["sa.is_available = true"]
-
-    if featured:
-        where_parts.append("sa.featured = true")
-
-    if creators:
-        params.append(creators)
-        where_parts.append(f"sa.creator_username = ANY(${param_idx})")
-        param_idx += 1
-
-    if category:
-        params.append(category)
-        where_parts.append(f"${param_idx} = ANY(sa.categories)")
-        param_idx += 1
-
-    where_clause = " AND ".join(where_parts)
-
-    # Weights
-    params.append(weights.semantic)
-    w_semantic = f"${param_idx}"
-    param_idx += 1
-
-    params.append(weights.lexical)
-    w_lexical = f"${param_idx}"
-    param_idx += 1
-
-    params.append(weights.category)
-    w_category = f"${param_idx}"
-    param_idx += 1
-
-    params.append(weights.recency)
-    w_recency = f"${param_idx}"
-    param_idx += 1
-
-    params.append(weights.popularity)
-    w_popularity = f"${param_idx}"
-    param_idx += 1
-
-    params.append(min_score)
-    min_score_param = f"${param_idx}"
-    param_idx += 1
-
-    params.append(page_size)
-    limit_param = f"${param_idx}"
-    param_idx += 1
-
-    params.append(offset)
-    offset_param = f"${param_idx}"
-    param_idx += 1
-
-    # Query using UnifiedContentEmbedding for search, StoreAgent for metadata
-    sql_query = f"""
-        WITH candidates AS (
-            -- Lexical matches via UnifiedContentEmbedding.search
-            SELECT uce."contentId" as "storeListingVersionId"
-            FROM {{schema_prefix}}"UnifiedContentEmbedding" uce
-            INNER JOIN {{schema_prefix}}"StoreAgent" sa
-                ON uce."contentId" = sa."storeListingVersionId"
-            WHERE uce."contentType" = 'STORE_AGENT'::{{schema_prefix}}"ContentType"
-            AND uce."userId" IS NULL
-            AND uce.search @@ plainto_tsquery('english', {query_param})
-            AND {where_clause}
-
-            UNION
-
-            -- Semantic matches via UnifiedContentEmbedding.embedding
-            SELECT uce."contentId" as "storeListingVersionId"
-            FROM (
-                SELECT uce."contentId", uce.embedding
-                FROM {{schema_prefix}}"UnifiedContentEmbedding" uce
-                INNER JOIN {{schema_prefix}}"StoreAgent" sa
-                    ON uce."contentId" = sa."storeListingVersionId"
-                WHERE uce."contentType" = 'STORE_AGENT'::{{schema_prefix}}"ContentType"
-                AND uce."userId" IS NULL
-                AND {where_clause}
-                ORDER BY uce.embedding <=> {embedding_param}::vector
-                LIMIT 200
-            ) uce
-        ),
-        search_scores AS (
-            SELECT
-                sa.slug,
-                sa.agent_name,
-                sa.agent_image,
-                sa.creator_username,
-                sa.creator_avatar,
-                sa.sub_heading,
-                sa.description,
-                sa.runs,
-                sa.rating,
-                sa.categories,
-                sa.featured,
-                sa.is_available,
-                sa.updated_at,
-                -- Searchable text for BM25 reranking
-                COALESCE(sa.agent_name, '') || ' ' || COALESCE(sa.sub_heading, '') || ' ' || COALESCE(sa.description, '') as searchable_text,
-                -- Semantic score
-                COALESCE(1 - (uce.embedding <=> {embedding_param}::vector), 0) as semantic_score,
-                -- Lexical score (raw, will normalize)
-                COALESCE(ts_rank_cd(uce.search, plainto_tsquery('english', {query_param})), 0) as lexical_raw,
-                -- Category match
-                CASE
-                    WHEN EXISTS (
-                        SELECT 1 FROM unnest(sa.categories) cat
-                        WHERE LOWER(cat) LIKE '%' || {query_lower_param} || '%'
-                    )
-                    THEN 1.0
-                    ELSE 0.0
-                END as category_score,
-                -- Recency
-                GREATEST(0, 1 - EXTRACT(EPOCH FROM (NOW() - sa.updated_at)) / (90 * 24 * 3600)) as recency_score,
-                -- Popularity (raw)
-                sa.runs as popularity_raw
-            FROM candidates c
-            INNER JOIN {{schema_prefix}}"StoreAgent" sa
-                ON c."storeListingVersionId" = sa."storeListingVersionId"
-            INNER JOIN {{schema_prefix}}"UnifiedContentEmbedding" uce
-                ON sa."storeListingVersionId" = uce."contentId"
-                AND uce."contentType" = 'STORE_AGENT'::{{schema_prefix}}"ContentType"
-        ),
-        max_vals AS (
-            SELECT
-                GREATEST(MAX(lexical_raw), 0.001) as max_lexical,
-                GREATEST(MAX(popularity_raw), 1) as max_popularity
-            FROM search_scores
-        ),
-        normalized AS (
-            SELECT
-                ss.*,
-                ss.lexical_raw / mv.max_lexical as lexical_score,
-                CASE
-                    WHEN ss.popularity_raw > 0
-                    THEN LN(1 + ss.popularity_raw) / LN(1 + mv.max_popularity)
-                    ELSE 0
-                END as popularity_score
-            FROM search_scores ss
-            CROSS JOIN max_vals mv
-        ),
-        scored AS (
-            SELECT
-                slug,
-                agent_name,
-                agent_image,
-                creator_username,
-                creator_avatar,
-                sub_heading,
-                description,
-                runs,
-                rating,
-                categories,
-                featured,
-                is_available,
-                updated_at,
-                searchable_text,
-                semantic_score,
-                lexical_score,
-                category_score,
-                recency_score,
-                popularity_score,
-                (
-                    {w_semantic} * semantic_score +
-                    {w_lexical} * lexical_score +
-                    {w_category} * category_score +
-                    {w_recency} * recency_score +
-                    {w_popularity} * popularity_score
-                ) as combined_score
-            FROM normalized
-        ),
-        filtered AS (
-            SELECT *, COUNT(*) OVER () as total_count
-            FROM scored
-            WHERE combined_score >= {min_score_param}
-        )
-        SELECT * FROM filtered
-        ORDER BY combined_score DESC
-        LIMIT {limit_param} OFFSET {offset_param}
-    """
-
-    results = await query_raw_with_schema(
-        sql_query, *params, set_public_search_path=True
-    )
-
-    total = results[0]["total_count"] if results else 0
-
-    # Apply BM25 reranking
-    if results:
-        results = bm25_rerank(
-            query=query,
-            results=results,
-            text_field="searchable_text",
-            bm25_weight=0.3,
-            original_score_field="combined_score",
-        )
-
-    for result in results:
-        result.pop("total_count", None)
-        result.pop("searchable_text", None)
-
-    logger.info(f"Hybrid search (store agents): {len(results)} results, {total} total")
-
-    return results, total
-
-
-async def hybrid_search_simple(
-    query: str,
-    page: int = 1,
-    page_size: int = 20,
-) -> tuple[list[dict[str, Any]], int]:
-    """Simplified hybrid search for store agents."""
-    return await hybrid_search(query=query, page=page, page_size=page_size)
-
-
-# Backward compatibility alias - HybridSearchWeights maps to StoreAgentSearchWeights
-# for existing code that expects the popularity parameter
-HybridSearchWeights = StoreAgentSearchWeights

autogpt_platform/backend/backend/api/features/store/hybrid_search_test.py

@@ -1,726 +0,0 @@
-"""
-Integration tests for hybrid search with schema handling.
-
-These tests verify that hybrid search works correctly across different database schemas.
-"""
-
-from unittest.mock import patch
-
-import pytest
-from prisma.enums import ContentType
-
-from backend.api.features.store import embeddings
-from backend.api.features.store.hybrid_search import (
-    HybridSearchWeights,
-    UnifiedSearchWeights,
-    hybrid_search,
-    unified_hybrid_search,
-)
-
-
-@pytest.mark.asyncio(loop_scope="session")
-@pytest.mark.integration
-async def test_hybrid_search_with_schema_handling():
-    """Test that hybrid search correctly handles database schema prefixes."""
-    # Test with a mock query to ensure schema handling works
-    query = "test agent"
-
-    with patch(
-        "backend.api.features.store.hybrid_search.query_raw_with_schema"
-    ) as mock_query:
-        # Mock the query result
-        mock_query.return_value = [
-            {
-                "slug": "test/agent",
-                "agent_name": "Test Agent",
-                "agent_image": "test.png",
-                "creator_username": "test",
-                "creator_avatar": "avatar.png",
-                "sub_heading": "Test sub-heading",
-                "description": "Test description",
-                "runs": 10,
-                "rating": 4.5,
-                "categories": ["test"],
-                "featured": False,
-                "is_available": True,
-                "updated_at": "2024-01-01T00:00:00Z",
-                "combined_score": 0.8,
-                "semantic_score": 0.7,
-                "lexical_score": 0.6,
-                "category_score": 0.5,
-                "recency_score": 0.4,
-                "total_count": 1,
-            }
-        ]
-
-        with patch(
-            "backend.api.features.store.hybrid_search.embed_query"
-        ) as mock_embed:
-            mock_embed.return_value = [0.1] * embeddings.EMBEDDING_DIM  # Mock embedding
-
-            results, total = await hybrid_search(
-                query=query,
-                page=1,
-                page_size=20,
-            )
-
-            # Verify the query was called
-            assert mock_query.called
-            # Verify the SQL template uses schema_prefix placeholder
-            call_args = mock_query.call_args
-            sql_template = call_args[0][0]
-            assert "{schema_prefix}" in sql_template
-
-            # Verify results
-            assert len(results) == 1
-            assert total == 1
-            assert results[0]["slug"] == "test/agent"
-
-
-@pytest.mark.asyncio(loop_scope="session")
-@pytest.mark.integration
-async def test_hybrid_search_with_public_schema():
-    """Test hybrid search when using public schema (no prefix needed)."""
-    with patch("backend.data.db.get_database_schema") as mock_schema:
-        mock_schema.return_value = "public"
-
-        with patch(
-            "backend.api.features.store.hybrid_search.query_raw_with_schema"
-        ) as mock_query:
-            mock_query.return_value = []
-
-            with patch(
-                "backend.api.features.store.hybrid_search.embed_query"
-            ) as mock_embed:
-                mock_embed.return_value = [0.1] * embeddings.EMBEDDING_DIM
-
-                results, total = await hybrid_search(
-                    query="test",
-                    page=1,
-                    page_size=20,
-                )
-
-                # Verify the mock was set up correctly
-                assert mock_schema.return_value == "public"
-
-                # Results should work even with empty results
-                assert results == []
-                assert total == 0
-
-
-@pytest.mark.asyncio(loop_scope="session")
-@pytest.mark.integration
-async def test_hybrid_search_with_custom_schema():
-    """Test hybrid search when using custom schema (e.g., 'platform')."""
-    with patch("backend.data.db.get_database_schema") as mock_schema:
-        mock_schema.return_value = "platform"
-
-        with patch(
-            "backend.api.features.store.hybrid_search.query_raw_with_schema"
-        ) as mock_query:
-            mock_query.return_value = []
-
-            with patch(
-                "backend.api.features.store.hybrid_search.embed_query"
-            ) as mock_embed:
-                mock_embed.return_value = [0.1] * embeddings.EMBEDDING_DIM
-
-                results, total = await hybrid_search(
-                    query="test",
-                    page=1,
-                    page_size=20,
-                )
-
-                # Verify the mock was set up correctly
-                assert mock_schema.return_value == "platform"
-
-                assert results == []
-                assert total == 0
-
-
-@pytest.mark.asyncio(loop_scope="session")
-@pytest.mark.integration
-async def test_hybrid_search_without_embeddings():
-    """Test hybrid search gracefully degrades when embeddings are unavailable."""
-    # Mock database to return some results
-    mock_results = [
-        {
-            "slug": "test-agent",
-            "agent_name": "Test Agent",
-            "agent_image": "test.png",
-            "creator_username": "creator",
-            "creator_avatar": "avatar.png",
-            "sub_heading": "Test heading",
-            "description": "Test description",
-            "runs": 100,
-            "rating": 4.5,
-            "categories": ["AI"],
-            "featured": False,
-            "is_available": True,
-            "updated_at": "2025-01-01T00:00:00Z",
-            "semantic_score": 0.0,  # Zero because no embedding
-            "lexical_score": 0.5,
-            "category_score": 0.0,
-            "recency_score": 0.1,
-            "popularity_score": 0.2,
-            "combined_score": 0.3,
-            "total_count": 1,
-        }
-    ]
-
-    with patch("backend.api.features.store.hybrid_search.embed_query") as mock_embed:
-        with patch(
-            "backend.api.features.store.hybrid_search.query_raw_with_schema"
-        ) as mock_query:
-            # Simulate embedding failure
-            mock_embed.return_value = None
-            mock_query.return_value = mock_results
-
-            # Should NOT raise - graceful degradation
-            results, total = await hybrid_search(
-                query="test",
-                page=1,
-                page_size=20,
-            )
-
-            # Verify it returns results even without embeddings
-            assert len(results) == 1
-            assert results[0]["slug"] == "test-agent"
-            assert total == 1
-
-
-@pytest.mark.asyncio(loop_scope="session")
-@pytest.mark.integration
-async def test_hybrid_search_with_filters():
-    """Test hybrid search with various filters."""
-    with patch(
-        "backend.api.features.store.hybrid_search.query_raw_with_schema"
-    ) as mock_query:
-        mock_query.return_value = []
-
-        with patch(
-            "backend.api.features.store.hybrid_search.embed_query"
-        ) as mock_embed:
-            mock_embed.return_value = [0.1] * embeddings.EMBEDDING_DIM
-
-            # Test with featured filter
-            results, total = await hybrid_search(
-                query="test",
-                featured=True,
-                creators=["user1", "user2"],
-                category="productivity",
-                page=1,
-                page_size=10,
-            )
-
-            # Verify filters were applied in the query
-            call_args = mock_query.call_args
-            params = call_args[0][1:]  # Skip SQL template
-
-            # Should have query, query_lower, creators array, category
-            assert len(params) >= 4
-
-
-@pytest.mark.asyncio(loop_scope="session")
-@pytest.mark.integration
-async def test_hybrid_search_weights():
-    """Test hybrid search with custom weights."""
-    custom_weights = HybridSearchWeights(
-        semantic=0.5,
-        lexical=0.3,
-        category=0.1,
-        recency=0.1,
-        popularity=0.0,
-    )
-
-    with patch(
-        "backend.api.features.store.hybrid_search.query_raw_with_schema"
-    ) as mock_query:
-        mock_query.return_value = []
-
-        with patch(
-            "backend.api.features.store.hybrid_search.embed_query"
-        ) as mock_embed:
-            mock_embed.return_value = [0.1] * embeddings.EMBEDDING_DIM
-
-            results, total = await hybrid_search(
-                query="test",
-                weights=custom_weights,
-                page=1,
-                page_size=20,
-            )
-
-            # Verify custom weights were used in the query
-            call_args = mock_query.call_args
-            sql_template = call_args[0][0]
-            params = call_args[0][1:]  # Get all parameters passed
-
-            # Check that SQL uses parameterized weights (not f-string interpolation)
-            assert "$" in sql_template  # Verify parameterization is used
-
-            # Check that custom weights are in the params
-            assert 0.5 in params  # semantic weight
-            assert 0.3 in params  # lexical weight
-            assert 0.1 in params  # category and recency weights
-
-
-@pytest.mark.asyncio(loop_scope="session")
-@pytest.mark.integration
-async def test_hybrid_search_min_score_filtering():
-    """Test hybrid search minimum score threshold."""
-    with patch(
-        "backend.api.features.store.hybrid_search.query_raw_with_schema"
-    ) as mock_query:
-        # Return results with varying scores
-        mock_query.return_value = [
-            {
-                "slug": "high-score/agent",
-                "agent_name": "High Score Agent",
-                "combined_score": 0.8,
-                "total_count": 1,
-                # ... other fields
-            }
-        ]
-
-        with patch(
-            "backend.api.features.store.hybrid_search.embed_query"
-        ) as mock_embed:
-            mock_embed.return_value = [0.1] * embeddings.EMBEDDING_DIM
-
-            # Test with custom min_score
-            results, total = await hybrid_search(
-                query="test",
-                min_score=0.5,  # High threshold
-                page=1,
-                page_size=20,
-            )
-
-            # Verify min_score was applied in query
-            call_args = mock_query.call_args
-            sql_template = call_args[0][0]
-            params = call_args[0][1:]  # Get all parameters
-
-            # Check that SQL uses parameterized min_score
-            assert "combined_score >=" in sql_template
-            assert "$" in sql_template  # Verify parameterization
-
-            # Check that custom min_score is in the params
-            assert 0.5 in params
-
-
-@pytest.mark.asyncio(loop_scope="session")
-@pytest.mark.integration
-async def test_hybrid_search_pagination():
-    """Test hybrid search pagination.
-
-    Pagination happens in SQL (LIMIT/OFFSET), then BM25 reranking is applied
-    to the paginated results.
-    """
-    # Create mock results that SQL would return for a page
-    mock_results = [
-        {
-            "slug": f"agent-{i}",
-            "agent_name": f"Agent {i}",
-            "agent_image": "test.png",
-            "creator_username": "test",
-            "creator_avatar": "avatar.png",
-            "sub_heading": "Test",
-            "description": "Test description",
-            "runs": 100 - i,
-            "rating": 4.5,
-            "categories": ["test"],
-            "featured": False,
-            "is_available": True,
-            "updated_at": "2024-01-01T00:00:00Z",
-            "searchable_text": f"Agent {i} test description",
-            "combined_score": 0.9 - (i * 0.01),
-            "semantic_score": 0.7,
-            "lexical_score": 0.6,
-            "category_score": 0.5,
-            "recency_score": 0.4,
-            "popularity_score": 0.3,
-            "total_count": 25,
-        }
-        for i in range(10)  # SQL returns page_size results
-    ]
-
-    with patch(
-        "backend.api.features.store.hybrid_search.query_raw_with_schema"
-    ) as mock_query:
-        mock_query.return_value = mock_results
-
-        with patch(
-            "backend.api.features.store.hybrid_search.embed_query"
-        ) as mock_embed:
-            mock_embed.return_value = [0.1] * embeddings.EMBEDDING_DIM
-
-            # Test page 2 with page_size 10
-            results, total = await hybrid_search(
-                query="test",
-                page=2,
-                page_size=10,
-            )
-
-            # Verify results returned
-            assert len(results) == 10
-            assert total == 25  # Total from SQL COUNT(*) OVER()
-
-            # Verify the SQL query uses page_size and offset
-            call_args = mock_query.call_args
-            params = call_args[0]
-            # Last two params are page_size and offset
-            page_size_param = params[-2]
-            offset_param = params[-1]
-            assert page_size_param == 10
-            assert offset_param == 10  # (page 2 - 1) * 10
-
-
-@pytest.mark.asyncio(loop_scope="session")
-@pytest.mark.integration
-async def test_hybrid_search_error_handling():
-    """Test hybrid search error handling."""
-    with patch(
-        "backend.api.features.store.hybrid_search.query_raw_with_schema"
-    ) as mock_query:
-        # Simulate database error
-        mock_query.side_effect = Exception("Database connection error")
-
-        with patch(
-            "backend.api.features.store.hybrid_search.embed_query"
-        ) as mock_embed:
-            mock_embed.return_value = [0.1] * embeddings.EMBEDDING_DIM
-
-            # Should raise exception
-            with pytest.raises(Exception) as exc_info:
-                await hybrid_search(
-                    query="test",
-                    page=1,
-                    page_size=20,
-                )
-
-            assert "Database connection error" in str(exc_info.value)
-
-
-# =============================================================================
-# Unified Hybrid Search Tests
-# =============================================================================
-
-
-@pytest.mark.asyncio(loop_scope="session")
-@pytest.mark.integration
-async def test_unified_hybrid_search_basic():
-    """Test basic unified hybrid search across all content types."""
-    mock_results = [
-        {
-            "content_type": "STORE_AGENT",
-            "content_id": "agent-1",
-            "searchable_text": "Test Agent Description",
-            "metadata": {"name": "Test Agent"},
-            "updated_at": "2025-01-01T00:00:00Z",
-            "semantic_score": 0.7,
-            "lexical_score": 0.8,
-            "category_score": 0.5,
-            "recency_score": 0.3,
-            "combined_score": 0.6,
-            "total_count": 2,
-        },
-        {
-            "content_type": "BLOCK",
-            "content_id": "block-1",
-            "searchable_text": "Test Block Description",
-            "metadata": {"name": "Test Block"},
-            "updated_at": "2025-01-01T00:00:00Z",
-            "semantic_score": 0.6,
-            "lexical_score": 0.7,
-            "category_score": 0.4,
-            "recency_score": 0.2,
-            "combined_score": 0.5,
-            "total_count": 2,
-        },
-    ]
-
-    with patch(
-        "backend.api.features.store.hybrid_search.query_raw_with_schema"
-    ) as mock_query:
-        with patch(
-            "backend.api.features.store.hybrid_search.embed_query"
-        ) as mock_embed:
-            mock_query.return_value = mock_results
-            mock_embed.return_value = [0.1] * embeddings.EMBEDDING_DIM
-
-            results, total = await unified_hybrid_search(
-                query="test",
-                page=1,
-                page_size=20,
-            )
-
-            assert len(results) == 2
-            assert total == 2
-            assert results[0]["content_type"] == "STORE_AGENT"
-            assert results[1]["content_type"] == "BLOCK"
-
-
-@pytest.mark.asyncio(loop_scope="session")
-@pytest.mark.integration
-async def test_unified_hybrid_search_filter_by_content_type():
-    """Test unified search filtering by specific content types."""
-    mock_results = [
-        {
-            "content_type": "BLOCK",
-            "content_id": "block-1",
-            "searchable_text": "Test Block",
-            "metadata": {},
-            "updated_at": "2025-01-01T00:00:00Z",
-            "semantic_score": 0.7,
-            "lexical_score": 0.8,
-            "category_score": 0.0,
-            "recency_score": 0.3,
-            "combined_score": 0.5,
-            "total_count": 1,
-        },
-    ]
-
-    with patch(
-        "backend.api.features.store.hybrid_search.query_raw_with_schema"
-    ) as mock_query:
-        with patch(
-            "backend.api.features.store.hybrid_search.embed_query"
-        ) as mock_embed:
-            mock_query.return_value = mock_results
-            mock_embed.return_value = [0.1] * embeddings.EMBEDDING_DIM
-
-            results, total = await unified_hybrid_search(
-                query="test",
-                content_types=[ContentType.BLOCK],
-                page=1,
-                page_size=20,
-            )
-
-            # Verify content_types parameter was passed correctly
-            call_args = mock_query.call_args
-            params = call_args[0][1:]
-            # The content types should be in the params as a list
-            assert ["BLOCK"] in params
-
-            assert len(results) == 1
-            assert total == 1
-
-
-@pytest.mark.asyncio(loop_scope="session")
-@pytest.mark.integration
-async def test_unified_hybrid_search_with_user_id():
-    """Test unified search with user_id for private content."""
-    mock_results = [
-        {
-            "content_type": "STORE_AGENT",
-            "content_id": "agent-1",
-            "searchable_text": "My Private Agent",
-            "metadata": {},
-            "updated_at": "2025-01-01T00:00:00Z",
-            "semantic_score": 0.7,
-            "lexical_score": 0.8,
-            "category_score": 0.0,
-            "recency_score": 0.3,
-            "combined_score": 0.6,
-            "total_count": 1,
-        },
-    ]
-
-    with patch(
-        "backend.api.features.store.hybrid_search.query_raw_with_schema"
-    ) as mock_query:
-        with patch(
-            "backend.api.features.store.hybrid_search.embed_query"
-        ) as mock_embed:
-            mock_query.return_value = mock_results
-            mock_embed.return_value = [0.1] * embeddings.EMBEDDING_DIM
-
-            results, total = await unified_hybrid_search(
-                query="test",
-                user_id="user-123",
-                page=1,
-                page_size=20,
-            )
-
-            # Verify SQL contains user_id filter
-            call_args = mock_query.call_args
-            sql_template = call_args[0][0]
-            params = call_args[0][1:]
-
-            assert 'uce."userId"' in sql_template
-            assert "user-123" in params
-
-
-@pytest.mark.asyncio(loop_scope="session")
-@pytest.mark.integration
-async def test_unified_hybrid_search_custom_weights():
-    """Test unified search with custom weights."""
-    custom_weights = UnifiedSearchWeights(
-        semantic=0.6,
-        lexical=0.2,
-        category=0.1,
-        recency=0.1,
-    )
-
-    with patch(
-        "backend.api.features.store.hybrid_search.query_raw_with_schema"
-    ) as mock_query:
-        with patch(
-            "backend.api.features.store.hybrid_search.embed_query"
-        ) as mock_embed:
-            mock_query.return_value = []
-            mock_embed.return_value = [0.1] * embeddings.EMBEDDING_DIM
-
-            results, total = await unified_hybrid_search(
-                query="test",
-                weights=custom_weights,
-                page=1,
-                page_size=20,
-            )
-
-            # Verify custom weights are in parameters
-            call_args = mock_query.call_args
-            params = call_args[0][1:]
-
-            assert 0.6 in params  # semantic weight
-            assert 0.2 in params  # lexical weight
-
-
-@pytest.mark.asyncio(loop_scope="session")
-@pytest.mark.integration
-async def test_unified_hybrid_search_graceful_degradation():
-    """Test unified search gracefully degrades when embeddings unavailable."""
-    mock_results = [
-        {
-            "content_type": "DOCUMENTATION",
-            "content_id": "doc-1",
-            "searchable_text": "API Documentation",
-            "metadata": {},
-            "updated_at": "2025-01-01T00:00:00Z",
-            "semantic_score": 0.0,  # Zero because no embedding
-            "lexical_score": 0.8,
-            "category_score": 0.0,
-            "recency_score": 0.2,
-            "combined_score": 0.5,
-            "total_count": 1,
-        },
-    ]
-
-    with patch(
-        "backend.api.features.store.hybrid_search.query_raw_with_schema"
-    ) as mock_query:
-        with patch(
-            "backend.api.features.store.hybrid_search.embed_query"
-        ) as mock_embed:
-            mock_query.return_value = mock_results
-            mock_embed.return_value = None  # Embedding failure
-
-            # Should NOT raise - graceful degradation
-            results, total = await unified_hybrid_search(
-                query="test",
-                page=1,
-                page_size=20,
-            )
-
-            assert len(results) == 1
-            assert total == 1
-
-
-@pytest.mark.asyncio(loop_scope="session")
-@pytest.mark.integration
-async def test_unified_hybrid_search_empty_query():
-    """Test unified search with empty query returns empty results."""
-    results, total = await unified_hybrid_search(
-        query="",
-        page=1,
-        page_size=20,
-    )
-
-    assert results == []
-    assert total == 0
-
-
-@pytest.mark.asyncio(loop_scope="session")
-@pytest.mark.integration
-async def test_unified_hybrid_search_pagination():
-    """Test unified search pagination with BM25 reranking.
-
-    Pagination happens in SQL (LIMIT/OFFSET), then BM25 reranking is applied
-    to the paginated results.
-    """
-    # Create mock results that SQL would return for a page
-    mock_results = [
-        {
-            "content_type": "STORE_AGENT",
-            "content_id": f"agent-{i}",
-            "searchable_text": f"Agent {i} description",
-            "metadata": {"name": f"Agent {i}"},
-            "updated_at": "2025-01-01T00:00:00Z",
-            "semantic_score": 0.7,
-            "lexical_score": 0.8 - (i * 0.01),
-            "category_score": 0.5,
-            "recency_score": 0.3,
-            "combined_score": 0.6 - (i * 0.01),
-            "total_count": 50,
-        }
-        for i in range(15)  # SQL returns page_size results
-    ]
-
-    with patch(
-        "backend.api.features.store.hybrid_search.query_raw_with_schema"
-    ) as mock_query:
-        with patch(
-            "backend.api.features.store.hybrid_search.embed_query"
-        ) as mock_embed:
-            mock_query.return_value = mock_results
-            mock_embed.return_value = [0.1] * embeddings.EMBEDDING_DIM
-
-            results, total = await unified_hybrid_search(
-                query="test",
-                page=3,
-                page_size=15,
-            )
-
-            # Verify results returned
-            assert len(results) == 15
-            assert total == 50  # Total from SQL COUNT(*) OVER()
-
-            # Verify the SQL query uses page_size and offset
-            call_args = mock_query.call_args
-            params = call_args[0]
-            # Last two params are page_size and offset
-            page_size_param = params[-2]
-            offset_param = params[-1]
-            assert page_size_param == 15
-            assert offset_param == 30  # (page 3 - 1) * 15
-
-
-@pytest.mark.asyncio(loop_scope="session")
-@pytest.mark.integration
-async def test_unified_hybrid_search_schema_prefix():
-    """Test unified search uses schema_prefix placeholder."""
-    with patch(
-        "backend.api.features.store.hybrid_search.query_raw_with_schema"
-    ) as mock_query:
-        with patch(
-            "backend.api.features.store.hybrid_search.embed_query"
-        ) as mock_embed:
-            mock_query.return_value = []
-            mock_embed.return_value = [0.1] * embeddings.EMBEDDING_DIM
-
-            await unified_hybrid_search(
-                query="test",
-                page=1,
-                page_size=20,
-            )
-
-            call_args = mock_query.call_args
-            sql_template = call_args[0][0]
-
-            # Verify schema_prefix placeholder is used for table references
-            assert "{schema_prefix}" in sql_template
-            assert '"UnifiedContentEmbedding"' in sql_template
-
-
-if __name__ == "__main__":
-    pytest.main([__file__, "-v", "-s"])

autogpt_platform/backend/backend/api/features/store/semantic_search_test.py

@@ -1,272 +0,0 @@
-"""Tests for the semantic_search function."""
-
-import pytest
-from prisma.enums import ContentType
-
-from backend.api.features.store.embeddings import EMBEDDING_DIM, semantic_search
-
-
-@pytest.mark.asyncio
-async def test_search_blocks_only(mocker):
-    """Test searching only BLOCK content type."""
-    # Mock embed_query to return a test embedding
-    mock_embedding = [0.1] * EMBEDDING_DIM
-    mocker.patch(
-        "backend.api.features.store.embeddings.embed_query",
-        return_value=mock_embedding,
-    )
-
-    # Mock query_raw_with_schema to return test results
-    mock_results = [
-        {
-            "content_id": "block-123",
-            "content_type": "BLOCK",
-            "searchable_text": "Calculator Block - Performs arithmetic operations",
-            "metadata": {"name": "Calculator", "categories": ["Math"]},
-            "similarity": 0.85,
-        }
-    ]
-    mocker.patch(
-        "backend.api.features.store.embeddings.query_raw_with_schema",
-        return_value=mock_results,
-    )
-
-    results = await semantic_search(
-        query="calculate numbers",
-        content_types=[ContentType.BLOCK],
-    )
-
-    assert len(results) == 1
-    assert results[0]["content_type"] == "BLOCK"
-    assert results[0]["content_id"] == "block-123"
-    assert results[0]["similarity"] == 0.85
-
-
-@pytest.mark.asyncio
-async def test_search_multiple_content_types(mocker):
-    """Test searching multiple content types simultaneously."""
-    mock_embedding = [0.1] * EMBEDDING_DIM
-    mocker.patch(
-        "backend.api.features.store.embeddings.embed_query",
-        return_value=mock_embedding,
-    )
-
-    mock_results = [
-        {
-            "content_id": "block-123",
-            "content_type": "BLOCK",
-            "searchable_text": "Calculator Block",
-            "metadata": {},
-            "similarity": 0.85,
-        },
-        {
-            "content_id": "doc-456",
-            "content_type": "DOCUMENTATION",
-            "searchable_text": "How to use Calculator",
-            "metadata": {},
-            "similarity": 0.75,
-        },
-    ]
-    mocker.patch(
-        "backend.api.features.store.embeddings.query_raw_with_schema",
-        return_value=mock_results,
-    )
-
-    results = await semantic_search(
-        query="calculator",
-        content_types=[ContentType.BLOCK, ContentType.DOCUMENTATION],
-    )
-
-    assert len(results) == 2
-    assert results[0]["content_type"] == "BLOCK"
-    assert results[1]["content_type"] == "DOCUMENTATION"
-
-
-@pytest.mark.asyncio
-async def test_search_with_min_similarity_threshold(mocker):
-    """Test that results below min_similarity are filtered out."""
-    mock_embedding = [0.1] * EMBEDDING_DIM
-    mocker.patch(
-        "backend.api.features.store.embeddings.embed_query",
-        return_value=mock_embedding,
-    )
-
-    # Only return results above 0.7 similarity
-    mock_results = [
-        {
-            "content_id": "block-123",
-            "content_type": "BLOCK",
-            "searchable_text": "Calculator Block",
-            "metadata": {},
-            "similarity": 0.85,
-        }
-    ]
-    mocker.patch(
-        "backend.api.features.store.embeddings.query_raw_with_schema",
-        return_value=mock_results,
-    )
-
-    results = await semantic_search(
-        query="calculate",
-        content_types=[ContentType.BLOCK],
-        min_similarity=0.7,
-    )
-
-    assert len(results) == 1
-    assert results[0]["similarity"] >= 0.7
-
-
-@pytest.mark.asyncio
-async def test_search_fallback_to_lexical(mocker):
-    """Test fallback to lexical search when embeddings fail."""
-    # Mock embed_query to return None (embeddings unavailable)
-    mocker.patch(
-        "backend.api.features.store.embeddings.embed_query",
-        return_value=None,
-    )
-
-    mock_lexical_results = [
-        {
-            "content_id": "block-123",
-            "content_type": "BLOCK",
-            "searchable_text": "Calculator Block performs calculations",
-            "metadata": {},
-            "similarity": 0.0,
-        }
-    ]
-    mocker.patch(
-        "backend.api.features.store.embeddings.query_raw_with_schema",
-        return_value=mock_lexical_results,
-    )
-
-    results = await semantic_search(
-        query="calculator",
-        content_types=[ContentType.BLOCK],
-    )
-
-    assert len(results) == 1
-    assert results[0]["similarity"] == 0.0  # Lexical search returns 0 similarity
-
-
-@pytest.mark.asyncio
-async def test_search_empty_query():
-    """Test that empty query returns no results."""
-    results = await semantic_search(query="")
-    assert results == []
-
-    results = await semantic_search(query="   ")
-    assert results == []
-
-
-@pytest.mark.asyncio
-async def test_search_with_user_id_filter(mocker):
-    """Test searching with user_id filter for private content."""
-    mock_embedding = [0.1] * EMBEDDING_DIM
-    mocker.patch(
-        "backend.api.features.store.embeddings.embed_query",
-        return_value=mock_embedding,
-    )
-
-    mock_results = [
-        {
-            "content_id": "agent-789",
-            "content_type": "LIBRARY_AGENT",
-            "searchable_text": "My Custom Agent",
-            "metadata": {},
-            "similarity": 0.9,
-        }
-    ]
-    mocker.patch(
-        "backend.api.features.store.embeddings.query_raw_with_schema",
-        return_value=mock_results,
-    )
-
-    results = await semantic_search(
-        query="custom agent",
-        content_types=[ContentType.LIBRARY_AGENT],
-        user_id="user-123",
-    )
-
-    assert len(results) == 1
-    assert results[0]["content_type"] == "LIBRARY_AGENT"
-
-
-@pytest.mark.asyncio
-async def test_search_limit_parameter(mocker):
-    """Test that limit parameter correctly limits results."""
-    mock_embedding = [0.1] * EMBEDDING_DIM
-    mocker.patch(
-        "backend.api.features.store.embeddings.embed_query",
-        return_value=mock_embedding,
-    )
-
-    # Return 5 results
-    mock_results = [
-        {
-            "content_id": f"block-{i}",
-            "content_type": "BLOCK",
-            "searchable_text": f"Block {i}",
-            "metadata": {},
-            "similarity": 0.8,
-        }
-        for i in range(5)
-    ]
-    mocker.patch(
-        "backend.api.features.store.embeddings.query_raw_with_schema",
-        return_value=mock_results,
-    )
-
-    results = await semantic_search(
-        query="block",
-        content_types=[ContentType.BLOCK],
-        limit=5,
-    )
-
-    assert len(results) == 5
-
-
-@pytest.mark.asyncio
-async def test_search_default_content_types(mocker):
-    """Test that default content_types includes BLOCK, STORE_AGENT, and DOCUMENTATION."""
-    mock_embedding = [0.1] * EMBEDDING_DIM
-    mocker.patch(
-        "backend.api.features.store.embeddings.embed_query",
-        return_value=mock_embedding,
-    )
-
-    mock_query_raw = mocker.patch(
-        "backend.api.features.store.embeddings.query_raw_with_schema",
-        return_value=[],
-    )
-
-    await semantic_search(query="test")
-
-    # Check that the SQL query includes all three default content types
-    call_args = mock_query_raw.call_args
-    assert "BLOCK" in str(call_args)
-    assert "STORE_AGENT" in str(call_args)
-    assert "DOCUMENTATION" in str(call_args)
-
-
-@pytest.mark.asyncio
-async def test_search_handles_database_error(mocker):
-    """Test that database errors are handled gracefully."""
-    mock_embedding = [0.1] * EMBEDDING_DIM
-    mocker.patch(
-        "backend.api.features.store.embeddings.embed_query",
-        return_value=mock_embedding,
-    )
-
-    # Simulate database error
-    mocker.patch(
-        "backend.api.features.store.embeddings.query_raw_with_schema",
-        side_effect=Exception("Database connection failed"),
-    )
-
-    results = await semantic_search(
-        query="test",
-        content_types=[ContentType.BLOCK],
-    )
-
-    # Should return empty list on error
-    assert results == []

autogpt_platform/backend/backend/api/utils/openapi.py

@@ -1,41 +0,0 @@
-from fastapi import FastAPI
-
-
-def sort_openapi(app: FastAPI) -> None:
-    """
-    Patch a FastAPI instance's `openapi()` method to sort the endpoints,
-    schemas, and responses.
-    """
-    wrapped_openapi = app.openapi
-
-    def custom_openapi():
-        if app.openapi_schema:
-            return app.openapi_schema
-
-        openapi_schema = wrapped_openapi()
-
-        # Sort endpoints
-        openapi_schema["paths"] = dict(sorted(openapi_schema["paths"].items()))
-
-        # Sort endpoints -> methods
-        for p in openapi_schema["paths"].keys():
-            openapi_schema["paths"][p] = dict(
-                sorted(openapi_schema["paths"][p].items())
-            )
-
-            # Sort endpoints -> methods -> responses
-            for m in openapi_schema["paths"][p].keys():
-                openapi_schema["paths"][p][m]["responses"] = dict(
-                    sorted(openapi_schema["paths"][p][m]["responses"].items())
-                )
-
-        # Sort schemas and responses as well
-        for k in openapi_schema["components"].keys():
-            openapi_schema["components"][k] = dict(
-                sorted(openapi_schema["components"][k].items())
-            )
-
-        app.openapi_schema = openapi_schema
-        return openapi_schema
-
-    app.openapi = custom_openapi

autogpt_platform/backend/backend/app.py

@@ -36,10 +36,10 @@ def main(**kwargs):
     Run all the processes required for the AutoGPT-server (REST and WebSocket APIs).
     """
 
-    from backend.api.rest_api import AgentServer
-    from backend.api.ws_api import WebsocketServer
     from backend.executor import DatabaseManager, ExecutionManager, Scheduler
     from backend.notifications import NotificationManager
+    from backend.server.rest_api import AgentServer
+    from backend.server.ws_api import WebsocketServer
 
     run_processes(
         DatabaseManager().set_log_level("warning"),

autogpt_platform/backend/backend/blocks/ai_condition.py

@@ -1,7 +1,6 @@
 from typing import Any
 
 from backend.blocks.llm import (
-    DEFAULT_LLM_MODEL,
     TEST_CREDENTIALS,
     TEST_CREDENTIALS_INPUT,
     AIBlockBase,
@@ -50,7 +49,7 @@ class AIConditionBlock(AIBlockBase):
         )
         model: LlmModel = SchemaField(
             title="LLM Model",
-            default=DEFAULT_LLM_MODEL,
+            default=LlmModel.GPT4O,
             description="The language model to use for evaluating the condition.",
             advanced=False,
         )
@@ -82,7 +81,7 @@ class AIConditionBlock(AIBlockBase):
                 "condition": "the input is an email address",
                 "yes_value": "Valid email",
                 "no_value": "Not an email",
-                "model": DEFAULT_LLM_MODEL,
+                "model": LlmModel.GPT4O,
                 "credentials": TEST_CREDENTIALS_INPUT,
             },
             test_credentials=TEST_CREDENTIALS,

autogpt_platform/backend/backend/blocks/ai_shortform_video_block.py

@@ -20,7 +20,6 @@ from backend.data.model import (
     SchemaField,
 )
 from backend.integrations.providers import ProviderName
-from backend.util.exceptions import BlockExecutionError
 from backend.util.request import Requests
 
 TEST_CREDENTIALS = APIKeyCredentials(
@@ -247,11 +246,7 @@ class AIShortformVideoCreatorBlock(Block):
             await asyncio.sleep(10)
 
         logger.error("Video creation timed out")
-        raise BlockExecutionError(
-            message="Video creation timed out",
-            block_name=self.name,
-            block_id=self.id,
-        )
+        raise TimeoutError("Video creation timed out")
 
     def __init__(self):
         super().__init__(
@@ -427,11 +422,7 @@ class AIAdMakerVideoCreatorBlock(Block):
             await asyncio.sleep(10)
 
         logger.error("Video creation timed out")
-        raise BlockExecutionError(
-            message="Video creation timed out",
-            block_name=self.name,
-            block_id=self.id,
-        )
+        raise TimeoutError("Video creation timed out")
 
     def __init__(self):
         super().__init__(
@@ -608,11 +599,7 @@ class AIScreenshotToVideoAdBlock(Block):
             await asyncio.sleep(10)
 
         logger.error("Video creation timed out")
-        raise BlockExecutionError(
-            message="Video creation timed out",
-            block_name=self.name,
-            block_id=self.id,
-        )
+        raise TimeoutError("Video creation timed out")
 
     def __init__(self):
         super().__init__(

autogpt_platform/backend/backend/blocks/airtable/_webhook.py

@@ -6,9 +6,6 @@ import hashlib
 import hmac
 import logging
 from enum import Enum
-from typing import cast
-
-from prisma.types import Serializable
 
 from backend.sdk import (
     BaseWebhooksManager,
@@ -87,9 +84,7 @@ class AirtableWebhookManager(BaseWebhooksManager):
         # update webhook config
         await update_webhook(
             webhook.id,
-            config=cast(
-                dict[str, Serializable], {"base_id": base_id, "cursor": response.cursor}
-            ),
+            config={"base_id": base_id, "cursor": response.cursor},
         )
 
         event_type = "notification"

autogpt_platform/backend/backend/blocks/branching.py

@@ -106,10 +106,7 @@ class ConditionBlock(Block):
             ComparisonOperator.LESS_THAN_OR_EQUAL: lambda a, b: a <= b,
         }
 
-        try:
-            result = comparison_funcs[operator](value1, value2)
-        except Exception as e:
-            raise ValueError(f"Comparison failed: {e}") from e
+        result = comparison_funcs[operator](value1, value2)
 
         yield "result", result
 

autogpt_platform/backend/backend/blocks/dataforseo/related_keywords.py

@@ -182,10 +182,13 @@ class DataForSeoRelatedKeywordsBlock(Block):
             if results and len(results) > 0:
                 # results is a list, get the first element
                 first_result = results[0] if isinstance(results, list) else results
-                # Handle missing key, null value, or valid list value
-                if isinstance(first_result, dict):
-                    items = first_result.get("items") or []
-                else:
+                items = (
+                    first_result.get("items", [])
+                    if isinstance(first_result, dict)
+                    else []
+                )
+                # Ensure items is never None
+                if items is None:
                     items = []
                 for item in items:
                     # Extract keyword_data from the item

autogpt_platform/backend/backend/blocks/firecrawl/extract.py

@@ -15,7 +15,6 @@ from backend.sdk import (
     SchemaField,
     cost,
 )
-from backend.util.exceptions import BlockExecutionError
 
 from ._config import firecrawl
 
@@ -60,18 +59,11 @@ class FirecrawlExtractBlock(Block):
     ) -> BlockOutput:
         app = FirecrawlApp(api_key=credentials.api_key.get_secret_value())
 
-        try:
-            extract_result = app.extract(
-                urls=input_data.urls,
-                prompt=input_data.prompt,
-                schema=input_data.output_schema,
-                enable_web_search=input_data.enable_web_search,
-            )
-        except Exception as e:
-            raise BlockExecutionError(
-                message=f"Extract failed: {e}",
-                block_name=self.name,
-                block_id=self.id,
-            ) from e
+        extract_result = app.extract(
+            urls=input_data.urls,
+            prompt=input_data.prompt,
+            schema=input_data.output_schema,
+            enable_web_search=input_data.enable_web_search,
+        )
 
         yield "data", extract_result.data

autogpt_platform/backend/backend/blocks/flux_kontext.py

@@ -19,7 +19,6 @@ from backend.data.model import (
     SchemaField,
 )
 from backend.integrations.providers import ProviderName
-from backend.util.exceptions import ModerationError
 from backend.util.file import MediaFileType, store_media_file
 
 TEST_CREDENTIALS = APIKeyCredentials(
@@ -154,8 +153,6 @@ class AIImageEditorBlock(Block):
             ),
             aspect_ratio=input_data.aspect_ratio.value,
             seed=input_data.seed,
-            user_id=user_id,
-            graph_exec_id=graph_exec_id,
         )
         yield "output_image", result
 
@@ -167,8 +164,6 @@ class AIImageEditorBlock(Block):
         input_image_b64: Optional[str],
         aspect_ratio: str,
         seed: Optional[int],
-        user_id: str,
-        graph_exec_id: str,
     ) -> MediaFileType:
         client = ReplicateClient(api_token=api_key.get_secret_value())
         input_params = {
@@ -178,21 +173,11 @@ class AIImageEditorBlock(Block):
             **({"seed": seed} if seed is not None else {}),
         }
 
-        try:
-            output: FileOutput | list[FileOutput] = await client.async_run(  # type: ignore
-                model_name,
-                input=input_params,
-                wait=False,
-            )
-        except Exception as e:
-            if "flagged as sensitive" in str(e).lower():
-                raise ModerationError(
-                    message="Content was flagged as sensitive by the model provider",
-                    user_id=user_id,
-                    graph_exec_id=graph_exec_id,
-                    moderation_type="model_provider",
-                )
-            raise ValueError(f"Model execution failed: {e}") from e
+        output: FileOutput | list[FileOutput] = await client.async_run(  # type: ignore
+            model_name,
+            input=input_params,
+            wait=False,
+        )
 
         if isinstance(output, list) and output:
             output = output[0]

autogpt_platform/backend/backend/blocks/helpers/review.py

@@ -1,184 +0,0 @@
-"""
-Shared helpers for Human-In-The-Loop (HITL) review functionality.
-Used by both the dedicated HumanInTheLoopBlock and blocks that require human review.
-"""
-
-import logging
-from typing import Any, Optional
-
-from prisma.enums import ReviewStatus
-from pydantic import BaseModel
-
-from backend.data.execution import ExecutionContext, ExecutionStatus
-from backend.data.human_review import ReviewResult
-from backend.executor.manager import async_update_node_execution_status
-from backend.util.clients import get_database_manager_async_client
-
-logger = logging.getLogger(__name__)
-
-
-class ReviewDecision(BaseModel):
-    """Result of a review decision."""
-
-    should_proceed: bool
-    message: str
-    review_result: ReviewResult
-
-
-class HITLReviewHelper:
-    """Helper class for Human-In-The-Loop review operations."""
-
-    @staticmethod
-    async def get_or_create_human_review(**kwargs) -> Optional[ReviewResult]:
-        """Create or retrieve a human review from the database."""
-        return await get_database_manager_async_client().get_or_create_human_review(
-            **kwargs
-        )
-
-    @staticmethod
-    async def update_node_execution_status(**kwargs) -> None:
-        """Update the execution status of a node."""
-        await async_update_node_execution_status(
-            db_client=get_database_manager_async_client(), **kwargs
-        )
-
-    @staticmethod
-    async def update_review_processed_status(
-        node_exec_id: str, processed: bool
-    ) -> None:
-        """Update the processed status of a review."""
-        return await get_database_manager_async_client().update_review_processed_status(
-            node_exec_id, processed
-        )
-
-    @staticmethod
-    async def _handle_review_request(
-        input_data: Any,
-        user_id: str,
-        node_exec_id: str,
-        graph_exec_id: str,
-        graph_id: str,
-        graph_version: int,
-        execution_context: ExecutionContext,
-        block_name: str = "Block",
-        editable: bool = False,
-    ) -> Optional[ReviewResult]:
-        """
-        Handle a review request for a block that requires human review.
-
-        Args:
-            input_data: The input data to be reviewed
-            user_id: ID of the user requesting the review
-            node_exec_id: ID of the node execution
-            graph_exec_id: ID of the graph execution
-            graph_id: ID of the graph
-            graph_version: Version of the graph
-            execution_context: Current execution context
-            block_name: Name of the block requesting review
-            editable: Whether the reviewer can edit the data
-
-        Returns:
-            ReviewResult if review is complete, None if waiting for human input
-
-        Raises:
-            Exception: If review creation or status update fails
-        """
-        # Skip review if safe mode is disabled - return auto-approved result
-        if not execution_context.safe_mode:
-            logger.info(
-                f"Block {block_name} skipping review for node {node_exec_id} - safe mode disabled"
-            )
-            return ReviewResult(
-                data=input_data,
-                status=ReviewStatus.APPROVED,
-                message="Auto-approved (safe mode disabled)",
-                processed=True,
-                node_exec_id=node_exec_id,
-            )
-
-        result = await HITLReviewHelper.get_or_create_human_review(
-            user_id=user_id,
-            node_exec_id=node_exec_id,
-            graph_exec_id=graph_exec_id,
-            graph_id=graph_id,
-            graph_version=graph_version,
-            input_data=input_data,
-            message=f"Review required for {block_name} execution",
-            editable=editable,
-        )
-
-        if result is None:
-            logger.info(
-                f"Block {block_name} pausing execution for node {node_exec_id} - awaiting human review"
-            )
-            await HITLReviewHelper.update_node_execution_status(
-                exec_id=node_exec_id,
-                status=ExecutionStatus.REVIEW,
-            )
-            return None  # Signal that execution should pause
-
-        # Mark review as processed if not already done
-        if not result.processed:
-            await HITLReviewHelper.update_review_processed_status(
-                node_exec_id=node_exec_id, processed=True
-            )
-
-        return result
-
-    @staticmethod
-    async def handle_review_decision(
-        input_data: Any,
-        user_id: str,
-        node_exec_id: str,
-        graph_exec_id: str,
-        graph_id: str,
-        graph_version: int,
-        execution_context: ExecutionContext,
-        block_name: str = "Block",
-        editable: bool = False,
-    ) -> Optional[ReviewDecision]:
-        """
-        Handle a review request and return the decision in a single call.
-
-        Args:
-            input_data: The input data to be reviewed
-            user_id: ID of the user requesting the review
-            node_exec_id: ID of the node execution
-            graph_exec_id: ID of the graph execution
-            graph_id: ID of the graph
-            graph_version: Version of the graph
-            execution_context: Current execution context
-            block_name: Name of the block requesting review
-            editable: Whether the reviewer can edit the data
-
-        Returns:
-            ReviewDecision if review is complete (approved/rejected),
-            None if execution should pause (awaiting review)
-        """
-        review_result = await HITLReviewHelper._handle_review_request(
-            input_data=input_data,
-            user_id=user_id,
-            node_exec_id=node_exec_id,
-            graph_exec_id=graph_exec_id,
-            graph_id=graph_id,
-            graph_version=graph_version,
-            execution_context=execution_context,
-            block_name=block_name,
-            editable=editable,
-        )
-
-        if review_result is None:
-            # Still awaiting review - return None to pause execution
-            return None
-
-        # Review is complete, determine outcome
-        should_proceed = review_result.status == ReviewStatus.APPROVED
-        message = review_result.message or (
-            "Execution approved by reviewer"
-            if should_proceed
-            else "Execution rejected by reviewer"
-        )
-
-        return ReviewDecision(
-            should_proceed=should_proceed, message=message, review_result=review_result
-        )

autogpt_platform/backend/backend/blocks/human_in_the_loop.py

@@ -1,9 +1,8 @@
 import logging
-from typing import Any
+from typing import Any, Literal
 
 from prisma.enums import ReviewStatus
 
-from backend.blocks.helpers.review import HITLReviewHelper
 from backend.data.block import (
     Block,
     BlockCategory,
@@ -12,9 +11,11 @@ from backend.data.block import (
     BlockSchemaOutput,
     BlockType,
 )
-from backend.data.execution import ExecutionContext
+from backend.data.execution import ExecutionContext, ExecutionStatus
 from backend.data.human_review import ReviewResult
 from backend.data.model import SchemaField
+from backend.executor.manager import async_update_node_execution_status
+from backend.util.clients import get_database_manager_async_client
 
 logger = logging.getLogger(__name__)
 
@@ -44,11 +45,11 @@ class HumanInTheLoopBlock(Block):
         )
 
     class Output(BlockSchemaOutput):
-        approved_data: Any = SchemaField(
-            description="The data when approved (may be modified by reviewer)"
+        reviewed_data: Any = SchemaField(
+            description="The data after human review (may be modified)"
         )
-        rejected_data: Any = SchemaField(
-            description="The data when rejected (may be modified by reviewer)"
+        status: Literal["approved", "rejected"] = SchemaField(
+            description="Status of the review: 'approved' or 'rejected'"
         )
         review_message: str = SchemaField(
             description="Any message provided by the reviewer", default=""
@@ -68,29 +69,36 @@ class HumanInTheLoopBlock(Block):
                 "editable": True,
             },
             test_output=[
-                ("approved_data", {"name": "John Doe", "age": 30}),
+                ("status", "approved"),
+                ("reviewed_data", {"name": "John Doe", "age": 30}),
             ],
             test_mock={
-                "handle_review_decision": lambda **kwargs: type(
-                    "ReviewDecision",
-                    (),
-                    {
-                        "should_proceed": True,
-                        "message": "Test approval message",
-                        "review_result": ReviewResult(
-                            data={"name": "John Doe", "age": 30},
-                            status=ReviewStatus.APPROVED,
-                            message="",
-                            processed=False,
-                            node_exec_id="test-node-exec-id",
-                        ),
-                    },
-                )(),
+                "get_or_create_human_review": lambda *_args, **_kwargs: ReviewResult(
+                    data={"name": "John Doe", "age": 30},
+                    status=ReviewStatus.APPROVED,
+                    message="",
+                    processed=False,
+                    node_exec_id="test-node-exec-id",
+                ),
+                "update_node_execution_status": lambda *_args, **_kwargs: None,
+                "update_review_processed_status": lambda *_args, **_kwargs: None,
             },
         )
 
-    async def handle_review_decision(self, **kwargs):
-        return await HITLReviewHelper.handle_review_decision(**kwargs)
+    async def get_or_create_human_review(self, **kwargs):
+        return await get_database_manager_async_client().get_or_create_human_review(
+            **kwargs
+        )
+
+    async def update_node_execution_status(self, **kwargs):
+        return await async_update_node_execution_status(
+            db_client=get_database_manager_async_client(), **kwargs
+        )
+
+    async def update_review_processed_status(self, node_exec_id: str, processed: bool):
+        return await get_database_manager_async_client().update_review_processed_status(
+            node_exec_id, processed
+        )
 
     async def run(
         self,
@@ -102,38 +110,60 @@ class HumanInTheLoopBlock(Block):
         graph_id: str,
         graph_version: int,
         execution_context: ExecutionContext,
-        **_kwargs,
+        **kwargs,
     ) -> BlockOutput:
         if not execution_context.safe_mode:
             logger.info(
                 f"HITL block skipping review for node {node_exec_id} - safe mode disabled"
             )
-            yield "approved_data", input_data.data
+            yield "status", "approved"
+            yield "reviewed_data", input_data.data
             yield "review_message", "Auto-approved (safe mode disabled)"
             return
 
-        decision = await self.handle_review_decision(
-            input_data=input_data.data,
-            user_id=user_id,
-            node_exec_id=node_exec_id,
-            graph_exec_id=graph_exec_id,
-            graph_id=graph_id,
-            graph_version=graph_version,
-            execution_context=execution_context,
-            block_name=self.name,
-            editable=input_data.editable,
-        )
+        try:
+            result = await self.get_or_create_human_review(
+                user_id=user_id,
+                node_exec_id=node_exec_id,
+                graph_exec_id=graph_exec_id,
+                graph_id=graph_id,
+                graph_version=graph_version,
+                input_data=input_data.data,
+                message=input_data.name,
+                editable=input_data.editable,
+            )
+        except Exception as e:
+            logger.error(f"Error in HITL block for node {node_exec_id}: {str(e)}")
+            raise
 
-        if decision is None:
-            return
+        if result is None:
+            logger.info(
+                f"HITL block pausing execution for node {node_exec_id} - awaiting human review"
+            )
+            try:
+                await self.update_node_execution_status(
+                    exec_id=node_exec_id,
+                    status=ExecutionStatus.REVIEW,
+                )
+                return
+            except Exception as e:
+                logger.error(
+                    f"Failed to update node status for HITL block {node_exec_id}: {str(e)}"
+                )
+                raise
 
-        status = decision.review_result.status
-        if status == ReviewStatus.APPROVED:
-            yield "approved_data", decision.review_result.data
-        elif status == ReviewStatus.REJECTED:
-            yield "rejected_data", decision.review_result.data
-        else:
-            raise RuntimeError(f"Unexpected review status: {status}")
+        if not result.processed:
+            await self.update_review_processed_status(
+                node_exec_id=node_exec_id, processed=True
+            )
 
-        if decision.message:
-            yield "review_message", decision.message
+            if result.status == ReviewStatus.APPROVED:
+                yield "status", "approved"
+                yield "reviewed_data", result.data
+                if result.message:
+                    yield "review_message", result.message
+
+            elif result.status == ReviewStatus.REJECTED:
+                yield "status", "rejected"
+                if result.message:
+                    yield "review_message", result.message

autogpt_platform/backend/backend/blocks/ideogram.py

@@ -2,6 +2,7 @@ from enum import Enum
 from typing import Any, Dict, Literal, Optional
 
 from pydantic import SecretStr
+from requests.exceptions import RequestException
 
 from backend.data.block import (
     Block,
@@ -331,8 +332,8 @@ class IdeogramModelBlock(Block):
         try:
             response = await Requests().post(url, headers=headers, json=data)
             return response.json()["data"][0]["url"]
-        except Exception as e:
-            raise ValueError(f"Failed to fetch image with V3 endpoint: {e}") from e
+        except RequestException as e:
+            raise Exception(f"Failed to fetch image with V3 endpoint: {str(e)}")
 
     async def _run_model_legacy(
         self,
@@ -384,8 +385,8 @@ class IdeogramModelBlock(Block):
         try:
             response = await Requests().post(url, headers=headers, json=data)
             return response.json()["data"][0]["url"]
-        except Exception as e:
-            raise ValueError(f"Failed to fetch image with legacy endpoint: {e}") from e
+        except RequestException as e:
+            raise Exception(f"Failed to fetch image with legacy endpoint: {str(e)}")
 
     async def upscale_image(self, api_key: SecretStr, image_url: str):
         url = "https://api.ideogram.ai/upscale"
@@ -412,5 +413,5 @@ class IdeogramModelBlock(Block):
 
             return (response.json())["data"][0]["url"]
 
-        except Exception as e:
-            raise ValueError(f"Failed to upscale image: {e}") from e
+        except RequestException as e:
+            raise Exception(f"Failed to upscale image: {str(e)}")

autogpt_platform/backend/backend/blocks/jina/search.py

@@ -16,7 +16,6 @@ from backend.data.block import (
     BlockSchemaOutput,
 )
 from backend.data.model import SchemaField
-from backend.util.exceptions import BlockExecutionError
 
 
 class SearchTheWebBlock(Block, GetRequest):
@@ -57,17 +56,7 @@ class SearchTheWebBlock(Block, GetRequest):
 
         # Prepend the Jina Search URL to the encoded query
         jina_search_url = f"https://s.jina.ai/{encoded_query}"
-
-        try:
-            results = await self.get_request(
-                jina_search_url, headers=headers, json=False
-            )
-        except Exception as e:
-            raise BlockExecutionError(
-                message=f"Search failed: {e}",
-                block_name=self.name,
-                block_id=self.id,
-            ) from e
+        results = await self.get_request(jina_search_url, headers=headers, json=False)
 
         # Output the search results
         yield "results", results

autogpt_platform/backend/backend/blocks/llm.py

@@ -92,9 +92,8 @@ class LlmModel(str, Enum, metaclass=LlmModelMeta):
     O1 = "o1"
     O1_MINI = "o1-mini"
     # GPT-5 models
-    GPT5_2 = "gpt-5.2-2025-12-11"
-    GPT5_1 = "gpt-5.1-2025-11-13"
     GPT5 = "gpt-5-2025-08-07"
+    GPT5_1 = "gpt-5.1-2025-11-13"
     GPT5_MINI = "gpt-5-mini-2025-08-07"
     GPT5_NANO = "gpt-5-nano-2025-08-07"
     GPT5_CHAT = "gpt-5-chat-latest"
@@ -195,9 +194,8 @@ MODEL_METADATA = {
     LlmModel.O1: ModelMetadata("openai", 200000, 100000),  # o1-2024-12-17
     LlmModel.O1_MINI: ModelMetadata("openai", 128000, 65536),  # o1-mini-2024-09-12
     # GPT-5 models
-    LlmModel.GPT5_2: ModelMetadata("openai", 400000, 128000),
-    LlmModel.GPT5_1: ModelMetadata("openai", 400000, 128000),
     LlmModel.GPT5: ModelMetadata("openai", 400000, 128000),
+    LlmModel.GPT5_1: ModelMetadata("openai", 400000, 128000),
     LlmModel.GPT5_MINI: ModelMetadata("openai", 400000, 128000),
     LlmModel.GPT5_NANO: ModelMetadata("openai", 400000, 128000),
     LlmModel.GPT5_CHAT: ModelMetadata("openai", 400000, 16384),
@@ -305,8 +303,6 @@ MODEL_METADATA = {
     LlmModel.V0_1_0_MD: ModelMetadata("v0", 128000, 64000),
 }
 
-DEFAULT_LLM_MODEL = LlmModel.GPT5_2
-
 for model in LlmModel:
     if model not in MODEL_METADATA:
         raise ValueError(f"Missing MODEL_METADATA metadata for model: {model}")
@@ -794,7 +790,7 @@ class AIStructuredResponseGeneratorBlock(AIBlockBase):
         )
         model: LlmModel = SchemaField(
             title="LLM Model",
-            default=DEFAULT_LLM_MODEL,
+            default=LlmModel.GPT4O,
             description="The language model to use for answering the prompt.",
             advanced=False,
         )
@@ -859,7 +855,7 @@ class AIStructuredResponseGeneratorBlock(AIBlockBase):
             input_schema=AIStructuredResponseGeneratorBlock.Input,
             output_schema=AIStructuredResponseGeneratorBlock.Output,
             test_input={
-                "model": DEFAULT_LLM_MODEL,
+                "model": LlmModel.GPT4O,
                 "credentials": TEST_CREDENTIALS_INPUT,
                 "expected_format": {
                     "key1": "value1",
@@ -1225,7 +1221,7 @@ class AITextGeneratorBlock(AIBlockBase):
         )
         model: LlmModel = SchemaField(
             title="LLM Model",
-            default=DEFAULT_LLM_MODEL,
+            default=LlmModel.GPT4O,
             description="The language model to use for answering the prompt.",
             advanced=False,
         )
@@ -1321,7 +1317,7 @@ class AITextSummarizerBlock(AIBlockBase):
         )
         model: LlmModel = SchemaField(
             title="LLM Model",
-            default=DEFAULT_LLM_MODEL,
+            default=LlmModel.GPT4O,
             description="The language model to use for summarizing the text.",
         )
         focus: str = SchemaField(
@@ -1538,7 +1534,7 @@ class AIConversationBlock(AIBlockBase):
         )
         model: LlmModel = SchemaField(
             title="LLM Model",
-            default=DEFAULT_LLM_MODEL,
+            default=LlmModel.GPT4O,
             description="The language model to use for the conversation.",
         )
         credentials: AICredentials = AICredentialsField()
@@ -1576,7 +1572,7 @@ class AIConversationBlock(AIBlockBase):
                     },
                     {"role": "user", "content": "Where was it played?"},
                 ],
-                "model": DEFAULT_LLM_MODEL,
+                "model": LlmModel.GPT4O,
                 "credentials": TEST_CREDENTIALS_INPUT,
             },
             test_credentials=TEST_CREDENTIALS,
@@ -1639,7 +1635,7 @@ class AIListGeneratorBlock(AIBlockBase):
         )
         model: LlmModel = SchemaField(
             title="LLM Model",
-            default=DEFAULT_LLM_MODEL,
+            default=LlmModel.GPT4O,
             description="The language model to use for generating the list.",
             advanced=True,
         )
@@ -1696,7 +1692,7 @@ class AIListGeneratorBlock(AIBlockBase):
                     "drawing explorers to uncover its mysteries. Each planet showcases the limitless possibilities of "
                     "fictional worlds."
                 ),
-                "model": DEFAULT_LLM_MODEL,
+                "model": LlmModel.GPT4O,
                 "credentials": TEST_CREDENTIALS_INPUT,
                 "max_retries": 3,
                 "force_json_output": False,

autogpt_platform/backend/backend/blocks/replicate/replicate_block.py

@@ -18,7 +18,6 @@ from backend.data.block import (
     BlockSchemaOutput,
 )
 from backend.data.model import APIKeyCredentials, CredentialsField, SchemaField
-from backend.util.exceptions import BlockExecutionError, BlockInputError
 
 logger = logging.getLogger(__name__)
 
@@ -112,27 +111,9 @@ class ReplicateModelBlock(Block):
             yield "status", "succeeded"
             yield "model_name", input_data.model_name
         except Exception as e:
-            error_msg = str(e)
-            logger.error(f"Error running Replicate model: {error_msg}")
-
-            # Input validation errors (422, 400) → BlockInputError
-            if (
-                "422" in error_msg
-                or "Input validation failed" in error_msg
-                or "400" in error_msg
-            ):
-                raise BlockInputError(
-                    message=f"Invalid model inputs: {error_msg}",
-                    block_name=self.name,
-                    block_id=self.id,
-                ) from e
-            # Everything else → BlockExecutionError
-            else:
-                raise BlockExecutionError(
-                    message=f"Replicate model error: {error_msg}",
-                    block_name=self.name,
-                    block_id=self.id,
-                ) from e
+            error_msg = f"Unexpected error running Replicate model: {str(e)}"
+            logger.error(error_msg)
+            raise RuntimeError(error_msg)
 
     async def run_model(self, model_ref: str, model_inputs: dict, api_key: SecretStr):
         """

autogpt_platform/backend/backend/blocks/search.py

@@ -18,7 +18,6 @@ from backend.data.model import (
     SchemaField,
 )
 from backend.integrations.providers import ProviderName
-from backend.util.request import DEFAULT_USER_AGENT
 
 
 class GetWikipediaSummaryBlock(Block, GetRequest):
@@ -40,32 +39,16 @@ class GetWikipediaSummaryBlock(Block, GetRequest):
             output_schema=GetWikipediaSummaryBlock.Output,
             test_input={"topic": "Artificial Intelligence"},
             test_output=("summary", "summary content"),
-            test_mock={
-                "get_request": lambda url, headers, json: {"extract": "summary content"}
-            },
+            test_mock={"get_request": lambda url, json: {"extract": "summary content"}},
         )
 
     async def run(self, input_data: Input, **kwargs) -> BlockOutput:
         topic = input_data.topic
-        # URL-encode the topic to handle spaces and special characters
-        encoded_topic = quote(topic, safe="")
-        url = f"https://en.wikipedia.org/api/rest_v1/page/summary/{encoded_topic}"
-
-        # Set headers per Wikimedia robot policy (https://w.wiki/4wJS)
-        # - User-Agent: Required, must identify the bot
-        # - Accept-Encoding: gzip recommended to reduce bandwidth
-        headers = {
-            "User-Agent": DEFAULT_USER_AGENT,
-            "Accept-Encoding": "gzip, deflate",
-        }
-
-        try:
-            response = await self.get_request(url, headers=headers, json=True)
-            if "extract" not in response:
-                raise ValueError(f"Unable to parse Wikipedia response: {response}")
-            yield "summary", response["extract"]
-        except Exception as e:
-            raise ValueError(f"Failed to fetch Wikipedia summary: {e}") from e
+        url = f"https://en.wikipedia.org/api/rest_v1/page/summary/{topic}"
+        response = await self.get_request(url, json=True)
+        if "extract" not in response:
+            raise RuntimeError(f"Unable to parse Wikipedia response: {response}")
+        yield "summary", response["extract"]
 
 
 TEST_CREDENTIALS = APIKeyCredentials(

autogpt_platform/backend/backend/blocks/smart_decision_maker.py

@@ -1,11 +1,8 @@
 import logging
 import re
 from collections import Counter
-from concurrent.futures import Future
 from typing import TYPE_CHECKING, Any
 
-from pydantic import BaseModel
-
 import backend.blocks.llm as llm
 from backend.blocks.agent import AgentExecutorBlock
 from backend.data.block import (
@@ -23,41 +20,16 @@ from backend.data.dynamic_fields import (
     is_dynamic_field,
     is_tool_pin,
 )
-from backend.data.execution import ExecutionContext
 from backend.data.model import NodeExecutionStats, SchemaField
 from backend.util import json
 from backend.util.clients import get_database_manager_async_client
-from backend.util.prompt import MAIN_OBJECTIVE_PREFIX
 
 if TYPE_CHECKING:
     from backend.data.graph import Link, Node
-    from backend.executor.manager import ExecutionProcessor
 
 logger = logging.getLogger(__name__)
 
 
-class ToolInfo(BaseModel):
-    """Processed tool call information."""
-
-    tool_call: Any  # The original tool call object from LLM response
-    tool_name: str  # The function name
-    tool_def: dict[str, Any]  # The tool definition from tool_functions
-    input_data: dict[str, Any]  # Processed input data ready for tool execution
-    field_mapping: dict[str, str]  # Field name mapping for the tool
-
-
-class ExecutionParams(BaseModel):
-    """Tool execution parameters."""
-
-    user_id: str
-    graph_id: str
-    node_id: str
-    graph_version: int
-    graph_exec_id: str
-    node_exec_id: str
-    execution_context: "ExecutionContext"
-
-
 def _get_tool_requests(entry: dict[str, Any]) -> list[str]:
     """
     Return a list of tool_call_ids if the entry is a tool request.
@@ -133,50 +105,6 @@ def _create_tool_response(call_id: str, output: Any) -> dict[str, Any]:
     return {"role": "tool", "tool_call_id": call_id, "content": content}
 
 
-def _combine_tool_responses(tool_outputs: list[dict[str, Any]]) -> list[dict[str, Any]]:
-    """
-    Combine multiple Anthropic tool responses into a single user message.
-    For non-Anthropic formats, returns the original list unchanged.
-    """
-    if len(tool_outputs) <= 1:
-        return tool_outputs
-
-    # Anthropic responses have role="user", type="message", and content is a list with tool_result items
-    anthropic_responses = [
-        output
-        for output in tool_outputs
-        if (
-            output.get("role") == "user"
-            and output.get("type") == "message"
-            and isinstance(output.get("content"), list)
-            and any(
-                item.get("type") == "tool_result"
-                for item in output.get("content", [])
-                if isinstance(item, dict)
-            )
-        )
-    ]
-
-    if len(anthropic_responses) > 1:
-        combined_content = [
-            item for response in anthropic_responses for item in response["content"]
-        ]
-
-        combined_response = {
-            "role": "user",
-            "type": "message",
-            "content": combined_content,
-        }
-
-        non_anthropic_responses = [
-            output for output in tool_outputs if output not in anthropic_responses
-        ]
-
-        return [combined_response] + non_anthropic_responses
-
-    return tool_outputs
-
-
 def _convert_raw_response_to_dict(raw_response: Any) -> dict[str, Any]:
     """
     Safely convert raw_response to dictionary format for conversation history.
@@ -226,7 +154,7 @@ class SmartDecisionMakerBlock(Block):
         )
         model: llm.LlmModel = SchemaField(
             title="LLM Model",
-            default=llm.DEFAULT_LLM_MODEL,
+            default=llm.LlmModel.GPT4O,
             description="The language model to use for answering the prompt.",
             advanced=False,
         )
@@ -276,17 +204,6 @@ class SmartDecisionMakerBlock(Block):
             default="localhost:11434",
             description="Ollama host for local  models",
         )
-        agent_mode_max_iterations: int = SchemaField(
-            title="Agent Mode Max Iterations",
-            description="Maximum iterations for agent mode. 0 = traditional mode (single LLM call, yield tool calls for external execution), -1 = infinite agent mode (loop until finished), 1+ = agent mode with max iterations limit.",
-            advanced=True,
-            default=0,
-        )
-        conversation_compaction: bool = SchemaField(
-            default=True,
-            title="Context window auto-compaction",
-            description="Automatically compact the context window once it hits the limit",
-        )
 
         @classmethod
         def get_missing_links(cls, data: BlockInput, links: list["Link"]) -> set[str]:
@@ -391,12 +308,8 @@ class SmartDecisionMakerBlock(Block):
         """
         block = sink_node.block
 
-        # Use custom name from node metadata if set, otherwise fall back to block.name
-        custom_name = sink_node.metadata.get("customized_name")
-        tool_name = custom_name if custom_name else block.name
-
         tool_function: dict[str, Any] = {
-            "name": SmartDecisionMakerBlock.cleanup(tool_name),
+            "name": SmartDecisionMakerBlock.cleanup(block.name),
             "description": block.description,
         }
         sink_block_input_schema = block.input_schema
@@ -493,24 +406,14 @@ class SmartDecisionMakerBlock(Block):
                 f"Sink graph metadata not found: {graph_id} {graph_version}"
             )
 
-        # Use custom name from node metadata if set, otherwise fall back to graph name
-        custom_name = sink_node.metadata.get("customized_name")
-        tool_name = custom_name if custom_name else sink_graph_meta.name
-
         tool_function: dict[str, Any] = {
-            "name": SmartDecisionMakerBlock.cleanup(tool_name),
+            "name": SmartDecisionMakerBlock.cleanup(sink_graph_meta.name),
             "description": sink_graph_meta.description,
         }
 
         properties = {}
-        field_mapping = {}
 
         for link in links:
-            field_name = link.sink_name
-
-            clean_field_name = SmartDecisionMakerBlock.cleanup(field_name)
-            field_mapping[clean_field_name] = field_name
-
             sink_block_input_schema = sink_node.input_default["input_schema"]
             sink_block_properties = sink_block_input_schema.get("properties", {}).get(
                 link.sink_name, {}
@@ -520,7 +423,7 @@ class SmartDecisionMakerBlock(Block):
                 if "description" in sink_block_properties
                 else f"The {link.sink_name} of the tool"
             )
-            properties[clean_field_name] = {
+            properties[link.sink_name] = {
                 "type": "string",
                 "description": description,
                 "default": json.dumps(sink_block_properties.get("default", None)),
@@ -533,7 +436,7 @@ class SmartDecisionMakerBlock(Block):
             "strict": True,
         }
 
-        tool_function["_field_mapping"] = field_mapping
+        # Store node info for later use in output processing
         tool_function["_sink_node_id"] = sink_node.id
 
         return {"type": "function", "function": tool_function}
@@ -603,7 +506,6 @@ class SmartDecisionMakerBlock(Block):
         Returns the response if successful, raises ValueError if validation fails.
         """
         resp = await llm.llm_call(
-            compress_prompt_to_fit=input_data.conversation_compaction,
             credentials=credentials,
             llm_model=input_data.model,
             prompt=current_prompt,
@@ -691,291 +593,6 @@ class SmartDecisionMakerBlock(Block):
 
         return resp
 
-    def _process_tool_calls(
-        self, response, tool_functions: list[dict[str, Any]]
-    ) -> list[ToolInfo]:
-        """Process tool calls and extract tool definitions, arguments, and input data.
-
-        Returns a list of tool info dicts with:
-        - tool_call: The original tool call object
-        - tool_name: The function name
-        - tool_def: The tool definition from tool_functions
-        - input_data: Processed input data dict (includes None values)
-        - field_mapping: Field name mapping for the tool
-        """
-        if not response.tool_calls:
-            return []
-
-        processed_tools = []
-        for tool_call in response.tool_calls:
-            tool_name = tool_call.function.name
-            tool_args = json.loads(tool_call.function.arguments)
-
-            tool_def = next(
-                (
-                    tool
-                    for tool in tool_functions
-                    if tool["function"]["name"] == tool_name
-                ),
-                None,
-            )
-            if not tool_def:
-                if len(tool_functions) == 1:
-                    tool_def = tool_functions[0]
-                else:
-                    continue
-
-            # Build input data for the tool
-            input_data = {}
-            field_mapping = tool_def["function"].get("_field_mapping", {})
-            if "function" in tool_def and "parameters" in tool_def["function"]:
-                expected_args = tool_def["function"]["parameters"].get("properties", {})
-                for clean_arg_name in expected_args:
-                    original_field_name = field_mapping.get(
-                        clean_arg_name, clean_arg_name
-                    )
-                    arg_value = tool_args.get(clean_arg_name)
-                    # Include all expected parameters, even if None (for backward compatibility with tests)
-                    input_data[original_field_name] = arg_value
-
-            processed_tools.append(
-                ToolInfo(
-                    tool_call=tool_call,
-                    tool_name=tool_name,
-                    tool_def=tool_def,
-                    input_data=input_data,
-                    field_mapping=field_mapping,
-                )
-            )
-
-        return processed_tools
-
-    def _update_conversation(
-        self, prompt: list[dict], response, tool_outputs: list | None = None
-    ):
-        """Update conversation history with response and tool outputs."""
-        # Don't add separate reasoning message with tool calls (breaks Anthropic's tool_use->tool_result pairing)
-        assistant_message = _convert_raw_response_to_dict(response.raw_response)
-        has_tool_calls = isinstance(assistant_message.get("content"), list) and any(
-            item.get("type") == "tool_use"
-            for item in assistant_message.get("content", [])
-        )
-
-        if response.reasoning and not has_tool_calls:
-            prompt.append(
-                {"role": "assistant", "content": f"[Reasoning]: {response.reasoning}"}
-            )
-
-        prompt.append(assistant_message)
-
-        if tool_outputs:
-            prompt.extend(tool_outputs)
-
-    async def _execute_single_tool_with_manager(
-        self,
-        tool_info: ToolInfo,
-        execution_params: ExecutionParams,
-        execution_processor: "ExecutionProcessor",
-    ) -> dict:
-        """Execute a single tool using the execution manager for proper integration."""
-        # Lazy imports to avoid circular dependencies
-        from backend.data.execution import NodeExecutionEntry
-
-        tool_call = tool_info.tool_call
-        tool_def = tool_info.tool_def
-        raw_input_data = tool_info.input_data
-
-        # Get sink node and field mapping
-        sink_node_id = tool_def["function"]["_sink_node_id"]
-
-        # Use proper database operations for tool execution
-        db_client = get_database_manager_async_client()
-
-        # Get target node
-        target_node = await db_client.get_node(sink_node_id)
-        if not target_node:
-            raise ValueError(f"Target node {sink_node_id} not found")
-
-        # Create proper node execution using upsert_execution_input
-        node_exec_result = None
-        final_input_data = None
-
-        # Add all inputs to the execution
-        if not raw_input_data:
-            raise ValueError(f"Tool call has no input data: {tool_call}")
-
-        for input_name, input_value in raw_input_data.items():
-            node_exec_result, final_input_data = await db_client.upsert_execution_input(
-                node_id=sink_node_id,
-                graph_exec_id=execution_params.graph_exec_id,
-                input_name=input_name,
-                input_data=input_value,
-            )
-
-        assert node_exec_result is not None, "node_exec_result should not be None"
-
-        # Create NodeExecutionEntry for execution manager
-        node_exec_entry = NodeExecutionEntry(
-            user_id=execution_params.user_id,
-            graph_exec_id=execution_params.graph_exec_id,
-            graph_id=execution_params.graph_id,
-            graph_version=execution_params.graph_version,
-            node_exec_id=node_exec_result.node_exec_id,
-            node_id=sink_node_id,
-            block_id=target_node.block_id,
-            inputs=final_input_data or {},
-            execution_context=execution_params.execution_context,
-        )
-
-        # Use the execution manager to execute the tool node
-        try:
-            # Get NodeExecutionProgress from the execution manager's running nodes
-            node_exec_progress = execution_processor.running_node_execution[
-                sink_node_id
-            ]
-
-            # Use the execution manager's own graph stats
-            graph_stats_pair = (
-                execution_processor.execution_stats,
-                execution_processor.execution_stats_lock,
-            )
-
-            # Create a completed future for the task tracking system
-            node_exec_future = Future()
-            node_exec_progress.add_task(
-                node_exec_id=node_exec_result.node_exec_id,
-                task=node_exec_future,
-            )
-
-            # Execute the node directly since we're in the SmartDecisionMaker context
-            node_exec_future.set_result(
-                await execution_processor.on_node_execution(
-                    node_exec=node_exec_entry,
-                    node_exec_progress=node_exec_progress,
-                    nodes_input_masks=None,
-                    graph_stats_pair=graph_stats_pair,
-                )
-            )
-
-            # Get outputs from database after execution completes using database manager client
-            node_outputs = await db_client.get_execution_outputs_by_node_exec_id(
-                node_exec_result.node_exec_id
-            )
-
-            # Create tool response
-            tool_response_content = (
-                json.dumps(node_outputs)
-                if node_outputs
-                else "Tool executed successfully"
-            )
-            return _create_tool_response(tool_call.id, tool_response_content)
-
-        except Exception as e:
-            logger.error(f"Tool execution with manager failed: {e}")
-            # Return error response
-            return _create_tool_response(
-                tool_call.id, f"Tool execution failed: {str(e)}"
-            )
-
-    async def _execute_tools_agent_mode(
-        self,
-        input_data,
-        credentials,
-        tool_functions: list[dict[str, Any]],
-        prompt: list[dict],
-        graph_exec_id: str,
-        node_id: str,
-        node_exec_id: str,
-        user_id: str,
-        graph_id: str,
-        graph_version: int,
-        execution_context: ExecutionContext,
-        execution_processor: "ExecutionProcessor",
-    ):
-        """Execute tools in agent mode with a loop until finished."""
-        max_iterations = input_data.agent_mode_max_iterations
-        iteration = 0
-
-        # Execution parameters for tool execution
-        execution_params = ExecutionParams(
-            user_id=user_id,
-            graph_id=graph_id,
-            node_id=node_id,
-            graph_version=graph_version,
-            graph_exec_id=graph_exec_id,
-            node_exec_id=node_exec_id,
-            execution_context=execution_context,
-        )
-
-        current_prompt = list(prompt)
-
-        while max_iterations < 0 or iteration < max_iterations:
-            iteration += 1
-            logger.debug(f"Agent mode iteration {iteration}")
-
-            # Prepare prompt for this iteration
-            iteration_prompt = list(current_prompt)
-
-            # On the last iteration, add a special system message to encourage completion
-            if max_iterations > 0 and iteration == max_iterations:
-                last_iteration_message = {
-                    "role": "system",
-                    "content": f"{MAIN_OBJECTIVE_PREFIX}This is your last iteration ({iteration}/{max_iterations}). "
-                    "Try to complete the task with the information you have. If you cannot fully complete it, "
-                    "provide a summary of what you've accomplished and what remains to be done. "
-                    "Prefer finishing with a clear response rather than making additional tool calls.",
-                }
-                iteration_prompt.append(last_iteration_message)
-
-            # Get LLM response
-            try:
-                response = await self._attempt_llm_call_with_validation(
-                    credentials, input_data, iteration_prompt, tool_functions
-                )
-            except Exception as e:
-                yield "error", f"LLM call failed in agent mode iteration {iteration}: {str(e)}"
-                return
-
-            # Process tool calls
-            processed_tools = self._process_tool_calls(response, tool_functions)
-
-            # If no tool calls, we're done
-            if not processed_tools:
-                yield "finished", response.response
-                self._update_conversation(current_prompt, response)
-                yield "conversations", current_prompt
-                return
-
-            # Execute tools and collect responses
-            tool_outputs = []
-            for tool_info in processed_tools:
-                try:
-                    tool_response = await self._execute_single_tool_with_manager(
-                        tool_info, execution_params, execution_processor
-                    )
-                    tool_outputs.append(tool_response)
-                except Exception as e:
-                    logger.error(f"Tool execution failed: {e}")
-                    # Create error response for the tool
-                    error_response = _create_tool_response(
-                        tool_info.tool_call.id, f"Error: {str(e)}"
-                    )
-                    tool_outputs.append(error_response)
-
-            tool_outputs = _combine_tool_responses(tool_outputs)
-
-            self._update_conversation(current_prompt, response, tool_outputs)
-
-            # Yield intermediate conversation state
-            yield "conversations", current_prompt
-
-        # If we reach max iterations, yield the current state
-        if max_iterations < 0:
-            yield "finished", f"Agent mode completed after {iteration} iterations"
-        else:
-            yield "finished", f"Agent mode completed after {max_iterations} iterations (limit reached)"
-        yield "conversations", current_prompt
-
     async def run(
         self,
         input_data: Input,
@@ -986,31 +603,9 @@ class SmartDecisionMakerBlock(Block):
         graph_exec_id: str,
         node_exec_id: str,
         user_id: str,
-        graph_version: int,
-        execution_context: ExecutionContext,
-        execution_processor: "ExecutionProcessor",
-        nodes_to_skip: set[str] | None = None,
         **kwargs,
     ) -> BlockOutput:
-
         tool_functions = await self._create_tool_node_signatures(node_id)
-        original_tool_count = len(tool_functions)
-
-        # Filter out tools for nodes that should be skipped (e.g., missing optional credentials)
-        if nodes_to_skip:
-            tool_functions = [
-                tf
-                for tf in tool_functions
-                if tf.get("function", {}).get("_sink_node_id") not in nodes_to_skip
-            ]
-
-            # Only raise error if we had tools but they were all filtered out
-            if original_tool_count > 0 and not tool_functions:
-                raise ValueError(
-                    "No available tools to execute - all downstream nodes are unavailable "
-                    "(possibly due to missing optional credentials)"
-                )
-
         yield "tool_functions", json.dumps(tool_functions)
 
         conversation_history = input_data.conversation_history or []
@@ -1053,52 +648,24 @@ class SmartDecisionMakerBlock(Block):
             input_data.prompt = llm.fmt.format_string(input_data.prompt, values)
             input_data.sys_prompt = llm.fmt.format_string(input_data.sys_prompt, values)
 
+        prefix = "[Main Objective Prompt]: "
+
         if input_data.sys_prompt and not any(
-            p["role"] == "system" and p["content"].startswith(MAIN_OBJECTIVE_PREFIX)
-            for p in prompt
+            p["role"] == "system" and p["content"].startswith(prefix) for p in prompt
         ):
-            prompt.append(
-                {
-                    "role": "system",
-                    "content": MAIN_OBJECTIVE_PREFIX + input_data.sys_prompt,
-                }
-            )
+            prompt.append({"role": "system", "content": prefix + input_data.sys_prompt})
 
         if input_data.prompt and not any(
-            p["role"] == "user" and p["content"].startswith(MAIN_OBJECTIVE_PREFIX)
-            for p in prompt
+            p["role"] == "user" and p["content"].startswith(prefix) for p in prompt
         ):
-            prompt.append(
-                {"role": "user", "content": MAIN_OBJECTIVE_PREFIX + input_data.prompt}
-            )
+            prompt.append({"role": "user", "content": prefix + input_data.prompt})
 
-        # Execute tools based on the selected mode
-        if input_data.agent_mode_max_iterations != 0:
-            # In agent mode, execute tools directly in a loop until finished
-            async for result in self._execute_tools_agent_mode(
-                input_data=input_data,
-                credentials=credentials,
-                tool_functions=tool_functions,
-                prompt=prompt,
-                graph_exec_id=graph_exec_id,
-                node_id=node_id,
-                node_exec_id=node_exec_id,
-                user_id=user_id,
-                graph_id=graph_id,
-                graph_version=graph_version,
-                execution_context=execution_context,
-                execution_processor=execution_processor,
-            ):
-                yield result
-            return
-
-        # One-off mode: single LLM call and yield tool calls for external execution
         current_prompt = list(prompt)
         max_attempts = max(1, int(input_data.retry))
         response = None
 
         last_error = None
-        for _ in range(max_attempts):
+        for attempt in range(max_attempts):
             try:
                 response = await self._attempt_llm_call_with_validation(
                     credentials, input_data, current_prompt, tool_functions
@@ -1161,9 +728,8 @@ class SmartDecisionMakerBlock(Block):
                 original_field_name = field_mapping.get(clean_arg_name, clean_arg_name)
                 arg_value = tool_args.get(clean_arg_name)
 
-                # Use original_field_name directly (not sanitized) to match link sink_name
-                # The field_mapping already translates from LLM's cleaned names to original names
-                emit_key = f"tools_^_{sink_node_id}_~_{original_field_name}"
+                sanitized_arg_name = self.cleanup(original_field_name)
+                emit_key = f"tools_^_{sink_node_id}_~_{sanitized_arg_name}"
 
                 logger.debug(
                     "[SmartDecisionMakerBlock|geid:%s|neid:%s] emit %s",

autogpt_platform/backend/backend/blocks/test/test_blocks_dos_vulnerability.py

@@ -196,15 +196,6 @@ class TestXMLParserBlockSecurity:
             async for _ in block.run(XMLParserBlock.Input(input_xml=large_xml)):
                 pass
 
-    async def test_rejects_text_outside_root(self):
-        """Ensure parser surfaces readable errors for invalid root text."""
-        block = XMLParserBlock()
-        invalid_xml = "<root><child>value</child></root> trailing"
-
-        with pytest.raises(ValueError, match="text outside the root element"):
-            async for _ in block.run(XMLParserBlock.Input(input_xml=invalid_xml)):
-                pass
-
 
 class TestStoreMediaFileSecurity:
     """Test file storage security limits."""

autogpt_platform/backend/backend/blocks/test/test_llm.py

@@ -28,7 +28,7 @@ class TestLLMStatsTracking:
 
             response = await llm.llm_call(
                 credentials=llm.TEST_CREDENTIALS,
-                llm_model=llm.DEFAULT_LLM_MODEL,
+                llm_model=llm.LlmModel.GPT4O,
                 prompt=[{"role": "user", "content": "Hello"}],
                 max_tokens=100,
             )
@@ -65,7 +65,7 @@ class TestLLMStatsTracking:
         input_data = llm.AIStructuredResponseGeneratorBlock.Input(
             prompt="Test prompt",
             expected_format={"key1": "desc1", "key2": "desc2"},
-            model=llm.DEFAULT_LLM_MODEL,
+            model=llm.LlmModel.GPT4O,
             credentials=llm.TEST_CREDENTIALS_INPUT,  # type: ignore  # type: ignore
         )
 
@@ -109,7 +109,7 @@ class TestLLMStatsTracking:
         # Run the block
         input_data = llm.AITextGeneratorBlock.Input(
             prompt="Generate text",
-            model=llm.DEFAULT_LLM_MODEL,
+            model=llm.LlmModel.GPT4O,
             credentials=llm.TEST_CREDENTIALS_INPUT,  # type: ignore
         )
 
@@ -170,7 +170,7 @@ class TestLLMStatsTracking:
         input_data = llm.AIStructuredResponseGeneratorBlock.Input(
             prompt="Test prompt",
             expected_format={"key1": "desc1", "key2": "desc2"},
-            model=llm.DEFAULT_LLM_MODEL,
+            model=llm.LlmModel.GPT4O,
             credentials=llm.TEST_CREDENTIALS_INPUT,  # type: ignore
             retry=2,
         )
@@ -228,7 +228,7 @@ class TestLLMStatsTracking:
 
         input_data = llm.AITextSummarizerBlock.Input(
             text=long_text,
-            model=llm.DEFAULT_LLM_MODEL,
+            model=llm.LlmModel.GPT4O,
             credentials=llm.TEST_CREDENTIALS_INPUT,  # type: ignore
             max_tokens=100,  # Small chunks
             chunk_overlap=10,
@@ -299,7 +299,7 @@ class TestLLMStatsTracking:
             # Test with very short text (should only need 1 chunk + 1 final summary)
             input_data = llm.AITextSummarizerBlock.Input(
                 text="This is a short text.",
-                model=llm.DEFAULT_LLM_MODEL,
+                model=llm.LlmModel.GPT4O,
                 credentials=llm.TEST_CREDENTIALS_INPUT,  # type: ignore
                 max_tokens=1000,  # Large enough to avoid chunking
             )
@@ -346,7 +346,7 @@ class TestLLMStatsTracking:
                 {"role": "assistant", "content": "Hi there!"},
                 {"role": "user", "content": "How are you?"},
             ],
-            model=llm.DEFAULT_LLM_MODEL,
+            model=llm.LlmModel.GPT4O,
             credentials=llm.TEST_CREDENTIALS_INPUT,  # type: ignore
         )
 
@@ -387,7 +387,7 @@ class TestLLMStatsTracking:
         # Run the block
         input_data = llm.AIListGeneratorBlock.Input(
             focus="test items",
-            model=llm.DEFAULT_LLM_MODEL,
+            model=llm.LlmModel.GPT4O,
             credentials=llm.TEST_CREDENTIALS_INPUT,  # type: ignore
             max_retries=3,
         )
@@ -469,7 +469,7 @@ class TestLLMStatsTracking:
         input_data = llm.AIStructuredResponseGeneratorBlock.Input(
             prompt="Test",
             expected_format={"result": "desc"},
-            model=llm.DEFAULT_LLM_MODEL,
+            model=llm.LlmModel.GPT4O,
             credentials=llm.TEST_CREDENTIALS_INPUT,  # type: ignore
         )
 
@@ -513,7 +513,7 @@ class TestAITextSummarizerValidation:
         # Create input data
         input_data = llm.AITextSummarizerBlock.Input(
             text="Some text to summarize",
-            model=llm.DEFAULT_LLM_MODEL,
+            model=llm.LlmModel.GPT4O,
             credentials=llm.TEST_CREDENTIALS_INPUT,  # type: ignore
             style=llm.SummaryStyle.BULLET_POINTS,
         )
@@ -558,7 +558,7 @@ class TestAITextSummarizerValidation:
         # Create input data
         input_data = llm.AITextSummarizerBlock.Input(
             text="Some text to summarize",
-            model=llm.DEFAULT_LLM_MODEL,
+            model=llm.LlmModel.GPT4O,
             credentials=llm.TEST_CREDENTIALS_INPUT,  # type: ignore
             style=llm.SummaryStyle.BULLET_POINTS,
             max_tokens=1000,
@@ -593,7 +593,7 @@ class TestAITextSummarizerValidation:
         # Create input data
         input_data = llm.AITextSummarizerBlock.Input(
             text="Some text to summarize",
-            model=llm.DEFAULT_LLM_MODEL,
+            model=llm.LlmModel.GPT4O,
             credentials=llm.TEST_CREDENTIALS_INPUT,  # type: ignore
         )
 
@@ -623,7 +623,7 @@ class TestAITextSummarizerValidation:
         # Create input data
         input_data = llm.AITextSummarizerBlock.Input(
             text="Some text to summarize",
-            model=llm.DEFAULT_LLM_MODEL,
+            model=llm.LlmModel.GPT4O,
             credentials=llm.TEST_CREDENTIALS_INPUT,  # type: ignore
             max_tokens=1000,
         )
@@ -654,7 +654,7 @@ class TestAITextSummarizerValidation:
         # Create input data
         input_data = llm.AITextSummarizerBlock.Input(
             text="Some text to summarize",
-            model=llm.DEFAULT_LLM_MODEL,
+            model=llm.LlmModel.GPT4O,
             credentials=llm.TEST_CREDENTIALS_INPUT,  # type: ignore
         )
 

autogpt_platform/backend/backend/blocks/test/test_smart_decision_maker.py

@@ -1,14 +1,10 @@
 import logging
-import threading
-from collections import defaultdict
-from unittest.mock import AsyncMock, MagicMock, patch
 
 import pytest
 
-from backend.api.model import CreateGraph
-from backend.api.rest_api import AgentServer
-from backend.data.execution import ExecutionContext
 from backend.data.model import ProviderName, User
+from backend.server.model import CreateGraph
+from backend.server.rest_api import AgentServer
 from backend.usecases.sample import create_test_graph, create_test_user
 from backend.util.test import SpinTestServer, wait_execution
 
@@ -21,10 +17,10 @@ async def create_graph(s: SpinTestServer, g, u: User):
 
 
 async def create_credentials(s: SpinTestServer, u: User):
-    import backend.blocks.llm as llm_module
+    import backend.blocks.llm as llm
 
     provider = ProviderName.OPENAI
-    credentials = llm_module.TEST_CREDENTIALS
+    credentials = llm.TEST_CREDENTIALS
     return await s.agent_server.test_create_credentials(u.id, provider, credentials)
 
 
@@ -200,6 +196,8 @@ async def test_smart_decision_maker_function_signature(server: SpinTestServer):
 @pytest.mark.asyncio
 async def test_smart_decision_maker_tracks_llm_stats():
     """Test that SmartDecisionMakerBlock correctly tracks LLM usage stats."""
+    from unittest.mock import MagicMock, patch
+
     import backend.blocks.llm as llm_module
     from backend.blocks.smart_decision_maker import SmartDecisionMakerBlock
 
@@ -218,6 +216,7 @@ async def test_smart_decision_maker_tracks_llm_stats():
     }
 
     # Mock the _create_tool_node_signatures method to avoid database calls
+    from unittest.mock import AsyncMock
 
     with patch(
         "backend.blocks.llm.llm_call",
@@ -233,21 +232,12 @@ async def test_smart_decision_maker_tracks_llm_stats():
         # Create test input
         input_data = SmartDecisionMakerBlock.Input(
             prompt="Should I continue with this task?",
-            model=llm_module.DEFAULT_LLM_MODEL,
+            model=llm_module.LlmModel.GPT4O,
             credentials=llm_module.TEST_CREDENTIALS_INPUT,  # type: ignore
-            agent_mode_max_iterations=0,
         )
 
         # Execute the block
         outputs = {}
-        # Create execution context
-
-        mock_execution_context = ExecutionContext(safe_mode=False)
-
-        # Create a mock execution processor for tests
-
-        mock_execution_processor = MagicMock()
-
         async for output_name, output_data in block.run(
             input_data,
             credentials=llm_module.TEST_CREDENTIALS,
@@ -256,9 +246,6 @@ async def test_smart_decision_maker_tracks_llm_stats():
             graph_exec_id="test-exec-id",
             node_exec_id="test-node-exec-id",
             user_id="test-user-id",
-            graph_version=1,
-            execution_context=mock_execution_context,
-            execution_processor=mock_execution_processor,
         ):
             outputs[output_name] = output_data
 
@@ -276,6 +263,8 @@ async def test_smart_decision_maker_tracks_llm_stats():
 @pytest.mark.asyncio
 async def test_smart_decision_maker_parameter_validation():
     """Test that SmartDecisionMakerBlock correctly validates tool call parameters."""
+    from unittest.mock import MagicMock, patch
+
     import backend.blocks.llm as llm_module
     from backend.blocks.smart_decision_maker import SmartDecisionMakerBlock
 
@@ -322,6 +311,8 @@ async def test_smart_decision_maker_parameter_validation():
     mock_response_with_typo.reasoning = None
     mock_response_with_typo.raw_response = {"role": "assistant", "content": None}
 
+    from unittest.mock import AsyncMock
+
     with patch(
         "backend.blocks.llm.llm_call",
         new_callable=AsyncMock,
@@ -335,20 +326,11 @@ async def test_smart_decision_maker_parameter_validation():
 
         input_data = SmartDecisionMakerBlock.Input(
             prompt="Search for keywords",
-            model=llm_module.DEFAULT_LLM_MODEL,
+            model=llm_module.LlmModel.GPT4O,
             credentials=llm_module.TEST_CREDENTIALS_INPUT,  # type: ignore
             retry=2,  # Set retry to 2 for testing
-            agent_mode_max_iterations=0,
         )
 
-        # Create execution context
-
-        mock_execution_context = ExecutionContext(safe_mode=False)
-
-        # Create a mock execution processor for tests
-
-        mock_execution_processor = MagicMock()
-
         # Should raise ValueError after retries due to typo'd parameter name
         with pytest.raises(ValueError) as exc_info:
             outputs = {}
@@ -360,9 +342,6 @@ async def test_smart_decision_maker_parameter_validation():
                 graph_exec_id="test-exec-id",
                 node_exec_id="test-node-exec-id",
                 user_id="test-user-id",
-                graph_version=1,
-                execution_context=mock_execution_context,
-                execution_processor=mock_execution_processor,
             ):
                 outputs[output_name] = output_data
 
@@ -389,6 +368,8 @@ async def test_smart_decision_maker_parameter_validation():
     mock_response_missing_required.reasoning = None
     mock_response_missing_required.raw_response = {"role": "assistant", "content": None}
 
+    from unittest.mock import AsyncMock
+
     with patch(
         "backend.blocks.llm.llm_call",
         new_callable=AsyncMock,
@@ -402,19 +383,10 @@ async def test_smart_decision_maker_parameter_validation():
 
         input_data = SmartDecisionMakerBlock.Input(
             prompt="Search for keywords",
-            model=llm_module.DEFAULT_LLM_MODEL,
+            model=llm_module.LlmModel.GPT4O,
             credentials=llm_module.TEST_CREDENTIALS_INPUT,  # type: ignore
-            agent_mode_max_iterations=0,
         )
 
-        # Create execution context
-
-        mock_execution_context = ExecutionContext(safe_mode=False)
-
-        # Create a mock execution processor for tests
-
-        mock_execution_processor = MagicMock()
-
         # Should raise ValueError due to missing required parameter
         with pytest.raises(ValueError) as exc_info:
             outputs = {}
@@ -426,9 +398,6 @@ async def test_smart_decision_maker_parameter_validation():
                 graph_exec_id="test-exec-id",
                 node_exec_id="test-node-exec-id",
                 user_id="test-user-id",
-                graph_version=1,
-                execution_context=mock_execution_context,
-                execution_processor=mock_execution_processor,
             ):
                 outputs[output_name] = output_data
 
@@ -449,6 +418,8 @@ async def test_smart_decision_maker_parameter_validation():
     mock_response_valid.reasoning = None
     mock_response_valid.raw_response = {"role": "assistant", "content": None}
 
+    from unittest.mock import AsyncMock
+
     with patch(
         "backend.blocks.llm.llm_call",
         new_callable=AsyncMock,
@@ -462,21 +433,12 @@ async def test_smart_decision_maker_parameter_validation():
 
         input_data = SmartDecisionMakerBlock.Input(
             prompt="Search for keywords",
-            model=llm_module.DEFAULT_LLM_MODEL,
+            model=llm_module.LlmModel.GPT4O,
             credentials=llm_module.TEST_CREDENTIALS_INPUT,  # type: ignore
-            agent_mode_max_iterations=0,
         )
 
         # Should succeed - optional parameter missing is OK
         outputs = {}
-        # Create execution context
-
-        mock_execution_context = ExecutionContext(safe_mode=False)
-
-        # Create a mock execution processor for tests
-
-        mock_execution_processor = MagicMock()
-
         async for output_name, output_data in block.run(
             input_data,
             credentials=llm_module.TEST_CREDENTIALS,
@@ -485,9 +447,6 @@ async def test_smart_decision_maker_parameter_validation():
             graph_exec_id="test-exec-id",
             node_exec_id="test-node-exec-id",
             user_id="test-user-id",
-            graph_version=1,
-            execution_context=mock_execution_context,
-            execution_processor=mock_execution_processor,
         ):
             outputs[output_name] = output_data
 
@@ -513,6 +472,8 @@ async def test_smart_decision_maker_parameter_validation():
     mock_response_all_params.reasoning = None
     mock_response_all_params.raw_response = {"role": "assistant", "content": None}
 
+    from unittest.mock import AsyncMock
+
     with patch(
         "backend.blocks.llm.llm_call",
         new_callable=AsyncMock,
@@ -526,21 +487,12 @@ async def test_smart_decision_maker_parameter_validation():
 
         input_data = SmartDecisionMakerBlock.Input(
             prompt="Search for keywords",
-            model=llm_module.DEFAULT_LLM_MODEL,
+            model=llm_module.LlmModel.GPT4O,
             credentials=llm_module.TEST_CREDENTIALS_INPUT,  # type: ignore
-            agent_mode_max_iterations=0,
         )
 
         # Should succeed with all parameters
         outputs = {}
-        # Create execution context
-
-        mock_execution_context = ExecutionContext(safe_mode=False)
-
-        # Create a mock execution processor for tests
-
-        mock_execution_processor = MagicMock()
-
         async for output_name, output_data in block.run(
             input_data,
             credentials=llm_module.TEST_CREDENTIALS,
@@ -549,9 +501,6 @@ async def test_smart_decision_maker_parameter_validation():
             graph_exec_id="test-exec-id",
             node_exec_id="test-node-exec-id",
             user_id="test-user-id",
-            graph_version=1,
-            execution_context=mock_execution_context,
-            execution_processor=mock_execution_processor,
         ):
             outputs[output_name] = output_data
 
@@ -564,6 +513,8 @@ async def test_smart_decision_maker_parameter_validation():
 @pytest.mark.asyncio
 async def test_smart_decision_maker_raw_response_conversion():
     """Test that SmartDecisionMaker correctly handles different raw_response types with retry mechanism."""
+    from unittest.mock import MagicMock, patch
+
     import backend.blocks.llm as llm_module
     from backend.blocks.smart_decision_maker import SmartDecisionMakerBlock
 
@@ -633,6 +584,7 @@ async def test_smart_decision_maker_raw_response_conversion():
     )
 
     # Mock llm_call to return different responses on different calls
+    from unittest.mock import AsyncMock
 
     with patch(
         "backend.blocks.llm.llm_call", new_callable=AsyncMock
@@ -648,22 +600,13 @@ async def test_smart_decision_maker_raw_response_conversion():
 
         input_data = SmartDecisionMakerBlock.Input(
             prompt="Test prompt",
-            model=llm_module.DEFAULT_LLM_MODEL,
+            model=llm_module.LlmModel.GPT4O,
             credentials=llm_module.TEST_CREDENTIALS_INPUT,  # type: ignore
             retry=2,
-            agent_mode_max_iterations=0,
         )
 
         # Should succeed after retry, demonstrating our helper function works
         outputs = {}
-        # Create execution context
-
-        mock_execution_context = ExecutionContext(safe_mode=False)
-
-        # Create a mock execution processor for tests
-
-        mock_execution_processor = MagicMock()
-
         async for output_name, output_data in block.run(
             input_data,
             credentials=llm_module.TEST_CREDENTIALS,
@@ -672,9 +615,6 @@ async def test_smart_decision_maker_raw_response_conversion():
             graph_exec_id="test-exec-id",
             node_exec_id="test-node-exec-id",
             user_id="test-user-id",
-            graph_version=1,
-            execution_context=mock_execution_context,
-            execution_processor=mock_execution_processor,
         ):
             outputs[output_name] = output_data
 
@@ -710,6 +650,8 @@ async def test_smart_decision_maker_raw_response_conversion():
         "I'll help you with that."  # Ollama returns string
     )
 
+    from unittest.mock import AsyncMock
+
     with patch(
         "backend.blocks.llm.llm_call",
         new_callable=AsyncMock,
@@ -722,20 +664,11 @@ async def test_smart_decision_maker_raw_response_conversion():
     ):
         input_data = SmartDecisionMakerBlock.Input(
             prompt="Simple prompt",
-            model=llm_module.DEFAULT_LLM_MODEL,
+            model=llm_module.LlmModel.GPT4O,
             credentials=llm_module.TEST_CREDENTIALS_INPUT,  # type: ignore
-            agent_mode_max_iterations=0,
         )
 
         outputs = {}
-        # Create execution context
-
-        mock_execution_context = ExecutionContext(safe_mode=False)
-
-        # Create a mock execution processor for tests
-
-        mock_execution_processor = MagicMock()
-
         async for output_name, output_data in block.run(
             input_data,
             credentials=llm_module.TEST_CREDENTIALS,
@@ -744,9 +677,6 @@ async def test_smart_decision_maker_raw_response_conversion():
             graph_exec_id="test-exec-id",
             node_exec_id="test-node-exec-id",
             user_id="test-user-id",
-            graph_version=1,
-            execution_context=mock_execution_context,
-            execution_processor=mock_execution_processor,
         ):
             outputs[output_name] = output_data
 
@@ -766,6 +696,8 @@ async def test_smart_decision_maker_raw_response_conversion():
         "content": "Test response",
     }  # Dict format
 
+    from unittest.mock import AsyncMock
+
     with patch(
         "backend.blocks.llm.llm_call",
         new_callable=AsyncMock,
@@ -778,20 +710,11 @@ async def test_smart_decision_maker_raw_response_conversion():
     ):
         input_data = SmartDecisionMakerBlock.Input(
             prompt="Another test",
-            model=llm_module.DEFAULT_LLM_MODEL,
+            model=llm_module.LlmModel.GPT4O,
             credentials=llm_module.TEST_CREDENTIALS_INPUT,  # type: ignore
-            agent_mode_max_iterations=0,
         )
 
         outputs = {}
-        # Create execution context
-
-        mock_execution_context = ExecutionContext(safe_mode=False)
-
-        # Create a mock execution processor for tests
-
-        mock_execution_processor = MagicMock()
-
         async for output_name, output_data in block.run(
             input_data,
             credentials=llm_module.TEST_CREDENTIALS,
@@ -800,410 +723,8 @@ async def test_smart_decision_maker_raw_response_conversion():
             graph_exec_id="test-exec-id",
             node_exec_id="test-node-exec-id",
             user_id="test-user-id",
-            graph_version=1,
-            execution_context=mock_execution_context,
-            execution_processor=mock_execution_processor,
         ):
             outputs[output_name] = output_data
 
         assert "finished" in outputs
         assert outputs["finished"] == "Test response"
-
-
-@pytest.mark.asyncio
-async def test_smart_decision_maker_agent_mode():
-    """Test that agent mode executes tools directly and loops until finished."""
-    import backend.blocks.llm as llm_module
-    from backend.blocks.smart_decision_maker import SmartDecisionMakerBlock
-
-    block = SmartDecisionMakerBlock()
-
-    # Mock tool call that requires multiple iterations
-    mock_tool_call_1 = MagicMock()
-    mock_tool_call_1.id = "call_1"
-    mock_tool_call_1.function.name = "search_keywords"
-    mock_tool_call_1.function.arguments = (
-        '{"query": "test", "max_keyword_difficulty": 50}'
-    )
-
-    mock_response_1 = MagicMock()
-    mock_response_1.response = None
-    mock_response_1.tool_calls = [mock_tool_call_1]
-    mock_response_1.prompt_tokens = 50
-    mock_response_1.completion_tokens = 25
-    mock_response_1.reasoning = "Using search tool"
-    mock_response_1.raw_response = {
-        "role": "assistant",
-        "content": None,
-        "tool_calls": [{"id": "call_1", "type": "function"}],
-    }
-
-    # Final response with no tool calls (finished)
-    mock_response_2 = MagicMock()
-    mock_response_2.response = "Task completed successfully"
-    mock_response_2.tool_calls = []
-    mock_response_2.prompt_tokens = 30
-    mock_response_2.completion_tokens = 15
-    mock_response_2.reasoning = None
-    mock_response_2.raw_response = {
-        "role": "assistant",
-        "content": "Task completed successfully",
-    }
-
-    # Mock the LLM call to return different responses on each iteration
-    llm_call_mock = AsyncMock()
-    llm_call_mock.side_effect = [mock_response_1, mock_response_2]
-
-    # Mock tool node signatures
-    mock_tool_signatures = [
-        {
-            "type": "function",
-            "function": {
-                "name": "search_keywords",
-                "_sink_node_id": "test-sink-node-id",
-                "_field_mapping": {},
-                "parameters": {
-                    "properties": {
-                        "query": {"type": "string"},
-                        "max_keyword_difficulty": {"type": "integer"},
-                    },
-                    "required": ["query", "max_keyword_difficulty"],
-                },
-            },
-        }
-    ]
-
-    # Mock database and execution components
-    mock_db_client = AsyncMock()
-    mock_node = MagicMock()
-    mock_node.block_id = "test-block-id"
-    mock_db_client.get_node.return_value = mock_node
-
-    # Mock upsert_execution_input to return proper NodeExecutionResult and input data
-    mock_node_exec_result = MagicMock()
-    mock_node_exec_result.node_exec_id = "test-tool-exec-id"
-    mock_input_data = {"query": "test", "max_keyword_difficulty": 50}
-    mock_db_client.upsert_execution_input.return_value = (
-        mock_node_exec_result,
-        mock_input_data,
-    )
-
-    # No longer need mock_execute_node since we use execution_processor.on_node_execution
-
-    with patch("backend.blocks.llm.llm_call", llm_call_mock), patch.object(
-        block, "_create_tool_node_signatures", return_value=mock_tool_signatures
-    ), patch(
-        "backend.blocks.smart_decision_maker.get_database_manager_async_client",
-        return_value=mock_db_client,
-    ), patch(
-        "backend.executor.manager.async_update_node_execution_status",
-        new_callable=AsyncMock,
-    ), patch(
-        "backend.integrations.creds_manager.IntegrationCredentialsManager"
-    ):
-
-        # Create a mock execution context
-
-        mock_execution_context = ExecutionContext(
-            safe_mode=False,
-        )
-
-        # Create a mock execution processor for agent mode tests
-
-        mock_execution_processor = AsyncMock()
-        # Configure the execution processor mock with required attributes
-        mock_execution_processor.running_node_execution = defaultdict(MagicMock)
-        mock_execution_processor.execution_stats = MagicMock()
-        mock_execution_processor.execution_stats_lock = threading.Lock()
-
-        # Mock the on_node_execution method to return successful stats
-        mock_node_stats = MagicMock()
-        mock_node_stats.error = None  # No error
-        mock_execution_processor.on_node_execution = AsyncMock(
-            return_value=mock_node_stats
-        )
-
-        # Mock the get_execution_outputs_by_node_exec_id method
-        mock_db_client.get_execution_outputs_by_node_exec_id.return_value = {
-            "result": {"status": "success", "data": "search completed"}
-        }
-
-        # Test agent mode with max_iterations = 3
-        input_data = SmartDecisionMakerBlock.Input(
-            prompt="Complete this task using tools",
-            model=llm_module.DEFAULT_LLM_MODEL,
-            credentials=llm_module.TEST_CREDENTIALS_INPUT,  # type: ignore
-            agent_mode_max_iterations=3,  # Enable agent mode with 3 max iterations
-        )
-
-        outputs = {}
-        async for output_name, output_data in block.run(
-            input_data,
-            credentials=llm_module.TEST_CREDENTIALS,
-            graph_id="test-graph-id",
-            node_id="test-node-id",
-            graph_exec_id="test-exec-id",
-            node_exec_id="test-node-exec-id",
-            user_id="test-user-id",
-            graph_version=1,
-            execution_context=mock_execution_context,
-            execution_processor=mock_execution_processor,
-        ):
-            outputs[output_name] = output_data
-
-        # Verify agent mode behavior
-        assert "tool_functions" in outputs  # tool_functions is yielded in both modes
-        assert "finished" in outputs
-        assert outputs["finished"] == "Task completed successfully"
-        assert "conversations" in outputs
-
-        # Verify the conversation includes tool responses
-        conversations = outputs["conversations"]
-        assert len(conversations) > 2  # Should have multiple conversation entries
-
-        # Verify LLM was called twice (once for tool call, once for finish)
-        assert llm_call_mock.call_count == 2
-
-        # Verify tool was executed via execution processor
-        assert mock_execution_processor.on_node_execution.call_count == 1
-
-
-@pytest.mark.asyncio
-async def test_smart_decision_maker_traditional_mode_default():
-    """Test that default behavior (agent_mode_max_iterations=0) works as traditional mode."""
-    import backend.blocks.llm as llm_module
-    from backend.blocks.smart_decision_maker import SmartDecisionMakerBlock
-
-    block = SmartDecisionMakerBlock()
-
-    # Mock tool call
-    mock_tool_call = MagicMock()
-    mock_tool_call.function.name = "search_keywords"
-    mock_tool_call.function.arguments = (
-        '{"query": "test", "max_keyword_difficulty": 50}'
-    )
-
-    mock_response = MagicMock()
-    mock_response.response = None
-    mock_response.tool_calls = [mock_tool_call]
-    mock_response.prompt_tokens = 50
-    mock_response.completion_tokens = 25
-    mock_response.reasoning = None
-    mock_response.raw_response = {"role": "assistant", "content": None}
-
-    mock_tool_signatures = [
-        {
-            "type": "function",
-            "function": {
-                "name": "search_keywords",
-                "_sink_node_id": "test-sink-node-id",
-                "_field_mapping": {},
-                "parameters": {
-                    "properties": {
-                        "query": {"type": "string"},
-                        "max_keyword_difficulty": {"type": "integer"},
-                    },
-                    "required": ["query", "max_keyword_difficulty"],
-                },
-            },
-        }
-    ]
-
-    with patch(
-        "backend.blocks.llm.llm_call",
-        new_callable=AsyncMock,
-        return_value=mock_response,
-    ), patch.object(
-        block, "_create_tool_node_signatures", return_value=mock_tool_signatures
-    ):
-
-        # Test default behavior (traditional mode)
-        input_data = SmartDecisionMakerBlock.Input(
-            prompt="Test prompt",
-            model=llm_module.DEFAULT_LLM_MODEL,
-            credentials=llm_module.TEST_CREDENTIALS_INPUT,  # type: ignore
-            agent_mode_max_iterations=0,  # Traditional mode
-        )
-
-        # Create execution context
-
-        mock_execution_context = ExecutionContext(safe_mode=False)
-
-        # Create a mock execution processor for tests
-
-        mock_execution_processor = MagicMock()
-
-        outputs = {}
-        async for output_name, output_data in block.run(
-            input_data,
-            credentials=llm_module.TEST_CREDENTIALS,
-            graph_id="test-graph-id",
-            node_id="test-node-id",
-            graph_exec_id="test-exec-id",
-            node_exec_id="test-node-exec-id",
-            user_id="test-user-id",
-            graph_version=1,
-            execution_context=mock_execution_context,
-            execution_processor=mock_execution_processor,
-        ):
-            outputs[output_name] = output_data
-
-        # Verify traditional mode behavior
-        assert (
-            "tool_functions" in outputs
-        )  # Should yield tool_functions in traditional mode
-        assert (
-            "tools_^_test-sink-node-id_~_query" in outputs
-        )  # Should yield individual tool parameters
-        assert "tools_^_test-sink-node-id_~_max_keyword_difficulty" in outputs
-        assert "conversations" in outputs
-
-
-@pytest.mark.asyncio
-async def test_smart_decision_maker_uses_customized_name_for_blocks():
-    """Test that SmartDecisionMakerBlock uses customized_name from node metadata for tool names."""
-    from unittest.mock import MagicMock
-
-    from backend.blocks.basic import StoreValueBlock
-    from backend.blocks.smart_decision_maker import SmartDecisionMakerBlock
-    from backend.data.graph import Link, Node
-
-    # Create a mock node with customized_name in metadata
-    mock_node = MagicMock(spec=Node)
-    mock_node.id = "test-node-id"
-    mock_node.block_id = StoreValueBlock().id
-    mock_node.metadata = {"customized_name": "My Custom Tool Name"}
-    mock_node.block = StoreValueBlock()
-
-    # Create a mock link
-    mock_link = MagicMock(spec=Link)
-    mock_link.sink_name = "input"
-
-    # Call the function directly
-    result = await SmartDecisionMakerBlock._create_block_function_signature(
-        mock_node, [mock_link]
-    )
-
-    # Verify the tool name uses the customized name (cleaned up)
-    assert result["type"] == "function"
-    assert result["function"]["name"] == "my_custom_tool_name"  # Cleaned version
-    assert result["function"]["_sink_node_id"] == "test-node-id"
-
-
-@pytest.mark.asyncio
-async def test_smart_decision_maker_falls_back_to_block_name():
-    """Test that SmartDecisionMakerBlock falls back to block.name when no customized_name."""
-    from unittest.mock import MagicMock
-
-    from backend.blocks.basic import StoreValueBlock
-    from backend.blocks.smart_decision_maker import SmartDecisionMakerBlock
-    from backend.data.graph import Link, Node
-
-    # Create a mock node without customized_name
-    mock_node = MagicMock(spec=Node)
-    mock_node.id = "test-node-id"
-    mock_node.block_id = StoreValueBlock().id
-    mock_node.metadata = {}  # No customized_name
-    mock_node.block = StoreValueBlock()
-
-    # Create a mock link
-    mock_link = MagicMock(spec=Link)
-    mock_link.sink_name = "input"
-
-    # Call the function directly
-    result = await SmartDecisionMakerBlock._create_block_function_signature(
-        mock_node, [mock_link]
-    )
-
-    # Verify the tool name uses the block's default name
-    assert result["type"] == "function"
-    assert result["function"]["name"] == "storevalueblock"  # Default block name cleaned
-    assert result["function"]["_sink_node_id"] == "test-node-id"
-
-
-@pytest.mark.asyncio
-async def test_smart_decision_maker_uses_customized_name_for_agents():
-    """Test that SmartDecisionMakerBlock uses customized_name from metadata for agent nodes."""
-    from unittest.mock import AsyncMock, MagicMock, patch
-
-    from backend.blocks.smart_decision_maker import SmartDecisionMakerBlock
-    from backend.data.graph import Link, Node
-
-    # Create a mock node with customized_name in metadata
-    mock_node = MagicMock(spec=Node)
-    mock_node.id = "test-agent-node-id"
-    mock_node.metadata = {"customized_name": "My Custom Agent"}
-    mock_node.input_default = {
-        "graph_id": "test-graph-id",
-        "graph_version": 1,
-        "input_schema": {"properties": {"test_input": {"description": "Test input"}}},
-    }
-
-    # Create a mock link
-    mock_link = MagicMock(spec=Link)
-    mock_link.sink_name = "test_input"
-
-    # Mock the database client
-    mock_graph_meta = MagicMock()
-    mock_graph_meta.name = "Original Agent Name"
-    mock_graph_meta.description = "Agent description"
-
-    mock_db_client = AsyncMock()
-    mock_db_client.get_graph_metadata.return_value = mock_graph_meta
-
-    with patch(
-        "backend.blocks.smart_decision_maker.get_database_manager_async_client",
-        return_value=mock_db_client,
-    ):
-        result = await SmartDecisionMakerBlock._create_agent_function_signature(
-            mock_node, [mock_link]
-        )
-
-    # Verify the tool name uses the customized name (cleaned up)
-    assert result["type"] == "function"
-    assert result["function"]["name"] == "my_custom_agent"  # Cleaned version
-    assert result["function"]["_sink_node_id"] == "test-agent-node-id"
-
-
-@pytest.mark.asyncio
-async def test_smart_decision_maker_agent_falls_back_to_graph_name():
-    """Test that agent node falls back to graph name when no customized_name."""
-    from unittest.mock import AsyncMock, MagicMock, patch
-
-    from backend.blocks.smart_decision_maker import SmartDecisionMakerBlock
-    from backend.data.graph import Link, Node
-
-    # Create a mock node without customized_name
-    mock_node = MagicMock(spec=Node)
-    mock_node.id = "test-agent-node-id"
-    mock_node.metadata = {}  # No customized_name
-    mock_node.input_default = {
-        "graph_id": "test-graph-id",
-        "graph_version": 1,
-        "input_schema": {"properties": {"test_input": {"description": "Test input"}}},
-    }
-
-    # Create a mock link
-    mock_link = MagicMock(spec=Link)
-    mock_link.sink_name = "test_input"
-
-    # Mock the database client
-    mock_graph_meta = MagicMock()
-    mock_graph_meta.name = "Original Agent Name"
-    mock_graph_meta.description = "Agent description"
-
-    mock_db_client = AsyncMock()
-    mock_db_client.get_graph_metadata.return_value = mock_graph_meta
-
-    with patch(
-        "backend.blocks.smart_decision_maker.get_database_manager_async_client",
-        return_value=mock_db_client,
-    ):
-        result = await SmartDecisionMakerBlock._create_agent_function_signature(
-            mock_node, [mock_link]
-        )
-
-    # Verify the tool name uses the graph's default name
-    assert result["type"] == "function"
-    assert result["function"]["name"] == "original_agent_name"  # Graph name cleaned
-    assert result["function"]["_sink_node_id"] == "test-agent-node-id"

autogpt_platform/backend/backend/blocks/test/test_smart_decision_maker_dict.py

@@ -15,7 +15,6 @@ async def test_smart_decision_maker_handles_dynamic_dict_fields():
     mock_node.block = CreateDictionaryBlock()
     mock_node.block_id = CreateDictionaryBlock().id
     mock_node.input_default = {}
-    mock_node.metadata = {}
 
     # Create mock links with dynamic dictionary fields
     mock_links = [
@@ -78,7 +77,6 @@ async def test_smart_decision_maker_handles_dynamic_list_fields():
     mock_node.block = AddToListBlock()
     mock_node.block_id = AddToListBlock().id
     mock_node.input_default = {}
-    mock_node.metadata = {}
 
     # Create mock links with dynamic list fields
     mock_links = [

autogpt_platform/backend/backend/blocks/test/test_smart_decision_maker_dynamic_fields.py

@@ -1,7 +1,7 @@
 """Comprehensive tests for SmartDecisionMakerBlock dynamic field handling."""
 
 import json
-from unittest.mock import AsyncMock, MagicMock, Mock, patch
+from unittest.mock import AsyncMock, Mock, patch
 
 import pytest
 
@@ -44,7 +44,6 @@ async def test_create_block_function_signature_with_dict_fields():
     mock_node.block = CreateDictionaryBlock()
     mock_node.block_id = CreateDictionaryBlock().id
     mock_node.input_default = {}
-    mock_node.metadata = {}
 
     # Create mock links with dynamic dictionary fields (source sanitized, sink original)
     mock_links = [
@@ -107,7 +106,6 @@ async def test_create_block_function_signature_with_list_fields():
     mock_node.block = AddToListBlock()
     mock_node.block_id = AddToListBlock().id
     mock_node.input_default = {}
-    mock_node.metadata = {}
 
     # Create mock links with dynamic list fields
     mock_links = [
@@ -161,7 +159,6 @@ async def test_create_block_function_signature_with_object_fields():
     mock_node.block = MatchTextPatternBlock()
     mock_node.block_id = MatchTextPatternBlock().id
     mock_node.input_default = {}
-    mock_node.metadata = {}
 
     # Create mock links with dynamic object fields
     mock_links = [
@@ -211,13 +208,11 @@ async def test_create_tool_node_signatures():
         mock_dict_node.block = CreateDictionaryBlock()
         mock_dict_node.block_id = CreateDictionaryBlock().id
         mock_dict_node.input_default = {}
-        mock_dict_node.metadata = {}
 
         mock_list_node = Mock()
         mock_list_node.block = AddToListBlock()
         mock_list_node.block_id = AddToListBlock().id
         mock_list_node.input_default = {}
-        mock_list_node.metadata = {}
 
         # Mock links with dynamic fields
         dict_link1 = Mock(
@@ -313,47 +308,10 @@ async def test_output_yielding_with_dynamic_fields():
     ) as mock_llm:
         mock_llm.return_value = mock_response
 
-        # Mock the database manager to avoid HTTP calls during tool execution
-        with patch(
-            "backend.blocks.smart_decision_maker.get_database_manager_async_client"
-        ) as mock_db_manager, patch.object(
+        # Mock the function signature creation
+        with patch.object(
             block, "_create_tool_node_signatures", new_callable=AsyncMock
         ) as mock_sig:
-            # Set up the mock database manager
-            mock_db_client = AsyncMock()
-            mock_db_manager.return_value = mock_db_client
-
-            # Mock the node retrieval
-            mock_target_node = Mock()
-            mock_target_node.id = "test-sink-node-id"
-            mock_target_node.block_id = "CreateDictionaryBlock"
-            mock_target_node.block = Mock()
-            mock_target_node.block.name = "Create Dictionary"
-            mock_db_client.get_node.return_value = mock_target_node
-
-            # Mock the execution result creation
-            mock_node_exec_result = Mock()
-            mock_node_exec_result.node_exec_id = "mock-node-exec-id"
-            mock_final_input_data = {
-                "values_#_name": "Alice",
-                "values_#_age": 30,
-                "values_#_email": "alice@example.com",
-            }
-            mock_db_client.upsert_execution_input.return_value = (
-                mock_node_exec_result,
-                mock_final_input_data,
-            )
-
-            # Mock the output retrieval
-            mock_outputs = {
-                "values_#_name": "Alice",
-                "values_#_age": 30,
-                "values_#_email": "alice@example.com",
-            }
-            mock_db_client.get_execution_outputs_by_node_exec_id.return_value = (
-                mock_outputs
-            )
-
             mock_sig.return_value = [
                 {
                     "type": "function",
@@ -378,17 +336,11 @@ async def test_output_yielding_with_dynamic_fields():
             input_data = block.input_schema(
                 prompt="Create a user dictionary",
                 credentials=llm.TEST_CREDENTIALS_INPUT,
-                model=llm.DEFAULT_LLM_MODEL,
-                agent_mode_max_iterations=0,  # Use traditional mode to test output yielding
+                model=llm.LlmModel.GPT4O,
             )
 
             # Run the block
             outputs = {}
-            from backend.data.execution import ExecutionContext
-
-            mock_execution_context = ExecutionContext(safe_mode=False)
-            mock_execution_processor = MagicMock()
-
             async for output_name, output_value in block.run(
                 input_data,
                 credentials=llm.TEST_CREDENTIALS,
@@ -397,9 +349,6 @@ async def test_output_yielding_with_dynamic_fields():
                 graph_exec_id="test_exec",
                 node_exec_id="test_node_exec",
                 user_id="test_user",
-                graph_version=1,
-                execution_context=mock_execution_context,
-                execution_processor=mock_execution_processor,
             ):
                 outputs[output_name] = output_value
 
@@ -428,7 +377,6 @@ async def test_mixed_regular_and_dynamic_fields():
     mock_node.block.name = "TestBlock"
     mock_node.block.description = "A test block"
     mock_node.block.input_schema = Mock()
-    mock_node.metadata = {}
 
     # Mock the get_field_schema to return a proper schema for regular fields
     def get_field_schema(field_name):
@@ -563,108 +511,45 @@ async def test_validation_errors_dont_pollute_conversation():
                 }
             ]
 
-            # Mock the database manager to avoid HTTP calls during tool execution
-            with patch(
-                "backend.blocks.smart_decision_maker.get_database_manager_async_client"
-            ) as mock_db_manager:
-                # Set up the mock database manager for agent mode
-                mock_db_client = AsyncMock()
-                mock_db_manager.return_value = mock_db_client
+            # Create input data
+            from backend.blocks import llm
 
-                # Mock the node retrieval
-                mock_target_node = Mock()
-                mock_target_node.id = "test-sink-node-id"
-                mock_target_node.block_id = "TestBlock"
-                mock_target_node.block = Mock()
-                mock_target_node.block.name = "Test Block"
-                mock_db_client.get_node.return_value = mock_target_node
+            input_data = block.input_schema(
+                prompt="Test prompt",
+                credentials=llm.TEST_CREDENTIALS_INPUT,
+                model=llm.LlmModel.GPT4O,
+                retry=3,  # Allow retries
+            )
 
-                # Mock the execution result creation
-                mock_node_exec_result = Mock()
-                mock_node_exec_result.node_exec_id = "mock-node-exec-id"
-                mock_final_input_data = {"correct_param": "value"}
-                mock_db_client.upsert_execution_input.return_value = (
-                    mock_node_exec_result,
-                    mock_final_input_data,
-                )
+            # Run the block
+            outputs = {}
+            async for output_name, output_value in block.run(
+                input_data,
+                credentials=llm.TEST_CREDENTIALS,
+                graph_id="test_graph",
+                node_id="test_node",
+                graph_exec_id="test_exec",
+                node_exec_id="test_node_exec",
+                user_id="test_user",
+            ):
+                outputs[output_name] = output_value
 
-                # Mock the output retrieval
-                mock_outputs = {"correct_param": "value"}
-                mock_db_client.get_execution_outputs_by_node_exec_id.return_value = (
-                    mock_outputs
-                )
+            # Verify we had 2 LLM calls (initial + retry)
+            assert call_count == 2
 
-                # Create input data
-                from backend.blocks import llm
+            # Check the final conversation output
+            final_conversation = outputs.get("conversations", [])
 
-                input_data = block.input_schema(
-                    prompt="Test prompt",
-                    credentials=llm.TEST_CREDENTIALS_INPUT,
-                    model=llm.DEFAULT_LLM_MODEL,
-                    retry=3,  # Allow retries
-                    agent_mode_max_iterations=1,
-                )
+            # The final conversation should NOT contain the validation error message
+            error_messages = [
+                msg
+                for msg in final_conversation
+                if msg.get("role") == "user"
+                and "parameter errors" in msg.get("content", "")
+            ]
+            assert (
+                len(error_messages) == 0
+            ), "Validation error leaked into final conversation"
 
-                # Run the block
-                outputs = {}
-                from backend.data.execution import ExecutionContext
-
-                mock_execution_context = ExecutionContext(safe_mode=False)
-
-                # Create a proper mock execution processor for agent mode
-                from collections import defaultdict
-
-                mock_execution_processor = AsyncMock()
-                mock_execution_processor.execution_stats = MagicMock()
-                mock_execution_processor.execution_stats_lock = MagicMock()
-
-                # Create a mock NodeExecutionProgress for the sink node
-                mock_node_exec_progress = MagicMock()
-                mock_node_exec_progress.add_task = MagicMock()
-                mock_node_exec_progress.pop_output = MagicMock(
-                    return_value=None
-                )  # No outputs to process
-
-                # Set up running_node_execution as a defaultdict that returns our mock for any key
-                mock_execution_processor.running_node_execution = defaultdict(
-                    lambda: mock_node_exec_progress
-                )
-
-                # Mock the on_node_execution method that gets called during tool execution
-                mock_node_stats = MagicMock()
-                mock_node_stats.error = None
-                mock_execution_processor.on_node_execution.return_value = (
-                    mock_node_stats
-                )
-
-                async for output_name, output_value in block.run(
-                    input_data,
-                    credentials=llm.TEST_CREDENTIALS,
-                    graph_id="test_graph",
-                    node_id="test_node",
-                    graph_exec_id="test_exec",
-                    node_exec_id="test_node_exec",
-                    user_id="test_user",
-                    graph_version=1,
-                    execution_context=mock_execution_context,
-                    execution_processor=mock_execution_processor,
-                ):
-                    outputs[output_name] = output_value
-
-                # Verify we had at least 1 LLM call
-                assert call_count >= 1
-
-                # Check the final conversation output
-                final_conversation = outputs.get("conversations", [])
-
-                # The final conversation should NOT contain validation error messages
-                # Even if retries don't happen in agent mode, we should not leak errors
-                error_messages = [
-                    msg
-                    for msg in final_conversation
-                    if msg.get("role") == "user"
-                    and "parameter errors" in msg.get("content", "")
-                ]
-                assert (
-                    len(error_messages) == 0
-                ), "Validation error leaked into final conversation"
+            # The final conversation should only have the successful response
+            assert final_conversation[-1]["content"] == "valid"

autogpt_platform/backend/backend/blocks/wordpress/__init__.py

@@ -1,3 +1,3 @@
-from .blog import WordPressCreatePostBlock, WordPressGetAllPostsBlock
+from .blog import WordPressCreatePostBlock
 
-__all__ = ["WordPressCreatePostBlock", "WordPressGetAllPostsBlock"]
+__all__ = ["WordPressCreatePostBlock"]

autogpt_platform/backend/backend/blocks/wordpress/_api.py

@@ -161,7 +161,7 @@ async def oauth_exchange_code_for_tokens(
         grant_type="authorization_code",
     ).model_dump(exclude_none=True)
 
-    response = await Requests(raise_for_status=False).post(
+    response = await Requests().post(
         f"{WORDPRESS_BASE_URL}oauth2/token",
         headers=headers,
         data=data,
@@ -205,7 +205,7 @@ async def oauth_refresh_tokens(
         grant_type="refresh_token",
     ).model_dump(exclude_none=True)
 
-    response = await Requests(raise_for_status=False).post(
+    response = await Requests().post(
         f"{WORDPRESS_BASE_URL}oauth2/token",
         headers=headers,
         data=data,
@@ -252,7 +252,7 @@ async def validate_token(
         "token": token,
     }
 
-    response = await Requests(raise_for_status=False).get(
+    response = await Requests().get(
         f"{WORDPRESS_BASE_URL}oauth2/token-info",
         params=params,
     )
@@ -296,7 +296,7 @@ async def make_api_request(
 
     url = f"{WORDPRESS_BASE_URL.rstrip('/')}{endpoint}"
 
-    request_method = getattr(Requests(raise_for_status=False), method.lower())
+    request_method = getattr(Requests(), method.lower())
     response = await request_method(
         url,
         headers=headers,
@@ -476,7 +476,6 @@ async def create_post(
         data["tags"] = ",".join(str(t) for t in data["tags"])
 
     # Make the API request
-    site = normalize_site(site)
     endpoint = f"/rest/v1.1/sites/{site}/posts/new"
 
     headers = {
@@ -484,7 +483,7 @@ async def create_post(
         "Content-Type": "application/x-www-form-urlencoded",
     }
 
-    response = await Requests(raise_for_status=False).post(
+    response = await Requests().post(
         f"{WORDPRESS_BASE_URL.rstrip('/')}{endpoint}",
         headers=headers,
         data=data,
@@ -500,132 +499,3 @@ async def create_post(
     )
     error_message = error_data.get("message", response.text)
     raise ValueError(f"Failed to create post: {response.status} - {error_message}")
-
-
-class Post(BaseModel):
-    """Response model for individual posts in a posts list response.
-
-    This is a simplified version compared to PostResponse, as the list endpoint
-    returns less detailed information than the create/get single post endpoints.
-    """
-
-    ID: int
-    site_ID: int
-    author: PostAuthor
-    date: datetime
-    modified: datetime
-    title: str
-    URL: str
-    short_URL: str
-    content: str | None = None
-    excerpt: str | None = None
-    slug: str
-    guid: str
-    status: str
-    sticky: bool
-    password: str | None = ""
-    parent: Union[Dict[str, Any], bool, None] = None
-    type: str
-    discussion: Dict[str, Union[str, bool, int]] | None = None
-    likes_enabled: bool | None = None
-    sharing_enabled: bool | None = None
-    like_count: int | None = None
-    i_like: bool | None = None
-    is_reblogged: bool | None = None
-    is_following: bool | None = None
-    global_ID: str | None = None
-    featured_image: str | None = None
-    post_thumbnail: Dict[str, Any] | None = None
-    format: str | None = None
-    geo: Union[Dict[str, Any], bool, None] = None
-    menu_order: int | None = None
-    page_template: str | None = None
-    publicize_URLs: List[str] | None = None
-    terms: Dict[str, Dict[str, Any]] | None = None
-    tags: Dict[str, Dict[str, Any]] | None = None
-    categories: Dict[str, Dict[str, Any]] | None = None
-    attachments: Dict[str, Dict[str, Any]] | None = None
-    attachment_count: int | None = None
-    metadata: List[Dict[str, Any]] | None = None
-    meta: Dict[str, Any] | None = None
-    capabilities: Dict[str, bool] | None = None
-    revisions: List[int] | None = None
-    other_URLs: Dict[str, Any] | None = None
-
-
-class PostsResponse(BaseModel):
-    """Response model for WordPress posts list."""
-
-    found: int
-    posts: List[Post]
-    meta: Dict[str, Any]
-
-
-def normalize_site(site: str) -> str:
-    """
-    Normalize a site identifier by stripping protocol and trailing slashes.
-
-    Args:
-        site: Site URL, domain, or ID (e.g., "https://myblog.wordpress.com/", "myblog.wordpress.com", "123456789")
-
-    Returns:
-        Normalized site identifier (domain or ID only)
-    """
-    site = site.strip()
-    if site.startswith("https://"):
-        site = site[8:]
-    elif site.startswith("http://"):
-        site = site[7:]
-    return site.rstrip("/")
-
-
-async def get_posts(
-    credentials: Credentials,
-    site: str,
-    status: PostStatus | None = None,
-    number: int = 100,
-    offset: int = 0,
-) -> PostsResponse:
-    """
-    Get posts from a WordPress site.
-
-    Args:
-        credentials: OAuth credentials
-        site: Site ID or domain (e.g., "myblog.wordpress.com" or "123456789")
-        status: Filter by post status using PostStatus enum, or None for all
-        number: Number of posts to retrieve (max 100)
-        offset: Number of posts to skip (for pagination)
-
-    Returns:
-        PostsResponse with the list of posts
-    """
-    site = normalize_site(site)
-    endpoint = f"/rest/v1.1/sites/{site}/posts"
-
-    headers = {
-        "Authorization": credentials.auth_header(),
-    }
-
-    params: Dict[str, Any] = {
-        "number": max(1, min(number, 100)),  # 1–100 posts per request
-        "offset": offset,
-    }
-
-    if status:
-        params["status"] = status.value
-    response = await Requests(raise_for_status=False).get(
-        f"{WORDPRESS_BASE_URL.rstrip('/')}{endpoint}",
-        headers=headers,
-        params=params,
-    )
-
-    if response.ok:
-        return PostsResponse.model_validate(response.json())
-
-    error_data = (
-        response.json()
-        if response.headers.get("content-type", "").startswith("application/json")
-        else {}
-    )
-    error_message = error_data.get("message", response.text)
-    raise ValueError(f"Failed to get posts: {response.status} - {error_message}")

autogpt_platform/backend/backend/blocks/wordpress/blog.py

@@ -9,15 +9,7 @@ from backend.sdk import (
     SchemaField,
 )
 
-from ._api import (
-    CreatePostRequest,
-    Post,
-    PostResponse,
-    PostsResponse,
-    PostStatus,
-    create_post,
-    get_posts,
-)
+from ._api import CreatePostRequest, PostResponse, PostStatus, create_post
 from ._config import wordpress
 
 
@@ -57,15 +49,8 @@ class WordPressCreatePostBlock(Block):
         media_urls: list[str] = SchemaField(
             description="URLs of images to sideload and attach to the post", default=[]
         )
-        publish_as_draft: bool = SchemaField(
-            description="If True, publishes the post as a draft. If False, publishes it publicly.",
-            default=False,
-        )
 
     class Output(BlockSchemaOutput):
-        site: str = SchemaField(
-            description="The site ID or domain (pass-through for chaining with other blocks)"
-        )
         post_id: int = SchemaField(description="The ID of the created post")
         post_url: str = SchemaField(description="The full URL of the created post")
         short_url: str = SchemaField(description="The shortened wp.me URL")
@@ -93,9 +78,7 @@ class WordPressCreatePostBlock(Block):
             tags=input_data.tags,
             featured_image=input_data.featured_image,
             media_urls=input_data.media_urls,
-            status=(
-                PostStatus.DRAFT if input_data.publish_as_draft else PostStatus.PUBLISH
-            ),
+            status=PostStatus.PUBLISH,
         )
 
         post_response: PostResponse = await create_post(
@@ -104,69 +87,7 @@ class WordPressCreatePostBlock(Block):
             post_data=post_request,
         )
 
-        yield "site", input_data.site
         yield "post_id", post_response.ID
         yield "post_url", post_response.URL
         yield "short_url", post_response.short_URL
         yield "post_data", post_response.model_dump()
-
-
-class WordPressGetAllPostsBlock(Block):
-    """
-    Fetches all posts from a WordPress.com site or Jetpack-enabled site.
-    Supports filtering by status and pagination.
-    """
-
-    class Input(BlockSchemaInput):
-        credentials: CredentialsMetaInput = wordpress.credentials_field()
-        site: str = SchemaField(
-            description="Site ID or domain (e.g., 'myblog.wordpress.com' or '123456789')"
-        )
-        status: PostStatus | None = SchemaField(
-            description="Filter by post status, or None for all",
-            default=None,
-        )
-        number: int = SchemaField(
-            description="Number of posts to retrieve (max 100 per request)", default=20
-        )
-        offset: int = SchemaField(
-            description="Number of posts to skip (for pagination)", default=0
-        )
-
-    class Output(BlockSchemaOutput):
-        site: str = SchemaField(
-            description="The site ID or domain (pass-through for chaining with other blocks)"
-        )
-        found: int = SchemaField(description="Total number of posts found")
-        posts: list[Post] = SchemaField(
-            description="List of post objects with their details"
-        )
-        post: Post = SchemaField(
-            description="Individual post object (yielded for each post)"
-        )
-
-    def __init__(self):
-        super().__init__(
-            id="97728fa7-7f6f-4789-ba0c-f2c114119536",
-            description="Fetch all posts from WordPress.com or Jetpack sites",
-            categories={BlockCategory.SOCIAL},
-            input_schema=self.Input,
-            output_schema=self.Output,
-        )
-
-    async def run(
-        self, input_data: Input, *, credentials: Credentials, **kwargs
-    ) -> BlockOutput:
-        posts_response: PostsResponse = await get_posts(
-            credentials=credentials,
-            site=input_data.site,
-            status=input_data.status,
-            number=input_data.number,
-            offset=input_data.offset,
-        )
-
-        yield "site", input_data.site
-        yield "found", posts_response.found
-        yield "posts", posts_response.posts
-        for post in posts_response.posts:
-            yield "post", post

autogpt_platform/backend/backend/blocks/xml_parser.py

@@ -1,5 +1,5 @@
 from gravitasml.parser import Parser
-from gravitasml.token import Token, tokenize
+from gravitasml.token import tokenize
 
 from backend.data.block import Block, BlockOutput, BlockSchemaInput, BlockSchemaOutput
 from backend.data.model import SchemaField
@@ -25,38 +25,6 @@ class XMLParserBlock(Block):
             ],
         )
 
-    @staticmethod
-    def _validate_tokens(tokens: list[Token]) -> None:
-        """Ensure the XML has a single root element and no stray text."""
-        if not tokens:
-            raise ValueError("XML input is empty.")
-
-        depth = 0
-        root_seen = False
-
-        for token in tokens:
-            if token.type == "TAG_OPEN":
-                if depth == 0 and root_seen:
-                    raise ValueError("XML must have a single root element.")
-                depth += 1
-                if depth == 1:
-                    root_seen = True
-            elif token.type == "TAG_CLOSE":
-                depth -= 1
-                if depth < 0:
-                    raise SyntaxError("Unexpected closing tag in XML input.")
-            elif token.type in {"TEXT", "ESCAPE"}:
-                if depth == 0 and token.value:
-                    raise ValueError(
-                        "XML contains text outside the root element; "
-                        "wrap content in a single root tag."
-                    )
-
-        if depth != 0:
-            raise SyntaxError("Unclosed tag detected in XML input.")
-        if not root_seen:
-            raise ValueError("XML must include a root element.")
-
     async def run(self, input_data: Input, **kwargs) -> BlockOutput:
         # Security fix: Add size limits to prevent XML bomb attacks
         MAX_XML_SIZE = 10 * 1024 * 1024  # 10MB limit for XML input
@@ -67,9 +35,7 @@ class XMLParserBlock(Block):
             )
 
         try:
-            tokens = list(tokenize(input_data.input_xml))
-            self._validate_tokens(tokens)
-
+            tokens = tokenize(input_data.input_xml)
             parser = Parser(tokens)
             parsed_result = parser.parse()
             yield "parsed_xml", parsed_result

autogpt_platform/backend/backend/blocks/youtube.py

@@ -111,8 +111,6 @@ class TranscribeYoutubeVideoBlock(Block):
                 return parsed_url.path.split("/")[2]
             if parsed_url.path[:3] == "/v/":
                 return parsed_url.path.split("/")[2]
-            if parsed_url.path.startswith("/shorts/"):
-                return parsed_url.path.split("/")[2]
         raise ValueError(f"Invalid YouTube URL: {url}")
 
     def get_transcript(

autogpt_platform/backend/backend/cli.py

@@ -244,7 +244,11 @@ def websocket(server_address: str, graph_exec_id: str):
 
     import websockets.asyncio.client
 
-    from backend.api.ws_api import WSMessage, WSMethod, WSSubscribeGraphExecutionRequest
+    from backend.server.ws_api import (
+        WSMessage,
+        WSMethod,
+        WSSubscribeGraphExecutionRequest,
+    )
 
     async def send_message(server_address: str):
         uri = f"ws://{server_address}"

autogpt_platform/backend/backend/cli/__init__.py

@@ -1 +0,0 @@
-"""CLI utilities for backend development & administration"""

autogpt_platform/backend/backend/cli/generate_openapi_json.py

@@ -1,57 +0,0 @@
-#!/usr/bin/env python3
-"""
-Script to generate OpenAPI JSON specification for the FastAPI app.
-
-This script imports the FastAPI app from backend.api.rest_api and outputs
-the OpenAPI specification as JSON to stdout or a specified file.
-
-Usage:
-  `poetry run python generate_openapi_json.py`
-  `poetry run python generate_openapi_json.py --output openapi.json`
-  `poetry run python generate_openapi_json.py --indent 4 --output openapi.json`
-"""
-
-import json
-import os
-from pathlib import Path
-
-import click
-
-
-@click.command()
-@click.option(
-    "--output",
-    type=click.Path(dir_okay=False, path_type=Path),
-    help="Output file path (default: stdout)",
-)
-@click.option(
-    "--pretty",
-    type=click.BOOL,
-    default=False,
-    help="Pretty-print JSON output (indented 2 spaces)",
-)
-def main(output: Path, pretty: bool):
-    """Generate and output the OpenAPI JSON specification."""
-    openapi_schema = get_openapi_schema()
-
-    json_output = json.dumps(openapi_schema, indent=2 if pretty else None)
-
-    if output:
-        output.write_text(json_output)
-        click.echo(f"✅ OpenAPI specification written to {output}\n\nPreview:")
-        click.echo(f"\n{json_output[:500]} ...")
-    else:
-        print(json_output)
-
-
-def get_openapi_schema():
-    """Get the OpenAPI schema from the FastAPI app"""
-    from backend.api.rest_api import app
-
-    return app.openapi()
-
-
-if __name__ == "__main__":
-    os.environ["LOG_LEVEL"] = "ERROR"  # disable stdout log output
-
-    main()

autogpt_platform/backend/backend/data/__init__.py

@@ -1,4 +1,4 @@
-from backend.api.features.library.model import LibraryAgentPreset
+from backend.server.v2.library.model import LibraryAgentPreset
 
 from .graph import NodeModel
 from .integrations import Webhook  # noqa: F401

autogpt_platform/backend/backend/data/api_key.py

@@ -1,24 +1,22 @@
 import logging
 import uuid
 from datetime import datetime, timezone
-from typing import Literal, Optional
+from typing import Optional
 
 from autogpt_libs.api_key.keysmith import APIKeySmith
 from prisma.enums import APIKeyPermission, APIKeyStatus
 from prisma.models import APIKey as PrismaAPIKey
 from prisma.types import APIKeyWhereUniqueInput
-from pydantic import Field
+from pydantic import BaseModel, Field
 
 from backend.data.includes import MAX_USER_API_KEYS_FETCH
 from backend.util.exceptions import NotAuthorizedError, NotFoundError
 
-from .base import APIAuthorizationInfo
-
 logger = logging.getLogger(__name__)
 keysmith = APIKeySmith()
 
 
-class APIKeyInfo(APIAuthorizationInfo):
+class APIKeyInfo(BaseModel):
     id: str
     name: str
     head: str = Field(
@@ -28,9 +26,12 @@ class APIKeyInfo(APIAuthorizationInfo):
         description=f"The last {APIKeySmith.TAIL_LENGTH} characters of the key"
     )
     status: APIKeyStatus
+    permissions: list[APIKeyPermission]
+    created_at: datetime
+    last_used_at: Optional[datetime] = None
+    revoked_at: Optional[datetime] = None
     description: Optional[str] = None
-
-    type: Literal["api_key"] = "api_key"  # type: ignore
+    user_id: str
 
     @staticmethod
     def from_db(api_key: PrismaAPIKey):
@@ -40,7 +41,7 @@ class APIKeyInfo(APIAuthorizationInfo):
             head=api_key.head,
             tail=api_key.tail,
             status=APIKeyStatus(api_key.status),
-            scopes=[APIKeyPermission(p) for p in api_key.permissions],
+            permissions=[APIKeyPermission(p) for p in api_key.permissions],
             created_at=api_key.createdAt,
             last_used_at=api_key.lastUsedAt,
             revoked_at=api_key.revokedAt,
@@ -210,7 +211,7 @@ async def suspend_api_key(key_id: str, user_id: str) -> APIKeyInfo:
 
 
 def has_permission(api_key: APIKeyInfo, required_permission: APIKeyPermission) -> bool:
-    return required_permission in api_key.scopes
+    return required_permission in api_key.permissions
 
 
 async def get_api_key_by_id(key_id: str, user_id: str) -> Optional[APIKeyInfo]:

autogpt_platform/backend/backend/data/auth/base.py

@@ -1,15 +0,0 @@
-from datetime import datetime
-from typing import Literal, Optional
-
-from prisma.enums import APIKeyPermission
-from pydantic import BaseModel
-
-
-class APIAuthorizationInfo(BaseModel):
-    user_id: str
-    scopes: list[APIKeyPermission]
-    type: Literal["oauth", "api_key"]
-    created_at: datetime
-    expires_at: Optional[datetime] = None
-    last_used_at: Optional[datetime] = None
-    revoked_at: Optional[datetime] = None

autogpt_platform/backend/backend/data/auth/oauth.py

@@ -1,872 +0,0 @@
-"""
-OAuth 2.0 Provider Data Layer
-
-Handles management of OAuth applications, authorization codes,
-access tokens, and refresh tokens.
-
-Hashing strategy:
-- Access tokens & Refresh tokens: SHA256 (deterministic, allows direct lookup by hash)
-- Client secrets: Scrypt with salt (lookup by client_id, then verify with salt)
-"""
-
-import hashlib
-import logging
-import secrets
-import uuid
-from datetime import datetime, timedelta, timezone
-from typing import Literal, Optional
-
-from autogpt_libs.api_key.keysmith import APIKeySmith
-from prisma.enums import APIKeyPermission as APIPermission
-from prisma.models import OAuthAccessToken as PrismaOAuthAccessToken
-from prisma.models import OAuthApplication as PrismaOAuthApplication
-from prisma.models import OAuthAuthorizationCode as PrismaOAuthAuthorizationCode
-from prisma.models import OAuthRefreshToken as PrismaOAuthRefreshToken
-from prisma.types import OAuthApplicationUpdateInput
-from pydantic import BaseModel, Field, SecretStr
-
-from .base import APIAuthorizationInfo
-
-logger = logging.getLogger(__name__)
-keysmith = APIKeySmith()  # Only used for client secret hashing (Scrypt)
-
-
-def _generate_token() -> str:
-    """Generate a cryptographically secure random token."""
-    return secrets.token_urlsafe(32)
-
-
-def _hash_token(token: str) -> str:
-    """Hash a token using SHA256 (deterministic, for direct lookup)."""
-    return hashlib.sha256(token.encode()).hexdigest()
-
-
-# Token TTLs
-AUTHORIZATION_CODE_TTL = timedelta(minutes=10)
-ACCESS_TOKEN_TTL = timedelta(hours=1)
-REFRESH_TOKEN_TTL = timedelta(days=30)
-
-ACCESS_TOKEN_PREFIX = "agpt_xt_"
-REFRESH_TOKEN_PREFIX = "agpt_rt_"
-
-
-# ============================================================================
-# Exception Classes
-# ============================================================================
-
-
-class OAuthError(Exception):
-    """Base OAuth error"""
-
-    pass
-
-
-class InvalidClientError(OAuthError):
-    """Invalid client_id or client_secret"""
-
-    pass
-
-
-class InvalidGrantError(OAuthError):
-    """Invalid or expired authorization code/refresh token"""
-
-    def __init__(self, reason: str):
-        self.reason = reason
-        super().__init__(f"Invalid grant: {reason}")
-
-
-class InvalidTokenError(OAuthError):
-    """Invalid, expired, or revoked token"""
-
-    def __init__(self, reason: str):
-        self.reason = reason
-        super().__init__(f"Invalid token: {reason}")
-
-
-# ============================================================================
-# Data Models
-# ============================================================================
-
-
-class OAuthApplicationInfo(BaseModel):
-    """OAuth application information (without client secret hash)"""
-
-    id: str
-    name: str
-    description: Optional[str] = None
-    logo_url: Optional[str] = None
-    client_id: str
-    redirect_uris: list[str]
-    grant_types: list[str]
-    scopes: list[APIPermission]
-    owner_id: str
-    is_active: bool
-    created_at: datetime
-    updated_at: datetime
-
-    @staticmethod
-    def from_db(app: PrismaOAuthApplication):
-        return OAuthApplicationInfo(
-            id=app.id,
-            name=app.name,
-            description=app.description,
-            logo_url=app.logoUrl,
-            client_id=app.clientId,
-            redirect_uris=app.redirectUris,
-            grant_types=app.grantTypes,
-            scopes=[APIPermission(s) for s in app.scopes],
-            owner_id=app.ownerId,
-            is_active=app.isActive,
-            created_at=app.createdAt,
-            updated_at=app.updatedAt,
-        )
-
-
-class OAuthApplicationInfoWithSecret(OAuthApplicationInfo):
-    """OAuth application with client secret hash (for validation)"""
-
-    client_secret_hash: str
-    client_secret_salt: str
-
-    @staticmethod
-    def from_db(app: PrismaOAuthApplication):
-        return OAuthApplicationInfoWithSecret(
-            **OAuthApplicationInfo.from_db(app).model_dump(),
-            client_secret_hash=app.clientSecret,
-            client_secret_salt=app.clientSecretSalt,
-        )
-
-    def verify_secret(self, plaintext_secret: str) -> bool:
-        """Verify a plaintext client secret against the stored hash"""
-        # Use keysmith.verify_key() with stored salt
-        return keysmith.verify_key(
-            plaintext_secret, self.client_secret_hash, self.client_secret_salt
-        )
-
-
-class OAuthAuthorizationCodeInfo(BaseModel):
-    """Authorization code information"""
-
-    id: str
-    code: str
-    created_at: datetime
-    expires_at: datetime
-    application_id: str
-    user_id: str
-    scopes: list[APIPermission]
-    redirect_uri: str
-    code_challenge: Optional[str] = None
-    code_challenge_method: Optional[str] = None
-    used_at: Optional[datetime] = None
-
-    @property
-    def is_used(self) -> bool:
-        return self.used_at is not None
-
-    @staticmethod
-    def from_db(code: PrismaOAuthAuthorizationCode):
-        return OAuthAuthorizationCodeInfo(
-            id=code.id,
-            code=code.code,
-            created_at=code.createdAt,
-            expires_at=code.expiresAt,
-            application_id=code.applicationId,
-            user_id=code.userId,
-            scopes=[APIPermission(s) for s in code.scopes],
-            redirect_uri=code.redirectUri,
-            code_challenge=code.codeChallenge,
-            code_challenge_method=code.codeChallengeMethod,
-            used_at=code.usedAt,
-        )
-
-
-class OAuthAccessTokenInfo(APIAuthorizationInfo):
-    """Access token information"""
-
-    id: str
-    expires_at: datetime  # type: ignore
-    application_id: str
-
-    type: Literal["oauth"] = "oauth"  # type: ignore
-
-    @staticmethod
-    def from_db(token: PrismaOAuthAccessToken):
-        return OAuthAccessTokenInfo(
-            id=token.id,
-            user_id=token.userId,
-            scopes=[APIPermission(s) for s in token.scopes],
-            created_at=token.createdAt,
-            expires_at=token.expiresAt,
-            last_used_at=None,
-            revoked_at=token.revokedAt,
-            application_id=token.applicationId,
-        )
-
-
-class OAuthAccessToken(OAuthAccessTokenInfo):
-    """Access token with plaintext token included (sensitive)"""
-
-    token: SecretStr = Field(description="Plaintext token (sensitive)")
-
-    @staticmethod
-    def from_db(token: PrismaOAuthAccessToken, plaintext_token: str):  # type: ignore
-        return OAuthAccessToken(
-            **OAuthAccessTokenInfo.from_db(token).model_dump(),
-            token=SecretStr(plaintext_token),
-        )
-
-
-class OAuthRefreshTokenInfo(BaseModel):
-    """Refresh token information"""
-
-    id: str
-    user_id: str
-    scopes: list[APIPermission]
-    created_at: datetime
-    expires_at: datetime
-    application_id: str
-    revoked_at: Optional[datetime] = None
-
-    @property
-    def is_revoked(self) -> bool:
-        return self.revoked_at is not None
-
-    @staticmethod
-    def from_db(token: PrismaOAuthRefreshToken):
-        return OAuthRefreshTokenInfo(
-            id=token.id,
-            user_id=token.userId,
-            scopes=[APIPermission(s) for s in token.scopes],
-            created_at=token.createdAt,
-            expires_at=token.expiresAt,
-            application_id=token.applicationId,
-            revoked_at=token.revokedAt,
-        )
-
-
-class OAuthRefreshToken(OAuthRefreshTokenInfo):
-    """Refresh token with plaintext token included (sensitive)"""
-
-    token: SecretStr = Field(description="Plaintext token (sensitive)")
-
-    @staticmethod
-    def from_db(token: PrismaOAuthRefreshToken, plaintext_token: str):  # type: ignore
-        return OAuthRefreshToken(
-            **OAuthRefreshTokenInfo.from_db(token).model_dump(),
-            token=SecretStr(plaintext_token),
-        )
-
-
-class TokenIntrospectionResult(BaseModel):
-    """Result of token introspection (RFC 7662)"""
-
-    active: bool
-    scopes: Optional[list[str]] = None
-    client_id: Optional[str] = None
-    user_id: Optional[str] = None
-    exp: Optional[int] = None  # Unix timestamp
-    token_type: Optional[Literal["access_token", "refresh_token"]] = None
-
-
-# ============================================================================
-# OAuth Application Management
-# ============================================================================
-
-
-async def get_oauth_application(client_id: str) -> Optional[OAuthApplicationInfo]:
-    """Get OAuth application by client ID (without secret)"""
-    app = await PrismaOAuthApplication.prisma().find_unique(
-        where={"clientId": client_id}
-    )
-    if not app:
-        return None
-    return OAuthApplicationInfo.from_db(app)
-
-
-async def get_oauth_application_with_secret(
-    client_id: str,
-) -> Optional[OAuthApplicationInfoWithSecret]:
-    """Get OAuth application by client ID (with secret hash for validation)"""
-    app = await PrismaOAuthApplication.prisma().find_unique(
-        where={"clientId": client_id}
-    )
-    if not app:
-        return None
-    return OAuthApplicationInfoWithSecret.from_db(app)
-
-
-async def validate_client_credentials(
-    client_id: str, client_secret: str
-) -> OAuthApplicationInfo:
-    """
-    Validate client credentials and return application info.
-
-    Raises:
-        InvalidClientError: If client_id or client_secret is invalid, or app is inactive
-    """
-    app = await get_oauth_application_with_secret(client_id)
-    if not app:
-        raise InvalidClientError("Invalid client_id")
-
-    if not app.is_active:
-        raise InvalidClientError("Application is not active")
-
-    # Verify client secret
-    if not app.verify_secret(client_secret):
-        raise InvalidClientError("Invalid client_secret")
-
-    # Return without secret hash
-    return OAuthApplicationInfo(**app.model_dump(exclude={"client_secret_hash"}))
-
-
-def validate_redirect_uri(app: OAuthApplicationInfo, redirect_uri: str) -> bool:
-    """Validate that redirect URI is registered for the application"""
-    return redirect_uri in app.redirect_uris
-
-
-def validate_scopes(
-    app: OAuthApplicationInfo, requested_scopes: list[APIPermission]
-) -> bool:
-    """Validate that all requested scopes are allowed for the application"""
-    return all(scope in app.scopes for scope in requested_scopes)
-
-
-# ============================================================================
-# Authorization Code Flow
-# ============================================================================
-
-
-def _generate_authorization_code() -> str:
-    """Generate a cryptographically secure authorization code"""
-    # 32 bytes = 256 bits of entropy
-    return secrets.token_urlsafe(32)
-
-
-async def create_authorization_code(
-    application_id: str,
-    user_id: str,
-    scopes: list[APIPermission],
-    redirect_uri: str,
-    code_challenge: Optional[str] = None,
-    code_challenge_method: Optional[Literal["S256", "plain"]] = None,
-) -> OAuthAuthorizationCodeInfo:
-    """
-    Create a new authorization code.
-    Expires in 10 minutes and can only be used once.
-    """
-    code = _generate_authorization_code()
-    now = datetime.now(timezone.utc)
-    expires_at = now + AUTHORIZATION_CODE_TTL
-
-    saved_code = await PrismaOAuthAuthorizationCode.prisma().create(
-        data={
-            "id": str(uuid.uuid4()),
-            "code": code,
-            "expiresAt": expires_at,
-            "applicationId": application_id,
-            "userId": user_id,
-            "scopes": [s for s in scopes],
-            "redirectUri": redirect_uri,
-            "codeChallenge": code_challenge,
-            "codeChallengeMethod": code_challenge_method,
-        }
-    )
-
-    return OAuthAuthorizationCodeInfo.from_db(saved_code)
-
-
-async def consume_authorization_code(
-    code: str,
-    application_id: str,
-    redirect_uri: str,
-    code_verifier: Optional[str] = None,
-) -> tuple[str, list[APIPermission]]:
-    """
-    Consume an authorization code and return (user_id, scopes).
-
-    This marks the code as used and validates:
-    - Code exists and matches application
-    - Code is not expired
-    - Code has not been used
-    - Redirect URI matches
-    - PKCE code verifier matches (if code challenge was provided)
-
-    Raises:
-        InvalidGrantError: If code is invalid, expired, used, or PKCE fails
-    """
-    auth_code = await PrismaOAuthAuthorizationCode.prisma().find_unique(
-        where={"code": code}
-    )
-
-    if not auth_code:
-        raise InvalidGrantError("authorization code not found")
-
-    # Validate application
-    if auth_code.applicationId != application_id:
-        raise InvalidGrantError(
-            "authorization code does not belong to this application"
-        )
-
-    # Check if already used
-    if auth_code.usedAt is not None:
-        raise InvalidGrantError(
-            f"authorization code already used at {auth_code.usedAt}"
-        )
-
-    # Check expiration
-    now = datetime.now(timezone.utc)
-    if auth_code.expiresAt < now:
-        raise InvalidGrantError("authorization code expired")
-
-    # Validate redirect URI
-    if auth_code.redirectUri != redirect_uri:
-        raise InvalidGrantError("redirect_uri mismatch")
-
-    # Validate PKCE if code challenge was provided
-    if auth_code.codeChallenge:
-        if not code_verifier:
-            raise InvalidGrantError("code_verifier required but not provided")
-
-        if not _verify_pkce(
-            code_verifier, auth_code.codeChallenge, auth_code.codeChallengeMethod
-        ):
-            raise InvalidGrantError("PKCE verification failed")
-
-    # Mark code as used
-    await PrismaOAuthAuthorizationCode.prisma().update(
-        where={"code": code},
-        data={"usedAt": now},
-    )
-
-    return auth_code.userId, [APIPermission(s) for s in auth_code.scopes]
-
-
-def _verify_pkce(
-    code_verifier: str, code_challenge: str, code_challenge_method: Optional[str]
-) -> bool:
-    """
-    Verify PKCE code verifier against code challenge.
-
-    Supports:
-    - S256: SHA256(code_verifier) == code_challenge
-    - plain: code_verifier == code_challenge
-    """
-    if code_challenge_method == "S256":
-        # Hash the verifier with SHA256 and base64url encode
-        hashed = hashlib.sha256(code_verifier.encode("ascii")).digest()
-        computed_challenge = (
-            secrets.token_urlsafe(len(hashed)).encode("ascii").decode("ascii")
-        )
-        # For proper base64url encoding
-        import base64
-
-        computed_challenge = (
-            base64.urlsafe_b64encode(hashed).decode("ascii").rstrip("=")
-        )
-        return secrets.compare_digest(computed_challenge, code_challenge)
-    elif code_challenge_method == "plain" or code_challenge_method is None:
-        # Plain comparison
-        return secrets.compare_digest(code_verifier, code_challenge)
-    else:
-        logger.warning(f"Unsupported code challenge method: {code_challenge_method}")
-        return False
-
-
-# ============================================================================
-# Access Token Management
-# ============================================================================
-
-
-async def create_access_token(
-    application_id: str, user_id: str, scopes: list[APIPermission]
-) -> OAuthAccessToken:
-    """
-    Create a new access token.
-    Returns OAuthAccessToken (with plaintext token).
-    """
-    plaintext_token = ACCESS_TOKEN_PREFIX + _generate_token()
-    token_hash = _hash_token(plaintext_token)
-    now = datetime.now(timezone.utc)
-    expires_at = now + ACCESS_TOKEN_TTL
-
-    saved_token = await PrismaOAuthAccessToken.prisma().create(
-        data={
-            "id": str(uuid.uuid4()),
-            "token": token_hash,  # SHA256 hash for direct lookup
-            "expiresAt": expires_at,
-            "applicationId": application_id,
-            "userId": user_id,
-            "scopes": [s for s in scopes],
-        }
-    )
-
-    return OAuthAccessToken.from_db(saved_token, plaintext_token=plaintext_token)
-
-
-async def validate_access_token(
-    token: str,
-) -> tuple[OAuthAccessTokenInfo, OAuthApplicationInfo]:
-    """
-    Validate an access token and return token info.
-
-    Raises:
-        InvalidTokenError: If token is invalid, expired, or revoked
-        InvalidClientError: If the client application is not marked as active
-    """
-    token_hash = _hash_token(token)
-
-    # Direct lookup by hash
-    access_token = await PrismaOAuthAccessToken.prisma().find_unique(
-        where={"token": token_hash}, include={"Application": True}
-    )
-
-    if not access_token:
-        raise InvalidTokenError("access token not found")
-
-    if not access_token.Application:  # should be impossible
-        raise InvalidClientError("Client application not found")
-
-    if not access_token.Application.isActive:
-        raise InvalidClientError("Client application is disabled")
-
-    if access_token.revokedAt is not None:
-        raise InvalidTokenError("access token has been revoked")
-
-    # Check expiration
-    now = datetime.now(timezone.utc)
-    if access_token.expiresAt < now:
-        raise InvalidTokenError("access token expired")
-
-    return (
-        OAuthAccessTokenInfo.from_db(access_token),
-        OAuthApplicationInfo.from_db(access_token.Application),
-    )
-
-
-async def revoke_access_token(
-    token: str, application_id: str
-) -> OAuthAccessTokenInfo | None:
-    """
-    Revoke an access token.
-
-    Args:
-        token: The plaintext access token to revoke
-        application_id: The application ID making the revocation request.
-            Only tokens belonging to this application will be revoked.
-
-    Returns:
-        OAuthAccessTokenInfo if token was found and revoked, None otherwise.
-
-    Note:
-        Always performs exactly 2 DB queries regardless of outcome to prevent
-        timing side-channel attacks that could reveal token existence.
-    """
-    try:
-        token_hash = _hash_token(token)
-
-        # Use update_many to filter by both token and applicationId
-        updated_count = await PrismaOAuthAccessToken.prisma().update_many(
-            where={
-                "token": token_hash,
-                "applicationId": application_id,
-                "revokedAt": None,
-            },
-            data={"revokedAt": datetime.now(timezone.utc)},
-        )
-
-        # Always perform second query to ensure constant time
-        result = await PrismaOAuthAccessToken.prisma().find_unique(
-            where={"token": token_hash}
-        )
-
-        # Only return result if we actually revoked something
-        if updated_count == 0:
-            return None
-
-        return OAuthAccessTokenInfo.from_db(result) if result else None
-    except Exception as e:
-        logger.exception(f"Error revoking access token: {e}")
-        return None
-
-
-# ============================================================================
-# Refresh Token Management
-# ============================================================================
-
-
-async def create_refresh_token(
-    application_id: str, user_id: str, scopes: list[APIPermission]
-) -> OAuthRefreshToken:
-    """
-    Create a new refresh token.
-    Returns OAuthRefreshToken (with plaintext token).
-    """
-    plaintext_token = REFRESH_TOKEN_PREFIX + _generate_token()
-    token_hash = _hash_token(plaintext_token)
-    now = datetime.now(timezone.utc)
-    expires_at = now + REFRESH_TOKEN_TTL
-
-    saved_token = await PrismaOAuthRefreshToken.prisma().create(
-        data={
-            "id": str(uuid.uuid4()),
-            "token": token_hash,  # SHA256 hash for direct lookup
-            "expiresAt": expires_at,
-            "applicationId": application_id,
-            "userId": user_id,
-            "scopes": [s for s in scopes],
-        }
-    )
-
-    return OAuthRefreshToken.from_db(saved_token, plaintext_token=plaintext_token)
-
-
-async def refresh_tokens(
-    refresh_token: str, application_id: str
-) -> tuple[OAuthAccessToken, OAuthRefreshToken]:
-    """
-    Use a refresh token to create new access and refresh tokens.
-    Returns (new_access_token, new_refresh_token) both with plaintext tokens included.
-
-    Raises:
-        InvalidGrantError: If refresh token is invalid, expired, or revoked
-    """
-    token_hash = _hash_token(refresh_token)
-
-    # Direct lookup by hash
-    rt = await PrismaOAuthRefreshToken.prisma().find_unique(where={"token": token_hash})
-
-    if not rt:
-        raise InvalidGrantError("refresh token not found")
-
-    # NOTE: no need to check Application.isActive, this is checked by the token endpoint
-
-    if rt.revokedAt is not None:
-        raise InvalidGrantError("refresh token has been revoked")
-
-    # Validate application
-    if rt.applicationId != application_id:
-        raise InvalidGrantError("refresh token does not belong to this application")
-
-    # Check expiration
-    now = datetime.now(timezone.utc)
-    if rt.expiresAt < now:
-        raise InvalidGrantError("refresh token expired")
-
-    # Revoke old refresh token
-    await PrismaOAuthRefreshToken.prisma().update(
-        where={"token": token_hash},
-        data={"revokedAt": now},
-    )
-
-    # Create new access and refresh tokens with same scopes
-    scopes = [APIPermission(s) for s in rt.scopes]
-    new_access_token = await create_access_token(
-        rt.applicationId,
-        rt.userId,
-        scopes,
-    )
-    new_refresh_token = await create_refresh_token(
-        rt.applicationId,
-        rt.userId,
-        scopes,
-    )
-
-    return new_access_token, new_refresh_token
-
-
-async def revoke_refresh_token(
-    token: str, application_id: str
-) -> OAuthRefreshTokenInfo | None:
-    """
-    Revoke a refresh token.
-
-    Args:
-        token: The plaintext refresh token to revoke
-        application_id: The application ID making the revocation request.
-            Only tokens belonging to this application will be revoked.
-
-    Returns:
-        OAuthRefreshTokenInfo if token was found and revoked, None otherwise.
-
-    Note:
-        Always performs exactly 2 DB queries regardless of outcome to prevent
-        timing side-channel attacks that could reveal token existence.
-    """
-    try:
-        token_hash = _hash_token(token)
-
-        # Use update_many to filter by both token and applicationId
-        updated_count = await PrismaOAuthRefreshToken.prisma().update_many(
-            where={
-                "token": token_hash,
-                "applicationId": application_id,
-                "revokedAt": None,
-            },
-            data={"revokedAt": datetime.now(timezone.utc)},
-        )
-
-        # Always perform second query to ensure constant time
-        result = await PrismaOAuthRefreshToken.prisma().find_unique(
-            where={"token": token_hash}
-        )
-
-        # Only return result if we actually revoked something
-        if updated_count == 0:
-            return None
-
-        return OAuthRefreshTokenInfo.from_db(result) if result else None
-    except Exception as e:
-        logger.exception(f"Error revoking refresh token: {e}")
-        return None
-
-
-# ============================================================================
-# Token Introspection
-# ============================================================================
-
-
-async def introspect_token(
-    token: str,
-    token_type_hint: Optional[Literal["access_token", "refresh_token"]] = None,
-) -> TokenIntrospectionResult:
-    """
-    Introspect a token and return its metadata (RFC 7662).
-
-    Returns TokenIntrospectionResult with active=True and metadata if valid,
-    or active=False if the token is invalid/expired/revoked.
-    """
-    # Try as access token first (or if hint says "access_token")
-    if token_type_hint != "refresh_token":
-        try:
-            token_info, app = await validate_access_token(token)
-            return TokenIntrospectionResult(
-                active=True,
-                scopes=list(s.value for s in token_info.scopes),
-                client_id=app.client_id if app else None,
-                user_id=token_info.user_id,
-                exp=int(token_info.expires_at.timestamp()),
-                token_type="access_token",
-            )
-        except InvalidTokenError:
-            pass  # Try as refresh token
-
-    # Try as refresh token
-    token_hash = _hash_token(token)
-    refresh_token = await PrismaOAuthRefreshToken.prisma().find_unique(
-        where={"token": token_hash}
-    )
-
-    if refresh_token and refresh_token.revokedAt is None:
-        # Check if valid (not expired)
-        now = datetime.now(timezone.utc)
-        if refresh_token.expiresAt > now:
-            app = await get_oauth_application_by_id(refresh_token.applicationId)
-            return TokenIntrospectionResult(
-                active=True,
-                scopes=list(s for s in refresh_token.scopes),
-                client_id=app.client_id if app else None,
-                user_id=refresh_token.userId,
-                exp=int(refresh_token.expiresAt.timestamp()),
-                token_type="refresh_token",
-            )
-
-    # Token not found or inactive
-    return TokenIntrospectionResult(active=False)
-
-
-async def get_oauth_application_by_id(app_id: str) -> Optional[OAuthApplicationInfo]:
-    """Get OAuth application by ID"""
-    app = await PrismaOAuthApplication.prisma().find_unique(where={"id": app_id})
-    if not app:
-        return None
-    return OAuthApplicationInfo.from_db(app)
-
-
-async def list_user_oauth_applications(user_id: str) -> list[OAuthApplicationInfo]:
-    """Get all OAuth applications owned by a user"""
-    apps = await PrismaOAuthApplication.prisma().find_many(
-        where={"ownerId": user_id},
-        order={"createdAt": "desc"},
-    )
-    return [OAuthApplicationInfo.from_db(app) for app in apps]
-
-
-async def update_oauth_application(
-    app_id: str,
-    *,
-    owner_id: str,
-    is_active: Optional[bool] = None,
-    logo_url: Optional[str] = None,
-) -> Optional[OAuthApplicationInfo]:
-    """
-    Update OAuth application active status.
-    Only the owner can update their app's status.
-
-    Returns the updated app info, or None if app not found or not owned by user.
-    """
-    # First verify ownership
-    app = await PrismaOAuthApplication.prisma().find_first(
-        where={"id": app_id, "ownerId": owner_id}
-    )
-    if not app:
-        return None
-
-    patch: OAuthApplicationUpdateInput = {}
-    if is_active is not None:
-        patch["isActive"] = is_active
-    if logo_url:
-        patch["logoUrl"] = logo_url
-    if not patch:
-        return OAuthApplicationInfo.from_db(app)  # return unchanged
-
-    updated_app = await PrismaOAuthApplication.prisma().update(
-        where={"id": app_id},
-        data=patch,
-    )
-    return OAuthApplicationInfo.from_db(updated_app) if updated_app else None
-
-
-# ============================================================================
-# Token Cleanup
-# ============================================================================
-
-
-async def cleanup_expired_oauth_tokens() -> dict[str, int]:
-    """
-    Delete expired OAuth tokens from the database.
-
-    This removes:
-    - Expired authorization codes (10 min TTL)
-    - Expired access tokens (1 hour TTL)
-    - Expired refresh tokens (30 day TTL)
-
-    Returns a dict with counts of deleted tokens by type.
-    """
-    now = datetime.now(timezone.utc)
-
-    # Delete expired authorization codes
-    codes_result = await PrismaOAuthAuthorizationCode.prisma().delete_many(
-        where={"expiresAt": {"lt": now}}
-    )
-
-    # Delete expired access tokens
-    access_result = await PrismaOAuthAccessToken.prisma().delete_many(
-        where={"expiresAt": {"lt": now}}
-    )
-
-    # Delete expired refresh tokens
-    refresh_result = await PrismaOAuthRefreshToken.prisma().delete_many(
-        where={"expiresAt": {"lt": now}}
-    )
-
-    deleted = {
-        "authorization_codes": codes_result,
-        "access_tokens": access_result,
-        "refresh_tokens": refresh_result,
-    }
-
-    total = sum(deleted.values())
-    if total > 0:
-        logger.info(f"Cleaned up {total} expired OAuth tokens: {deleted}")
-
-    return deleted

autogpt_platform/backend/backend/data/block.py

@@ -50,8 +50,6 @@ from .model import (
 logger = logging.getLogger(__name__)
 
 if TYPE_CHECKING:
-    from backend.data.execution import ExecutionContext
-
     from .graph import Link
 
 app_config = Config()
@@ -474,7 +472,6 @@ class Block(ABC, Generic[BlockSchemaInputType, BlockSchemaOutputType]):
         self.block_type = block_type
         self.webhook_config = webhook_config
         self.execution_stats: NodeExecutionStats = NodeExecutionStats()
-        self.requires_human_review: bool = False
 
         if self.webhook_config:
             if isinstance(self.webhook_config, BlockWebhookConfig):
@@ -617,88 +614,7 @@ class Block(ABC, Generic[BlockSchemaInputType, BlockSchemaOutputType]):
                     block_id=self.id,
                 ) from ex
 
-    async def is_block_exec_need_review(
-        self,
-        input_data: BlockInput,
-        *,
-        user_id: str,
-        node_exec_id: str,
-        graph_exec_id: str,
-        graph_id: str,
-        graph_version: int,
-        execution_context: "ExecutionContext",
-        **kwargs,
-    ) -> tuple[bool, BlockInput]:
-        """
-        Check if this block execution needs human review and handle the review process.
-
-        Returns:
-            Tuple of (should_pause, input_data_to_use)
-            - should_pause: True if execution should be paused for review
-            - input_data_to_use: The input data to use (may be modified by reviewer)
-        """
-        # Skip review if not required or safe mode is disabled
-        if not self.requires_human_review or not execution_context.safe_mode:
-            return False, input_data
-
-        from backend.blocks.helpers.review import HITLReviewHelper
-
-        # Handle the review request and get decision
-        decision = await HITLReviewHelper.handle_review_decision(
-            input_data=input_data,
-            user_id=user_id,
-            node_exec_id=node_exec_id,
-            graph_exec_id=graph_exec_id,
-            graph_id=graph_id,
-            graph_version=graph_version,
-            execution_context=execution_context,
-            block_name=self.name,
-            editable=True,
-        )
-
-        if decision is None:
-            # We're awaiting review - pause execution
-            return True, input_data
-
-        if not decision.should_proceed:
-            # Review was rejected, raise an error to stop execution
-            raise BlockExecutionError(
-                message=f"Block execution rejected by reviewer: {decision.message}",
-                block_name=self.name,
-                block_id=self.id,
-            )
-
-        # Review was approved - use the potentially modified data
-        # ReviewResult.data must be a dict for block inputs
-        reviewed_data = decision.review_result.data
-        if not isinstance(reviewed_data, dict):
-            raise BlockExecutionError(
-                message=f"Review data must be a dict for block input, got {type(reviewed_data).__name__}",
-                block_name=self.name,
-                block_id=self.id,
-            )
-        return False, reviewed_data
-
     async def _execute(self, input_data: BlockInput, **kwargs) -> BlockOutput:
-        # Check for review requirement only if running within a graph execution context
-        # Direct block execution (e.g., from chat) skips the review process
-        has_graph_context = all(
-            key in kwargs
-            for key in (
-                "node_exec_id",
-                "graph_exec_id",
-                "graph_id",
-                "execution_context",
-            )
-        )
-        if has_graph_context:
-            should_pause, input_data = await self.is_block_exec_need_review(
-                input_data, **kwargs
-            )
-            if should_pause:
-                return
-
-        # Validate the input data (original or reviewer-modified) once
         if error := self.input_schema.validate_data(input_data):
             raise BlockInputError(
                 message=f"Unable to execute block with invalid input data: {error}",
@@ -706,7 +622,6 @@ class Block(ABC, Generic[BlockSchemaInputType, BlockSchemaOutputType]):
                 block_id=self.id,
             )
 
-        # Use the validated input data
         async for output_name, output_data in self.run(
             self.input_schema(**{k: v for k, v in input_data.items() if v is not None}),
             **kwargs,

autogpt_platform/backend/backend/data/block_cost_config.py

@@ -59,13 +59,12 @@ from backend.integrations.credentials_store import (
 
 MODEL_COST: dict[LlmModel, int] = {
     LlmModel.O3: 4,
-    LlmModel.O3_MINI: 2,
-    LlmModel.O1: 16,
+    LlmModel.O3_MINI: 2,  # $1.10 / $4.40
+    LlmModel.O1: 16,  # $15 / $60
     LlmModel.O1_MINI: 4,
     # GPT-5 models
-    LlmModel.GPT5_2: 6,
-    LlmModel.GPT5_1: 5,
     LlmModel.GPT5: 2,
+    LlmModel.GPT5_1: 5,
     LlmModel.GPT5_MINI: 1,
     LlmModel.GPT5_NANO: 1,
     LlmModel.GPT5_CHAT: 5,
@@ -88,7 +87,7 @@ MODEL_COST: dict[LlmModel, int] = {
     LlmModel.AIML_API_LLAMA3_3_70B: 1,
     LlmModel.AIML_API_META_LLAMA_3_1_70B: 1,
     LlmModel.AIML_API_LLAMA_3_2_3B: 1,
-    LlmModel.LLAMA3_3_70B: 1,
+    LlmModel.LLAMA3_3_70B: 1,  # $0.59 / $0.79
     LlmModel.LLAMA3_1_8B: 1,
     LlmModel.OLLAMA_LLAMA3_3: 1,
     LlmModel.OLLAMA_LLAMA3_2: 1,

autogpt_platform/backend/backend/data/credit.py

@@ -16,7 +16,6 @@ from prisma.models import CreditRefundRequest, CreditTransaction, User, UserBala
 from prisma.types import CreditRefundRequestCreateInput, CreditTransactionWhereInput
 from pydantic import BaseModel
 
-from backend.api.features.admin.model import UserHistoryResponse
 from backend.data.block_cost_config import BLOCK_COSTS
 from backend.data.db import query_raw_with_schema
 from backend.data.includes import MAX_CREDIT_REFUND_REQUESTS_FETCH
@@ -30,6 +29,7 @@ from backend.data.model import (
 from backend.data.notifications import NotificationEventModel, RefundRequestData
 from backend.data.user import get_user_by_id, get_user_email_by_id
 from backend.notifications.notifications import queue_notification_async
+from backend.server.v2.admin.model import UserHistoryResponse
 from backend.util.exceptions import InsufficientBalanceError
 from backend.util.feature_flag import Flag, is_feature_enabled
 from backend.util.json import SafeJson, dumps
@@ -341,19 +341,6 @@ class UserCreditBase(ABC):
 
         if result:
             # UserBalance is already updated by the CTE
-
-            # Clear insufficient funds notification flags when credits are added
-            # so user can receive alerts again if they run out in the future.
-            if transaction.amount > 0 and transaction.type in [
-                CreditTransactionType.GRANT,
-                CreditTransactionType.TOP_UP,
-            ]:
-                from backend.executor.manager import (
-                    clear_insufficient_funds_notifications,
-                )
-
-                await clear_insufficient_funds_notifications(user_id)
-
             return result[0]["balance"]
 
     async def _add_transaction(
@@ -543,22 +530,6 @@ class UserCreditBase(ABC):
         if result:
             new_balance, tx_key = result[0]["balance"], result[0]["transactionKey"]
             # UserBalance is already updated by the CTE
-
-            # Clear insufficient funds notification flags when credits are added
-            # so user can receive alerts again if they run out in the future.
-            if (
-                amount > 0
-                and is_active
-                and transaction_type
-                in [CreditTransactionType.GRANT, CreditTransactionType.TOP_UP]
-            ):
-                # Lazy import to avoid circular dependency with executor.manager
-                from backend.executor.manager import (
-                    clear_insufficient_funds_notifications,
-                )
-
-                await clear_insufficient_funds_notifications(user_id)
-
             return new_balance, tx_key
 
         # If no result, either user doesn't exist or insufficient balance

autogpt_platform/backend/backend/data/db.py

@@ -38,20 +38,6 @@ POOL_TIMEOUT = os.getenv("DB_POOL_TIMEOUT")
 if POOL_TIMEOUT:
     DATABASE_URL = add_param(DATABASE_URL, "pool_timeout", POOL_TIMEOUT)
 
-# Add public schema to search_path for pgvector type access
-# The vector extension is in public schema, but search_path is determined by schema parameter
-# Extract the schema from DATABASE_URL or default to 'public' (matching get_database_schema())
-parsed_url = urlparse(DATABASE_URL)
-url_params = dict(parse_qsl(parsed_url.query))
-db_schema = url_params.get("schema", "public")
-# Build search_path, avoiding duplicates if db_schema is already 'public'
-search_path_schemas = list(
-    dict.fromkeys([db_schema, "public"])
-)  # Preserves order, removes duplicates
-search_path = ",".join(search_path_schemas)
-# This allows using ::vector without schema qualification
-DATABASE_URL = add_param(DATABASE_URL, "options", f"-c search_path={search_path}")
-
 HTTP_TIMEOUT = int(POOL_TIMEOUT) if POOL_TIMEOUT else None
 
 prisma = Prisma(
@@ -122,102 +108,21 @@ def get_database_schema() -> str:
     return query_params.get("schema", "public")
 
 
-async def _raw_with_schema(
-    query_template: str,
-    *args,
-    execute: bool = False,
-    client: Prisma | None = None,
-    set_public_search_path: bool = False,
-) -> list[dict] | int:
-    """Internal: Execute raw SQL with proper schema handling.
-
-    Use query_raw_with_schema() or execute_raw_with_schema() instead.
-
-    Args:
-        query_template: SQL query with {schema_prefix} placeholder
-        *args: Query parameters
-        execute: If False, executes SELECT query. If True, executes INSERT/UPDATE/DELETE.
-        client: Optional Prisma client for transactions (only used when execute=True).
-        set_public_search_path: If True, sets search_path to include public schema.
-                                Needed for pgvector types and other public schema objects.
-
-    Returns:
-        - list[dict] if execute=False (query results)
-        - int if execute=True (number of affected rows)
-    """
+async def query_raw_with_schema(query_template: str, *args) -> list[dict]:
+    """Execute raw SQL query with proper schema handling."""
     schema = get_database_schema()
-    schema_prefix = f'"{schema}".' if schema != "public" else ""
+    schema_prefix = f"{schema}." if schema != "public" else ""
     formatted_query = query_template.format(schema_prefix=schema_prefix)
 
     import prisma as prisma_module
 
-    db_client = client if client else prisma_module.get_client()
-
-    # Set search_path to include public schema if requested
-    # Prisma doesn't support the 'options' connection parameter, so we set it per-session
-    # This is idempotent and safe to call multiple times
-    if set_public_search_path:
-        await db_client.execute_raw(f"SET search_path = {schema}, public")  # type: ignore
-
-    if execute:
-        result = await db_client.execute_raw(formatted_query, *args)  # type: ignore
-    else:
-        result = await db_client.query_raw(formatted_query, *args)  # type: ignore
+    result = await prisma_module.get_client().query_raw(
+        formatted_query, *args  # type: ignore
+    )
 
     return result
 
 
-async def query_raw_with_schema(
-    query_template: str, *args, set_public_search_path: bool = False
-) -> list[dict]:
-    """Execute raw SQL SELECT query with proper schema handling.
-
-    Args:
-        query_template: SQL query with {schema_prefix} placeholder
-        *args: Query parameters
-        set_public_search_path: If True, sets search_path to include public schema.
-                                Needed for pgvector types and other public schema objects.
-
-    Returns:
-        List of result rows as dictionaries
-
-    Example:
-        results = await query_raw_with_schema(
-            'SELECT * FROM {schema_prefix}"User" WHERE id = $1',
-            user_id
-        )
-    """
-    return await _raw_with_schema(query_template, *args, execute=False, set_public_search_path=set_public_search_path)  # type: ignore
-
-
-async def execute_raw_with_schema(
-    query_template: str,
-    *args,
-    client: Prisma | None = None,
-    set_public_search_path: bool = False,
-) -> int:
-    """Execute raw SQL command (INSERT/UPDATE/DELETE) with proper schema handling.
-
-    Args:
-        query_template: SQL query with {schema_prefix} placeholder
-        *args: Query parameters
-        client: Optional Prisma client for transactions
-        set_public_search_path: If True, sets search_path to include public schema.
-                                Needed for pgvector types and other public schema objects.
-
-    Returns:
-        Number of affected rows
-
-    Example:
-        await execute_raw_with_schema(
-            'INSERT INTO {schema_prefix}"User" (id, name) VALUES ($1, $2)',
-            user_id, name,
-            client=tx  # Optional transaction client
-        )
-    """
-    return await _raw_with_schema(query_template, *args, execute=True, client=client, set_public_search_path=set_public_search_path)  # type: ignore
-
-
 class BaseDbModel(BaseModel):
     id: str = Field(default_factory=lambda: str(uuid4()))
 

autogpt_platform/backend/backend/data/execution.py

@@ -5,7 +5,6 @@ from enum import Enum
 from multiprocessing import Manager
 from queue import Empty
 from typing import (
-    TYPE_CHECKING,
     Annotated,
     Any,
     AsyncGenerator,
@@ -66,9 +65,6 @@ from .includes import (
 )
 from .model import CredentialsMetaInput, GraphExecutionStats, NodeExecutionStats
 
-if TYPE_CHECKING:
-    pass
-
 T = TypeVar("T")
 
 logger = logging.getLogger(__name__)
@@ -383,7 +379,6 @@ class GraphExecutionWithNodes(GraphExecution):
         self,
         execution_context: ExecutionContext,
         compiled_nodes_input_masks: Optional[NodesInputMasks] = None,
-        nodes_to_skip: Optional[set[str]] = None,
     ):
         return GraphExecutionEntry(
             user_id=self.user_id,
@@ -391,7 +386,6 @@ class GraphExecutionWithNodes(GraphExecution):
             graph_version=self.graph_version or 0,
             graph_exec_id=self.id,
             nodes_input_masks=compiled_nodes_input_masks,
-            nodes_to_skip=nodes_to_skip or set(),
             execution_context=execution_context,
         )
 
@@ -842,30 +836,6 @@ async def upsert_execution_output(
     await AgentNodeExecutionInputOutput.prisma().create(data=data)
 
 
-async def get_execution_outputs_by_node_exec_id(
-    node_exec_id: str,
-) -> dict[str, Any]:
-    """
-    Get all execution outputs for a specific node execution ID.
-
-    Args:
-        node_exec_id: The node execution ID to get outputs for
-
-    Returns:
-        Dictionary mapping output names to their data values
-    """
-    outputs = await AgentNodeExecutionInputOutput.prisma().find_many(
-        where={"referencedByOutputExecId": node_exec_id}
-    )
-
-    result = {}
-    for output in outputs:
-        if output.data is not None:
-            result[output.name] = type_utils.convert(output.data, JsonValue)
-
-    return result
-
-
 async def update_graph_execution_start_time(
     graph_exec_id: str,
 ) -> GraphExecution | None:
@@ -1147,8 +1117,6 @@ class GraphExecutionEntry(BaseModel):
     graph_id: str
     graph_version: int
     nodes_input_masks: Optional[NodesInputMasks] = None
-    nodes_to_skip: set[str] = Field(default_factory=set)
-    """Node IDs that should be skipped due to optional credentials not being configured."""
     execution_context: ExecutionContext = Field(default_factory=ExecutionContext)
 
 

autogpt_platform/backend/backend/data/graph.py

@@ -94,15 +94,6 @@ class Node(BaseDbModel):
     input_links: list[Link] = []
     output_links: list[Link] = []
 
-    @property
-    def credentials_optional(self) -> bool:
-        """
-        Whether credentials are optional for this node.
-        When True and credentials are not configured, the node will be skipped
-        during execution rather than causing a validation error.
-        """
-        return self.metadata.get("credentials_optional", False)
-
     @property
     def block(self) -> AnyBlockSchema | "_UnknownBlockBase":
         """Get the block for this node. Returns UnknownBlock if block is deleted/missing."""
@@ -244,10 +235,7 @@ class BaseGraph(BaseDbModel):
         return any(
             node.block_id
             for node in self.nodes
-            if (
-                node.block.block_type == BlockType.HUMAN_IN_THE_LOOP
-                or node.block.requires_human_review
-            )
+            if node.block.block_type == BlockType.HUMAN_IN_THE_LOOP
         )
 
     @property
@@ -338,35 +326,7 @@ class Graph(BaseGraph):
     @computed_field
     @property
     def credentials_input_schema(self) -> dict[str, Any]:
-        schema = self._credentials_input_schema.jsonschema()
-
-        # Determine which credential fields are required based on credentials_optional metadata
-        graph_credentials_inputs = self.aggregate_credentials_inputs()
-        required_fields = []
-
-        # Build a map of node_id -> node for quick lookup
-        all_nodes = {node.id: node for node in self.nodes}
-        for sub_graph in self.sub_graphs:
-            for node in sub_graph.nodes:
-                all_nodes[node.id] = node
-
-        for field_key, (
-            _field_info,
-            node_field_pairs,
-        ) in graph_credentials_inputs.items():
-            # A field is required if ANY node using it has credentials_optional=False
-            is_required = False
-            for node_id, _field_name in node_field_pairs:
-                node = all_nodes.get(node_id)
-                if node and not node.credentials_optional:
-                    is_required = True
-                    break
-
-            if is_required:
-                required_fields.append(field_key)
-
-        schema["required"] = required_fields
-        return schema
+        return self._credentials_input_schema.jsonschema()
 
     @property
     def _credentials_input_schema(self) -> type[BlockSchema]:

autogpt_platform/backend/backend/data/graph_test.py

@@ -1,35 +1,23 @@
 import json
 from typing import Any
-from unittest.mock import AsyncMock, patch
 from uuid import UUID
 
 import fastapi.exceptions
 import pytest
 from pytest_snapshot.plugin import Snapshot
 
-import backend.api.features.store.model as store
-from backend.api.model import CreateGraph
+import backend.server.v2.store.model as store
 from backend.blocks.basic import StoreValueBlock
 from backend.blocks.io import AgentInputBlock, AgentOutputBlock
 from backend.data.block import BlockSchema, BlockSchemaInput
 from backend.data.graph import Graph, Link, Node
 from backend.data.model import SchemaField
 from backend.data.user import DEFAULT_USER_ID
+from backend.server.model import CreateGraph
 from backend.usecases.sample import create_test_user
 from backend.util.test import SpinTestServer
 
 
-@pytest.fixture(scope="session", autouse=True)
-def mock_embedding_functions():
-    """Mock embedding functions for all tests to avoid database/API dependencies."""
-    with patch(
-        "backend.api.features.store.db.ensure_embedding",
-        new_callable=AsyncMock,
-        return_value=True,
-    ):
-        yield
-
-
 @pytest.mark.asyncio(loop_scope="session")
 async def test_graph_creation(server: SpinTestServer, snapshot: Snapshot):
     """
@@ -408,58 +396,3 @@ async def test_access_store_listing_graph(server: SpinTestServer):
         created_graph.id, created_graph.version, "3e53486c-cf57-477e-ba2a-cb02dc828e1b"
     )
     assert got_graph is not None
-
-
-# ============================================================================
-# Tests for Optional Credentials Feature
-# ============================================================================
-
-
-def test_node_credentials_optional_default():
-    """Test that credentials_optional defaults to False when not set in metadata."""
-    node = Node(
-        id="test_node",
-        block_id=StoreValueBlock().id,
-        input_default={},
-        metadata={},
-    )
-    assert node.credentials_optional is False
-
-
-def test_node_credentials_optional_true():
-    """Test that credentials_optional returns True when explicitly set."""
-    node = Node(
-        id="test_node",
-        block_id=StoreValueBlock().id,
-        input_default={},
-        metadata={"credentials_optional": True},
-    )
-    assert node.credentials_optional is True
-
-
-def test_node_credentials_optional_false():
-    """Test that credentials_optional returns False when explicitly set to False."""
-    node = Node(
-        id="test_node",
-        block_id=StoreValueBlock().id,
-        input_default={},
-        metadata={"credentials_optional": False},
-    )
-    assert node.credentials_optional is False
-
-
-def test_node_credentials_optional_with_other_metadata():
-    """Test that credentials_optional works correctly with other metadata present."""
-    node = Node(
-        id="test_node",
-        block_id=StoreValueBlock().id,
-        input_default={},
-        metadata={
-            "position": {"x": 100, "y": 200},
-            "customized_name": "My Custom Node",
-            "credentials_optional": True,
-        },
-    )
-    assert node.credentials_optional is True
-    assert node.metadata["position"] == {"x": 100, "y": 200}
-    assert node.metadata["customized_name"] == "My Custom Node"

autogpt_platform/backend/backend/data/human_review.py

@@ -13,7 +13,7 @@ from prisma.models import PendingHumanReview
 from prisma.types import PendingHumanReviewUpdateInput
 from pydantic import BaseModel
 
-from backend.api.features.executions.review.model import (
+from backend.server.v2.executions.review.model import (
     PendingHumanReviewModel,
     SafeJsonData,
 )
@@ -100,7 +100,7 @@ async def get_or_create_human_review(
         return None
     else:
         return ReviewResult(
-            data=review.payload,
+            data=review.payload if review.status == ReviewStatus.APPROVED else None,
             status=review.status,
             message=review.reviewMessage or "",
             processed=review.processed,