fix(agent/llm): Include id in tool_calls in prompt

OpenAI requires the `id` property on tool calls in assistant messages.
We weren't storing it in the `AssistantChatMessage` that is created from the LLM's response,
causing an error when adding those messages back to the prompt.

Fix:
* Add `id` to `AssistantToolCall` and `AssistantToolCallDict` types in model_providers/schema.py
* Amend `_tool_calls_compat_extract_calls` to generate an ID for extracted tool calls

---
Co-authored-by: kcze <kpczerwinski@gmail.com>

autogpts/autogpt/autogpt/core/resource/model_providers/openai.py

@@ -758,6 +758,7 @@ def _functions_compat_fix_kwargs(
 def _tool_calls_compat_extract_calls(response: str) -> Iterator[AssistantToolCall]:
     import json
     import re
+    import uuid
 
     logging.debug(f"Trying to extract tool calls from response:\n{response}")
 
@@ -770,6 +771,7 @@ def _tool_calls_compat_extract_calls(response: str) -> Iterator[AssistantToolCal
         tool_calls: list[AssistantToolCallDict] = json.loads(block.group(1))
 
     for t in tool_calls:
+        t["id"] = str(uuid.uuid4())
         t["function"]["arguments"] = str(t["function"]["arguments"])  # HACK
 
         yield AssistantToolCall.parse_obj(t)

autogpts/autogpt/autogpt/core/resource/model_providers/schema.py

@@ -77,13 +77,13 @@ class AssistantFunctionCallDict(TypedDict):
 
 
 class AssistantToolCall(BaseModel):
-    # id: str
+    id: str
     type: Literal["function"]
     function: AssistantFunctionCall
 
 
 class AssistantToolCallDict(TypedDict):
-    # id: str
+    id: str
     type: Literal["function"]
     function: AssistantFunctionCallDict