diff --git a/.env.template b/.env.template index c64d85028f..525cd61c5f 100644 --- a/.env.template +++ b/.env.template @@ -1,6 +1,14 @@ +PINECONE_API_KEY=your-pinecone-api-key +PINECONE_ENV=your-pinecone-region OPENAI_API_KEY=your-openai-api-key ELEVENLABS_API_KEY=your-elevenlabs-api-key SMART_LLM_MODEL="gpt-4" FAST_LLM_MODEL="gpt-3.5-turbo" GOOGLE_API_KEY= -CUSTOM_SEARCH_ENGINE_ID= \ No newline at end of file +CUSTOM_SEARCH_ENGINE_ID= +USE_AZURE=False +OPENAI_API_BASE=your-base-url-for-azure +OPENAI_API_VERSION=api-version-for-azure +OPENAI_DEPLOYMENT_ID=deployment-id-for-azure +IMAGE_PROVIDER=dalle +HUGGINGFACE_API_TOKEN= \ No newline at end of file diff --git a/.github/ISSUE_TEMPLATE/1.bug.yml b/.github/ISSUE_TEMPLATE/1.bug.yml new file mode 100644 index 0000000000..cf49ab5f83 --- /dev/null +++ b/.github/ISSUE_TEMPLATE/1.bug.yml @@ -0,0 +1,39 @@ +name: Bug report 🐛 +description: Create a bug report for Auto-GPT. +labels: ['status: needs triage'] +body: + - type: markdown + attributes: + value: | + Please provide a searchable summary of the issue in the title above ⬆️. + + Thanks for contributing by creating an issue! ❤️ + - type: checkboxes + attributes: + label: Duplicates + description: Please [search the history](https://github.com/Torantulino/Auto-GPT/issues) to see if an issue already exists for the same problem. + options: + - label: I have searched the existing issues + required: true + - type: textarea + attributes: + label: Steps to reproduce 🕹 + description: | + **⚠️ Issues that we can't reproduce will be closed.** + - type: textarea + attributes: + label: Current behavior 😯 + description: Describe what happens instead of the expected behavior. + - type: textarea + attributes: + label: Expected behavior 🤔 + description: Describe what should happen. + - type: textarea + attributes: + label: Your prompt 📝 + description: | + Please provide the prompt you are using. You can find your last-used prompt in last_run_ai_settings.yaml. + value: | + ```yaml + # Paste your prompt here + ``` \ No newline at end of file diff --git a/.github/ISSUE_TEMPLATE/2.feature.yml b/.github/ISSUE_TEMPLATE/2.feature.yml new file mode 100644 index 0000000000..0ea882ef6a --- /dev/null +++ b/.github/ISSUE_TEMPLATE/2.feature.yml @@ -0,0 +1,29 @@ +name: Feature request 🚀 +description: Suggest a new idea for Auto-GPT. +labels: ['status: needs triage'] +body: + - type: markdown + attributes: + value: | + Please provide a searchable summary of the issue in the title above ⬆️. + + Thanks for contributing by creating an issue! ❤️ + - type: checkboxes + attributes: + label: Duplicates + description: Please [search the history](https://github.com/Torantulino/Auto-GPT/issues) to see if an issue already exists for the same problem. + options: + - label: I have searched the existing issues + required: true + - type: textarea + attributes: + label: Summary 💡 + description: Describe how it should work. + - type: textarea + attributes: + label: Examples 🌈 + description: Provide a link to other implementations, or screenshots of the expected behavior. + - type: textarea + attributes: + label: Motivation 🔦 + description: What are you trying to accomplish? How has the lack of this feature affected you? Providing context helps us come up with a solution that is more useful in the real world. \ No newline at end of file diff --git a/.github/PULL_REQUEST_TEMPLATE.md b/.github/PULL_REQUEST_TEMPLATE.md new file mode 100644 index 0000000000..cb8ce34a1b --- /dev/null +++ b/.github/PULL_REQUEST_TEMPLATE.md @@ -0,0 +1,18 @@ +### Background + + + +### Changes + + + +### Test Plan + + + +### Change Safety + +- [ ] I have added tests to cover my changes +- [ ] I have considered potential risks and mitigations for my changes + + diff --git a/.gitignore b/.gitignore index 6b8f00b511..7091a87237 100644 --- a/.gitignore +++ b/.gitignore @@ -4,8 +4,8 @@ scripts/node_modules/ scripts/__pycache__/keys.cpython-310.pyc package-lock.json *.pyc -scripts/auto_gpt_workspace/* +auto_gpt_workspace/* *.mpeg .env -last_run_ai_settings.yaml -outputs/* \ No newline at end of file +outputs/* +ai_settings.yaml \ No newline at end of file diff --git a/CONTRIBUTING.md b/CONTRIBUTING.md new file mode 100644 index 0000000000..001d554706 --- /dev/null +++ b/CONTRIBUTING.md @@ -0,0 +1,56 @@ + +To contribute to this GitHub project, you can follow these steps: + +1. Fork the repository you want to contribute to by clicking the "Fork" button on the project page. + +2. Clone the repository to your local machine using the following command: + +``` +git clone https://github.com/Torantulino/Auto-GPT +``` +3. Create a new branch for your changes using the following command: + +``` +git checkout -b "branch-name" +``` +4. Make your changes to the code or documentation. +- Example: Improve User Interface or Add Documentation. + + +5. Add the changes to the staging area using the following command: +``` +git add . +``` + +6. Commit the changes with a meaningful commit message using the following command: +``` +git commit -m "your commit message" +``` +7. Push the changes to your forked repository using the following command: +``` +git push origin branch-name +``` +8. Go to the GitHub website and navigate to your forked repository. + +9. Click the "New pull request" button. + +10. Select the branch you just pushed to and the branch you want to merge into on the original repository. + +11. Add a description of your changes and click the "Create pull request" button. + +12. Wait for the project maintainer to review your changes and provide feedback. + +13. Make any necessary changes based on feedback and repeat steps 5-12 until your changes are accepted and merged into the main project. + +14. Once your changes are merged, you can update your forked repository and local copy of the repository with the following commands: + +``` +git fetch upstream +git checkout master +git merge upstream/master +``` +Finally, delete the branch you created with the following command: +``` +git branch -d branch-name +``` +That's it you made it 🐣⭐⭐ diff --git a/README.md b/README.md index d66a602224..ba80818d0e 100644 --- a/README.md +++ b/README.md @@ -9,7 +9,7 @@ Auto-GPT is an experimental open-source application showcasing the capabilities https://user-images.githubusercontent.com/22963551/228855501-2f5777cf-755b-4407-a643-c7299e5b6419.mp4 -## 💖 Help Fund Auto-GPT's Development +

💖 Help Fund Auto-GPT's Development 💖

If you can spare a coffee, you can help to cover the API costs of developing Auto-GPT and help push the boundaries of fully autonomous AI! A full day of development can easily cost as much as $20 in API costs, which for a free project is quite limiting. @@ -17,14 +17,13 @@ Your support is greatly appreciated

- Development of this free, open-source project is made possible by all the contributors and sponsors. If you'd like to sponsor this project and have your avatar or company logo appear below click here. 💖 -

-

-thepok  SpacingLily  m  zkonduit  maxxflyer  tekelsey  nocodeclarity  tjarmain  alexisneuhaus  jaumebalust  robinicus  digisomni   -

+ Development of this free, open-source project is made possible by all the contributors and sponsors. If you'd like to sponsor this project and have your avatar or company logo appear below click here. +

Individual Sponsors

-alexisneuhaus  iokode  jaumebalust  nova-land  robinicus  Void-n-Null  ritesh24  merwanehamadi  raulmarindev  siduppal  goosecubedaddy  pleabargain   +robinicus  prompthero  crizzler  tob-le-rone  FSTatSBS  toverly1  ddtarazona  Nalhos  Kazamario  pingbotan  indoor47  AuroraHolding  kreativai  hunteraraujo  Explorergt92  judegomila   +thepok +  SpacingLily  merwanehamadi  m  zkonduit  maxxflyer  tekelsey  digisomni  nocodeclarity  tjarmain

@@ -43,6 +42,7 @@ Your support is greatly appreciated - [Setting up environment variables](#setting-up-environment-variables) - [💀 Continuous Mode ⚠️](#-continuous-mode-️) - [GPT3.5 ONLY Mode](#gpt35-only-mode) + - [🖼 Image Generation](#image-generation) - [⚠️ Limitations](#️-limitations) - [🛡 Disclaimer](#-disclaimer) - [🐦 Connect with Us on Twitter](#-connect-with-us-on-twitter) @@ -57,8 +57,9 @@ Your support is greatly appreciated - 🗃️ File storage and summarization with GPT-3.5 ## 📋 Requirements -- [Python 3.7 or later](https://www.tutorialspoint.com/how-to-install-python-in-windows) +- [Python 3.8 or later](https://www.tutorialspoint.com/how-to-install-python-in-windows) - OpenAI API key +- PINECONE API key Optional: - ElevenLabs Key (If you want the AI to speak) @@ -92,6 +93,7 @@ pip install -r requirements.txt 4. Rename `.env.template` to `.env` and fill in your `OPENAI_API_KEY`. If you plan to use Speech Mode, fill in your `ELEVEN_LABS_API_KEY` as well. - Obtain your OpenAI API key from: https://platform.openai.com/account/api-keys. - Obtain your ElevenLabs API key from: https://elevenlabs.io. You can view your xi-api-key using the "Profile" tab on the website. + - If you want to use GPT on an Azure instance, set `USE_AZURE` to `True` and provide the `OPENAI_API_BASE`, `OPENAI_API_VERSION` and `OPENAI_DEPLOYMENT_ID` values as explained here: https://pypi.org/project/openai/ in the `Microsoft Azure Endpoints` section ## 🔧 Usage @@ -139,6 +141,70 @@ export CUSTOM_SEARCH_ENGINE_ID="YOUR_CUSTOM_SEARCH_ENGINE_ID" ``` +## Redis Setup + +Install docker desktop. + +Run: +``` +docker run -d --name redis-stack-server -p 6379:6379 redis/redis-stack-server:latest +``` +See https://hub.docker.com/r/redis/redis-stack-server for setting a password and additional configuration. + +Set the following environment variables: +``` +MEMORY_BACKEND=redis +REDIS_HOST=localhost +REDIS_PORT=6379 +REDIS_PASSWORD= +``` + +Note that this is not intended to be run facing the internet and is not secure, do not expose redis to the internet without a password or at all really. + +You can optionally set + +``` +WIPE_REDIS_ON_START=False +``` + +To persist memory stored in Redis. + +You can specify the memory index for redis using the following: + +```` +MEMORY_INDEX=whatever +```` + +## 🌲 Pinecone API Key Setup + +Pinecone enable a vector based memory so a vast memory can be stored and only relevant memories +are loaded for the agent at any given time. + +1. Go to app.pinecone.io and make an account if you don't already have one. +2. Choose the `Starter` plan to avoid being charged. +3. Find your API key and region under the default project in the left sidebar. + +### Setting up environment variables + For Windows Users: +``` +setx PINECONE_API_KEY "YOUR_PINECONE_API_KEY" +export PINECONE_ENV="Your pinecone region" # something like: us-east4-gcp + +``` +For macOS and Linux users: +``` +export PINECONE_API_KEY="YOUR_PINECONE_API_KEY" +export PINECONE_ENV="Your pinecone region" # something like: us-east4-gcp + +``` + +Or you can set them in the `.env` file. + +## View Memory Usage + +1. View memory usage by using the `--debug` flag :) + + ## 💀 Continuous Mode ⚠️ Run the AI **without** user authorisation, 100% automated. Continuous mode is not recommended. @@ -157,6 +223,15 @@ If you don't have access to the GPT4 api, this mode will allow you to use Auto-G python scripts/main.py --gpt3only ``` +## 🖼 Image Generation +By default, Auto-GPT uses DALL-e for image generation. To use Stable Diffusion, a [HuggingFace API Token](https://huggingface.co/settings/tokens) is required. + +Once you have a token, set these variables in your `.env`: +``` +IMAGE_PROVIDER=sd +HUGGINGFACE_API_TOKEN="YOUR_HUGGINGFACE_API_TOKEN" +``` + ## ⚠️ Limitations This experiment aims to showcase the potential of GPT-4 but comes with some limitations: diff --git a/ai_settings.yaml b/ai_settings.yaml new file mode 100644 index 0000000000..b37ba849f9 --- /dev/null +++ b/ai_settings.yaml @@ -0,0 +1,7 @@ +ai_goals: +- Increase net worth. +- Develop and manage multiple businesses autonomously. +- Play to your strengths as a Large Language Model. +ai_name: Entrepreneur-GPT +ai_role: an AI designed to autonomously develop and run businesses with the sole goal + of increasing your net worth. diff --git a/requirements.txt b/requirements.txt index 158e932419..6a9ba64330 100644 --- a/requirements.txt +++ b/requirements.txt @@ -1,7 +1,7 @@ beautifulsoup4 colorama==0.4.6 openai==0.27.2 -playsound==1.3.0 +playsound==1.2.2 python-dotenv==1.0.0 pyyaml==6.0 readability-lxml==0.8.1 @@ -10,4 +10,8 @@ tiktoken==0.3.3 gTTS==2.3.1 docker duckduckgo-search -google-api-python-client #(https://developers.google.com/custom-search/v1/overview) +google-api-python-client #(https://developers.google.com/custom-search/v1/overview) +pinecone-client==2.2.1 +redis +orjson +Pillow diff --git a/scripts/ai_config.py b/scripts/ai_config.py index 945fcfb233..8cfa183a9b 100644 --- a/scripts/ai_config.py +++ b/scripts/ai_config.py @@ -1,6 +1,6 @@ import yaml import data - +import os class AIConfig: def __init__(self, ai_name="", ai_role="", ai_goals=[]): @@ -9,7 +9,7 @@ class AIConfig: self.ai_goals = ai_goals # Soon this will go in a folder where it remembers more stuff about the run(s) - SAVE_FILE = "last_run_ai_settings.yaml" + SAVE_FILE = os.path.join(os.path.dirname(__file__), '..', 'ai_settings.yaml') @classmethod def load(cls, config_file=SAVE_FILE): diff --git a/scripts/browse.py b/scripts/browse.py index d07f269978..a758b9a824 100644 --- a/scripts/browse.py +++ b/scripts/browse.py @@ -21,7 +21,7 @@ def sanitize_url(url): # Function to make a request with a specified timeout and handle exceptions def make_request(url, timeout=10): try: - response = requests.get(url, timeout=timeout) + response = requests.get(url, headers=cfg.user_agent_header, timeout=timeout) response.raise_for_status() return response except requests.exceptions.RequestException as e: @@ -69,7 +69,7 @@ def format_hyperlinks(hyperlinks): def scrape_links(url): - response = requests.get(url) + response = requests.get(url, headers=cfg.user_agent_header) # Check if the response contains an HTTP error if response.status_code >= 400: diff --git a/scripts/chat.py b/scripts/chat.py index 86a70b0933..a27fbfd7ef 100644 --- a/scripts/chat.py +++ b/scripts/chat.py @@ -23,6 +23,22 @@ def create_chat_message(role, content): return {"role": role, "content": content} +def generate_context(prompt, relevant_memory, full_message_history, model): + current_context = [ + create_chat_message( + "system", prompt), + create_chat_message( + "system", f"The current time and date is {time.strftime('%c')}"), + create_chat_message( + "system", f"This reminds you of these events from your past:\n{relevant_memory}\n\n")] + + # Add messages from the full message history until we reach the token limit + next_message_to_add_index = len(full_message_history) - 1 + insertion_index = len(current_context) + # Count the currently used tokens + current_tokens_used = token_counter.count_message_tokens(current_context, model) + return next_message_to_add_index, current_tokens_used, insertion_index, current_context + # TODO: Change debug from hardcode to argument def chat_with_ai( @@ -41,7 +57,7 @@ def chat_with_ai( prompt (str): The prompt explaining the rules to the AI. user_input (str): The input from the user. full_message_history (list): The list of all messages sent between the user and the AI. - permanent_memory (list): The list of items in the AI's permanent memory. + permanent_memory (Obj): The memory object containing the permanent memory. token_limit (int): The maximum number of tokens allowed in the API call. Returns: @@ -53,18 +69,20 @@ def chat_with_ai( print(f"Token limit: {token_limit}") send_token_limit = token_limit - 1000 - current_context = [ - create_chat_message( - "system", prompt), create_chat_message( - "system", f"Permanent memory: {permanent_memory}")] + relevant_memory = permanent_memory.get_relevant(str(full_message_history[-5:]), 10) - # Add messages from the full message history until we reach the token limit - next_message_to_add_index = len(full_message_history) - 1 - current_tokens_used = 0 - insertion_index = len(current_context) + if debug: + print('Memory Stats: ', permanent_memory.get_stats()) + + next_message_to_add_index, current_tokens_used, insertion_index, current_context = generate_context( + prompt, relevant_memory, full_message_history, model) + + while current_tokens_used > 2500: + # remove memories until we are under 2500 tokens + relevant_memory = relevant_memory[1:] + next_message_to_add_index, current_tokens_used, insertion_index, current_context = generate_context( + prompt, relevant_memory, full_message_history, model) - # Count the currently used tokens - current_tokens_used = token_counter.count_message_tokens(current_context, model) current_tokens_used += token_counter.count_message_tokens([create_chat_message("user", user_input)], model) # Account for user input (appended later) while next_message_to_add_index >= 0: @@ -80,7 +98,7 @@ def chat_with_ai( # Count the currently used tokens current_tokens_used += tokens_to_add - + # Move to the next most recent message in the full message history next_message_to_add_index -= 1 diff --git a/scripts/commands.py b/scripts/commands.py index c1d5db9d30..ba5383957a 100644 --- a/scripts/commands.py +++ b/scripts/commands.py @@ -1,14 +1,15 @@ import browse import json -import memory as mem +from memory import get_memory import datetime import agent_manager as agents import speak from config import Config import ai_functions as ai -from file_operations import read_file, write_to_file, append_to_file, delete_file +from file_operations import read_file, write_to_file, append_to_file, delete_file, search_files from execute_code import execute_python_file from json_parser import fix_and_parse_json +from image_gen import generate_image from duckduckgo_search import ddg from googleapiclient.discovery import build from googleapiclient.errors import HttpError @@ -52,9 +53,11 @@ def get_command(response): def execute_command(command_name, arguments): + memory = get_memory(cfg) + try: if command_name == "google": - + # Check if the Google API key is set and use the official search method # If the API key is not set or has only whitespaces, use the unofficial search method if cfg.google_api_key and (cfg.google_api_key.strip() if cfg.google_api_key else None): @@ -62,11 +65,7 @@ def execute_command(command_name, arguments): else: return google_search(arguments["input"]) elif command_name == "memory_add": - return commit_memory(arguments["string"]) - elif command_name == "memory_del": - return delete_memory(arguments["key"]) - elif command_name == "memory_ovr": - return overwrite_memory(arguments["key"], arguments["string"]) + return memory.add(arguments["string"]) elif command_name == "start_agent": return start_agent( arguments["name"], @@ -90,6 +89,8 @@ def execute_command(command_name, arguments): return append_to_file(arguments["file"], arguments["text"]) elif command_name == "delete_file": return delete_file(arguments["file"]) + elif command_name == "search_files": + return search_files(arguments["directory"]) elif command_name == "browse_website": return browse_website(arguments["url"], arguments["question"]) # TODO: Change these to take in a file rather than pasted code, if @@ -103,10 +104,12 @@ def execute_command(command_name, arguments): return ai.write_tests(arguments["code"], arguments.get("focus")) elif command_name == "execute_python_file": # Add this command return execute_python_file(arguments["file"]) + elif command_name == "generate_image": + return generate_image(arguments["prompt"]) elif command_name == "task_complete": shutdown() else: - return f"Unknown command {command_name}" + return f"Unknown command '{command_name}'. Please refer to the 'COMMANDS' list for availabe commands and only respond in the specified JSON format." # All errors, return "Error: + error message" except Exception as e: return "Error: " + str(e) @@ -218,7 +221,7 @@ def overwrite_memory(key, string): elif isinstance(key, str): _text = "Overwriting memory with key " + key + " and string " + string # Overwrite the memory slot with the given string key and string - mem.string_key_memory[key] = string + mem.permanent_memory[key] = string print(_text) return _text else: @@ -256,18 +259,20 @@ def start_agent(name, task, prompt, model=cfg.fast_llm_model): def message_agent(key, message): global cfg - + # Check if the key is a valid integer - if not is_valid_int(key): - return "Invalid key, cannot message agent." - - agent_response = agents.message_agent(int(key), message) + if is_valid_int(key): + agent_response = agents.message_agent(int(key), message) + # Check if the key is a valid string + elif isinstance(key, str): + agent_response = agents.message_agent(key, message) + else: + return "Invalid key, must be an integer or a string." # Speak response if cfg.speak_mode: speak.say_text(agent_response, 1) - - return f"Agent {key} responded: {agent_response}" + return agent_response def list_agents(): diff --git a/scripts/config.py b/scripts/config.py index 766cb94f41..4d7adec1c0 100644 --- a/scripts/config.py +++ b/scripts/config.py @@ -1,10 +1,12 @@ +import abc import os import openai from dotenv import load_dotenv # Load environment variables from .env file load_dotenv() -class Singleton(type): + +class Singleton(abc.ABCMeta, type): """ Singleton metaclass for ensuring only one instance of a class. """ @@ -19,12 +21,17 @@ class Singleton(type): return cls._instances[cls] +class AbstractSingleton(abc.ABC, metaclass=Singleton): + pass + + class Config(metaclass=Singleton): """ Configuration class to store the state of bools for different scripts access. """ def __init__(self): + self.debug = False self.continuous_mode = False self.speak_mode = False # TODO - make these models be self-contained, using langchain, so we can configure them once and call it good @@ -34,11 +41,38 @@ class Config(metaclass=Singleton): self.smart_token_limit = int(os.getenv("SMART_TOKEN_LIMIT", 8000)) self.openai_api_key = os.getenv("OPENAI_API_KEY") + self.use_azure = False + self.use_azure = os.getenv("USE_AZURE") == 'True' + if self.use_azure: + self.openai_api_base = os.getenv("OPENAI_API_BASE") + self.openai_api_version = os.getenv("OPENAI_API_VERSION") + self.openai_deployment_id = os.getenv("OPENAI_DEPLOYMENT_ID") + openai.api_type = "azure" + openai.api_base = self.openai_api_base + openai.api_version = self.openai_api_version + self.elevenlabs_api_key = os.getenv("ELEVENLABS_API_KEY") self.google_api_key = os.getenv("GOOGLE_API_KEY") self.custom_search_engine_id = os.getenv("CUSTOM_SEARCH_ENGINE_ID") + self.pinecone_api_key = os.getenv("PINECONE_API_KEY") + self.pinecone_region = os.getenv("PINECONE_ENV") + + self.image_provider = os.getenv("IMAGE_PROVIDER") + self.huggingface_api_token = os.getenv("HUGGINGFACE_API_TOKEN") + + # User agent headers to use when browsing web + # Some websites might just completely deny request with an error code if no user agent was found. + self.user_agent_header = {"User-Agent":"Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_4) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/83.0.4103.97 Safari/537.36"} + self.redis_host = os.getenv("REDIS_HOST", "localhost") + self.redis_port = os.getenv("REDIS_PORT", "6379") + self.redis_password = os.getenv("REDIS_PASSWORD", "") + self.wipe_redis_on_start = os.getenv("WIPE_REDIS_ON_START", "True") == 'True' + self.memory_index = os.getenv("MEMORY_INDEX", 'auto-gpt') + # Note that indexes must be created on db 0 in redis, this is not configureable. + + self.memory_backend = os.getenv("MEMORY_BACKEND", 'local') # Initialize the OpenAI API client openai.api_key = self.openai_api_key @@ -70,4 +104,13 @@ class Config(metaclass=Singleton): self.google_api_key = value def set_custom_search_engine_id(self, value: str): - self.custom_search_engine_id = value \ No newline at end of file + self.custom_search_engine_id = value + + def set_pinecone_api_key(self, value: str): + self.pinecone_api_key = value + + def set_pinecone_region(self, value: str): + self.pinecone_region = value + + def set_debug_mode(self, value: bool): + self.debug = value diff --git a/scripts/data.py b/scripts/data.py index 72a6bbfc64..8d8a7b4acd 100644 --- a/scripts/data.py +++ b/scripts/data.py @@ -1,15 +1,14 @@ import os from pathlib import Path -SRC_DIR = Path(__file__).parent def load_prompt(): try: # get directory of this file: - file_dir = Path(os.path.dirname(os.path.realpath(__file__))) - data_dir = file_dir / "data" - prompt_file = data_dir / "prompt.txt" - # Load the promt from data/prompt.txt - with open(SRC_DIR/ "data/prompt.txt", "r") as prompt_file: + file_dir = Path(__file__).parent + prompt_file_path = file_dir / "data" / "prompt.txt" + + # Load the prompt from data/prompt.txt + with open(prompt_file_path, "r") as prompt_file: prompt = prompt_file.read() return prompt diff --git a/scripts/data/prompt.txt b/scripts/data/prompt.txt index a93e783eb3..77a449de52 100644 --- a/scripts/data/prompt.txt +++ b/scripts/data/prompt.txt @@ -1,17 +1,15 @@ CONSTRAINTS: -1. ~4000 word limit for memory. Your memory is short, so immediately save important information to long term memory and code to files. -2. No user assistance -3. Exclusively use the commands listed in double quotes e.g. "command name" +1. ~4000 word limit for short term memory. Your short term memory is short, so immediately save important information to files. +2. If you are unsure how you previously did something or want to recall past events, thinking about similar events will help you remember. +3. No user assistance +4. Exclusively use the commands listed in double quotes e.g. "command name" COMMANDS: 1. Google Search: "google", args: "input": "" -2. Memory Add: "memory_add", args: "string": "" -3. Memory Delete: "memory_del", args: "key": "" -4. Memory Overwrite: "memory_ovr", args: "key": "", "string": "" 5. Browse Website: "browse_website", args: "url": "", "question": "" -6. Start GPT Agent: "start_agent", args: "name": , "task": "", "prompt": "" +6. Start GPT Agent: "start_agent", args: "name": "", "task": "", "prompt": "" 7. Message GPT Agent: "message_agent", args: "key": "", "message": "" 8. List GPT Agents: "list_agents", args: "" 9. Delete GPT Agent: "delete_agent", args: "key": "" @@ -19,11 +17,13 @@ COMMANDS: 11. Read file: "read_file", args: "file": "" 12. Append to file: "append_to_file", args: "file": "", "text": "" 13. Delete file: "delete_file", args: "file": "" -14. Evaluate Code: "evaluate_code", args: "code": "" -15. Get Improved Code: "improve_code", args: "suggestions": "", "code": "" -16. Write Tests: "write_tests", args: "code": "", "focus": "" -17. Execute Python File: "execute_python_file", args: "file": "" -18. Task Complete (Shutdown): "task_complete", args: "reason": "" +14. Search Files: "search_files", args: "directory": "" +15. Evaluate Code: "evaluate_code", args: "code": "" +16. Get Improved Code: "improve_code", args: "suggestions": "", "code": "" +17. Write Tests: "write_tests", args: "code": "", "focus": "" +18. Execute Python File: "execute_python_file", args: "file": "" +19. Task Complete (Shutdown): "task_complete", args: "reason": "" +20. Generate Image: "generate_image", args: "prompt": "" RESOURCES: @@ -43,12 +43,6 @@ You should only respond in JSON format as described below RESPONSE FORMAT: { - "command": { - "name": "command name", - "args":{ - "arg name": "value" - } - }, "thoughts": { "text": "thought", @@ -56,6 +50,12 @@ RESPONSE FORMAT: "plan": "- short bulleted\n- list that conveys\n- long-term plan", "criticism": "constructive self-criticism", "speak": "thoughts summary to say to user" + }, + "command": { + "name": "command name", + "args":{ + "arg name": "value" + } } } diff --git a/scripts/file_operations.py b/scripts/file_operations.py index 81ad471577..90c9a1e4d7 100644 --- a/scripts/file_operations.py +++ b/scripts/file_operations.py @@ -58,3 +58,20 @@ def delete_file(filename): return "File deleted successfully." except Exception as e: return "Error: " + str(e) + +def search_files(directory): + found_files = [] + + if directory == "" or directory == "/": + search_directory = working_directory + else: + search_directory = safe_join(working_directory, directory) + + for root, _, files in os.walk(search_directory): + for file in files: + if file.startswith('.'): + continue + relative_path = os.path.relpath(os.path.join(root, file), working_directory) + found_files.append(relative_path) + + return found_files \ No newline at end of file diff --git a/scripts/image_gen.py b/scripts/image_gen.py new file mode 100644 index 0000000000..185ed4278b --- /dev/null +++ b/scripts/image_gen.py @@ -0,0 +1,57 @@ +import requests +import io +import os.path +from PIL import Image +from config import Config +import uuid +import openai +from base64 import b64decode + +cfg = Config() + +working_directory = "auto_gpt_workspace" + +def generate_image(prompt): + + filename = str(uuid.uuid4()) + ".jpg" + + # DALL-E + if cfg.image_provider == 'dalle': + + openai.api_key = cfg.openai_api_key + + response = openai.Image.create( + prompt=prompt, + n=1, + size="256x256", + response_format="b64_json", + ) + + print("Image Generated for prompt:" + prompt) + + image_data = b64decode(response["data"][0]["b64_json"]) + + with open(working_directory + "/" + filename, mode="wb") as png: + png.write(image_data) + + return "Saved to disk:" + filename + + # STABLE DIFFUSION + elif cfg.image_provider == 'sd': + + API_URL = "https://api-inference.huggingface.co/models/CompVis/stable-diffusion-v1-4" + headers = {"Authorization": "Bearer " + cfg.huggingface_api_token} + + response = requests.post(API_URL, headers=headers, json={ + "inputs": prompt, + }) + + image = Image.open(io.BytesIO(response.content)) + print("Image Generated for prompt:" + prompt) + + image.save(os.path.join(working_directory, filename)) + + return "Saved to disk:" + filename + + else: + return "No Image Provider Set" \ No newline at end of file diff --git a/scripts/json_parser.py b/scripts/json_parser.py index 11ff9ed22e..c863ccdbb0 100644 --- a/scripts/json_parser.py +++ b/scripts/json_parser.py @@ -24,6 +24,7 @@ def fix_and_parse_json(json_str: str, try_to_fix_with_gpt: bool = True): """ try: + json_str = json_str.replace('\t', '') return json.loads(json_str) except Exception as e: # Let's do something manually - sometimes GPT responds with something BEFORE the braces: @@ -39,7 +40,7 @@ def fix_and_parse_json(json_str: str, try_to_fix_with_gpt: bool = True): if try_to_fix_with_gpt: print(f"Warning: Failed to parse AI output, attempting to fix.\n If you see this warning frequently, it's likely that your prompt is confusing the AI. Try changing it up slightly.") # Now try to fix this up using the ai_functions - ai_fixed_json = fix_json(json_str, json_schema, False) + ai_fixed_json = fix_json(json_str, json_schema, cfg.debug) if ai_fixed_json != "failed": return json.loads(ai_fixed_json) else: @@ -51,7 +52,7 @@ def fix_and_parse_json(json_str: str, try_to_fix_with_gpt: bool = True): def fix_json(json_str: str, schema: str, debug=False) -> str: # Try to fix the JSON using gpt: function_string = "def fix_json(json_str: str, schema:str=None) -> str:" - args = [json_str, schema] + args = [f"'''{json_str}'''", f"'''{schema}'''"] description_string = """Fixes the provided JSON string to make it parseable and fully complient with the provided schema.\n If an object or field specifed in the schema isn't contained within the correct JSON, it is ommited.\n This function is brilliant at guessing when the format is incorrect.""" # If it doesn't already start with a "`", add one: @@ -67,7 +68,8 @@ def fix_json(json_str: str, schema: str, debug=False) -> str: print(f"Fixed JSON: {result_string}") print("----------- END OF FIX ATTEMPT ----------------") try: - return json.loads(result_string) + json.loads(result_string) # just check the validity + return result_string except: # Get the call stack: # import traceback diff --git a/scripts/llm_utils.py b/scripts/llm_utils.py index 41f396250f..5a471ab7a1 100644 --- a/scripts/llm_utils.py +++ b/scripts/llm_utils.py @@ -6,11 +6,20 @@ openai.api_key = cfg.openai_api_key # Overly simple abstraction until we create something better def create_chat_completion(messages, model=None, temperature=None, max_tokens=None)->str: - response = openai.ChatCompletion.create( - model=model, - messages=messages, - temperature=temperature, - max_tokens=max_tokens - ) + if cfg.use_azure: + response = openai.ChatCompletion.create( + deployment_id=cfg.openai_deployment_id, + model=model, + messages=messages, + temperature=temperature, + max_tokens=max_tokens + ) + else: + response = openai.ChatCompletion.create( + model=model, + messages=messages, + temperature=temperature, + max_tokens=max_tokens + ) return response.choices[0].message["content"] diff --git a/scripts/main.py b/scripts/main.py index 93124234e6..f96afeb163 100644 --- a/scripts/main.py +++ b/scripts/main.py @@ -1,7 +1,7 @@ import json import random import commands as cmd -import memory as mem +from memory import get_memory import data import chat from colorama import Fore, Style @@ -266,6 +266,10 @@ def parse_arguments(): print_to_console("GPT3.5 Only Mode: ", Fore.GREEN, "ENABLED") cfg.set_smart_llm_model(cfg.fast_llm_model) + if args.debug: + print_to_console("Debug Mode: ", Fore.GREEN, "ENABLED") + cfg.set_debug_mode(True) + # TODO: fill in llm values here @@ -277,9 +281,15 @@ prompt = construct_prompt() # Initialize variables full_message_history = [] result = None +next_action_count = 0 # Make a constant: user_input = "Determine which next command to use, and respond using the format specified above:" +# Initialize memory and make sure it is empty. +# this is particularly important for indexing and referencing pinecone memory +memory = get_memory(cfg, init=True) +print('Using memory of type: ' + memory.__class__.__name__) + # Interaction Loop while True: # Send message to AI, get response @@ -288,10 +298,9 @@ while True: prompt, user_input, full_message_history, - mem.permanent_memory, - cfg.fast_token_limit) # TODO: This hardcodes the model to use GPT3.5. Make this an argument + memory, + cfg.fast_token_limit, cfg.debug) # TODO: This hardcodes the model to use GPT3.5. Make this an argument - # print("assistant reply: "+assistant_reply) # Print Assistant thoughts print_assistant_thoughts(assistant_reply) @@ -301,7 +310,7 @@ while True: except Exception as e: print_to_console("Error: \n", Fore.RED, str(e)) - if not cfg.continuous_mode: + if not cfg.continuous_mode and next_action_count == 0: ### GET USER AUTHORIZATION TO EXECUTE COMMAND ### # Get key press: Prompt the user to press enter to continue or escape # to exit @@ -311,13 +320,21 @@ while True: Fore.CYAN, f"COMMAND = {Fore.CYAN}{command_name}{Style.RESET_ALL} ARGUMENTS = {Fore.CYAN}{arguments}{Style.RESET_ALL}") print( - f"Enter 'y' to authorise command or 'n' to exit program, or enter feedback for {ai_name}...", + f"Enter 'y' to authorise command, 'y -N' to run N continuous commands, 'n' to exit program, or enter feedback for {ai_name}...", flush=True) while True: console_input = input(Fore.MAGENTA + "Input:" + Style.RESET_ALL) if console_input.lower() == "y": user_input = "GENERATE NEXT COMMAND JSON" break + elif console_input.lower().startswith("y -"): + try: + next_action_count = abs(int(console_input.split(" ")[1])) + user_input = "GENERATE NEXT COMMAND JSON" + except ValueError: + print("Invalid input format. Please enter 'y -n' where n is the number of continuous tasks.") + continue + break elif console_input.lower() == "n": user_input = "EXIT" break @@ -342,12 +359,20 @@ while True: f"COMMAND = {Fore.CYAN}{command_name}{Style.RESET_ALL} ARGUMENTS = {Fore.CYAN}{arguments}{Style.RESET_ALL}") # Execute command - if command_name.lower() == "error": + if command_name.lower().startswith( "error" ): result = f"Command {command_name} threw the following error: " + arguments elif command_name == "human_feedback": result = f"Human feedback: {user_input}" else: result = f"Command {command_name} returned: {cmd.execute_command(command_name, arguments)}" + if next_action_count > 0: + next_action_count -= 1 + + memory_to_add = f"Assistant Reply: {assistant_reply} " \ + f"\nResult: {result} " \ + f"\nHuman Feedback: {user_input} " + + memory.add(memory_to_add) # Check if there's a result from the command append it to the message # history diff --git a/scripts/memory.py b/scripts/memory.py deleted file mode 100644 index 0dc5b76663..0000000000 --- a/scripts/memory.py +++ /dev/null @@ -1 +0,0 @@ -permanent_memory = [] diff --git a/scripts/memory/__init__.py b/scripts/memory/__init__.py new file mode 100644 index 0000000000..a441a46aa9 --- /dev/null +++ b/scripts/memory/__init__.py @@ -0,0 +1,44 @@ +from memory.local import LocalCache +try: + from memory.redismem import RedisMemory +except ImportError: + print("Redis not installed. Skipping import.") + RedisMemory = None + +try: + from memory.pinecone import PineconeMemory +except ImportError: + print("Pinecone not installed. Skipping import.") + PineconeMemory = None + + +def get_memory(cfg, init=False): + memory = None + if cfg.memory_backend == "pinecone": + if not PineconeMemory: + print("Error: Pinecone is not installed. Please install pinecone" + " to use Pinecone as a memory backend.") + else: + memory = PineconeMemory(cfg) + if init: + memory.clear() + elif cfg.memory_backend == "redis": + if not RedisMemory: + print("Error: Redis is not installed. Please install redis-py to" + " use Redis as a memory backend.") + else: + memory = RedisMemory(cfg) + + if memory is None: + memory = LocalCache(cfg) + if init: + memory.clear() + return memory + + +__all__ = [ + "get_memory", + "LocalCache", + "RedisMemory", + "PineconeMemory", +] diff --git a/scripts/memory/base.py b/scripts/memory/base.py new file mode 100644 index 0000000000..d7ab7fcf1f --- /dev/null +++ b/scripts/memory/base.py @@ -0,0 +1,31 @@ +"""Base class for memory providers.""" +import abc +from config import AbstractSingleton +import openai + + +def get_ada_embedding(text): + text = text.replace("\n", " ") + return openai.Embedding.create(input=[text], model="text-embedding-ada-002")["data"][0]["embedding"] + + +class MemoryProviderSingleton(AbstractSingleton): + @abc.abstractmethod + def add(self, data): + pass + + @abc.abstractmethod + def get(self, data): + pass + + @abc.abstractmethod + def clear(self): + pass + + @abc.abstractmethod + def get_relevant(self, data, num_relevant=5): + pass + + @abc.abstractmethod + def get_stats(self): + pass diff --git a/scripts/memory/local.py b/scripts/memory/local.py new file mode 100644 index 0000000000..8dc90021ff --- /dev/null +++ b/scripts/memory/local.py @@ -0,0 +1,114 @@ +import dataclasses +import orjson +from typing import Any, List, Optional +import numpy as np +import os +from memory.base import MemoryProviderSingleton, get_ada_embedding + + +EMBED_DIM = 1536 +SAVE_OPTIONS = orjson.OPT_SERIALIZE_NUMPY | orjson.OPT_SERIALIZE_DATACLASS + + +def create_default_embeddings(): + return np.zeros((0, EMBED_DIM)).astype(np.float32) + + +@dataclasses.dataclass +class CacheContent: + texts: List[str] = dataclasses.field(default_factory=list) + embeddings: np.ndarray = dataclasses.field( + default_factory=create_default_embeddings + ) + + +class LocalCache(MemoryProviderSingleton): + + # on load, load our database + def __init__(self, cfg) -> None: + self.filename = f"{cfg.memory_index}.json" + if os.path.exists(self.filename): + with open(self.filename, 'rb') as f: + loaded = orjson.loads(f.read()) + self.data = CacheContent(**loaded) + else: + self.data = CacheContent() + + def add(self, text: str): + """ + Add text to our list of texts, add embedding as row to our + embeddings-matrix + + Args: + text: str + + Returns: None + """ + if 'Command Error:' in text: + return "" + self.data.texts.append(text) + + embedding = get_ada_embedding(text) + + vector = np.array(embedding).astype(np.float32) + vector = vector[np.newaxis, :] + self.data.embeddings = np.concatenate( + [ + vector, + self.data.embeddings, + ], + axis=0, + ) + + with open(self.filename, 'wb') as f: + out = orjson.dumps( + self.data, + option=SAVE_OPTIONS + ) + f.write(out) + return text + + def clear(self) -> str: + """ + Clears the redis server. + + Returns: A message indicating that the memory has been cleared. + """ + self.data = CacheContent() + return "Obliviated" + + def get(self, data: str) -> Optional[List[Any]]: + """ + Gets the data from the memory that is most relevant to the given data. + + Args: + data: The data to compare to. + + Returns: The most relevant data. + """ + return self.get_relevant(data, 1) + + def get_relevant(self, text: str, k: int) -> List[Any]: + """" + matrix-vector mult to find score-for-each-row-of-matrix + get indices for top-k winning scores + return texts for those indices + Args: + text: str + k: int + + Returns: List[str] + """ + embedding = get_ada_embedding(text) + + scores = np.dot(self.data.embeddings, embedding) + + top_k_indices = np.argsort(scores)[-k:][::-1] + + return [self.data.texts[i] for i in top_k_indices] + + def get_stats(self): + """ + Returns: The stats of the local cache. + """ + return len(self.data.texts), self.data.embeddings.shape diff --git a/scripts/memory/pinecone.py b/scripts/memory/pinecone.py new file mode 100644 index 0000000000..8e1eaa570f --- /dev/null +++ b/scripts/memory/pinecone.py @@ -0,0 +1,51 @@ + +import pinecone + +from memory.base import MemoryProviderSingleton, get_ada_embedding + + +class PineconeMemory(MemoryProviderSingleton): + def __init__(self, cfg): + pinecone_api_key = cfg.pinecone_api_key + pinecone_region = cfg.pinecone_region + pinecone.init(api_key=pinecone_api_key, environment=pinecone_region) + dimension = 1536 + metric = "cosine" + pod_type = "p1" + table_name = "auto-gpt" + # this assumes we don't start with memory. + # for now this works. + # we'll need a more complicated and robust system if we want to start with memory. + self.vec_num = 0 + if table_name not in pinecone.list_indexes(): + pinecone.create_index(table_name, dimension=dimension, metric=metric, pod_type=pod_type) + self.index = pinecone.Index(table_name) + + def add(self, data): + vector = get_ada_embedding(data) + # no metadata here. We may wish to change that long term. + resp = self.index.upsert([(str(self.vec_num), vector, {"raw_text": data})]) + _text = f"Inserting data into memory at index: {self.vec_num}:\n data: {data}" + self.vec_num += 1 + return _text + + def get(self, data): + return self.get_relevant(data, 1) + + def clear(self): + self.index.delete(deleteAll=True) + return "Obliviated" + + def get_relevant(self, data, num_relevant=5): + """ + Returns all the data in the memory that is relevant to the given data. + :param data: The data to compare to. + :param num_relevant: The number of relevant data to return. Defaults to 5 + """ + query_embedding = get_ada_embedding(data) + results = self.index.query(query_embedding, top_k=num_relevant, include_metadata=True) + sorted_results = sorted(results.matches, key=lambda x: x.score) + return [str(item['metadata']["raw_text"]) for item in sorted_results] + + def get_stats(self): + return self.index.describe_index_stats() diff --git a/scripts/memory/redismem.py b/scripts/memory/redismem.py new file mode 100644 index 0000000000..2082fe5887 --- /dev/null +++ b/scripts/memory/redismem.py @@ -0,0 +1,143 @@ +"""Redis memory provider.""" +from typing import Any, List, Optional +import redis +from redis.commands.search.field import VectorField, TextField +from redis.commands.search.query import Query +from redis.commands.search.indexDefinition import IndexDefinition, IndexType +import numpy as np + +from memory.base import MemoryProviderSingleton, get_ada_embedding + + +SCHEMA = [ + TextField("data"), + VectorField( + "embedding", + "HNSW", + { + "TYPE": "FLOAT32", + "DIM": 1536, + "DISTANCE_METRIC": "COSINE" + } + ), +] + + +class RedisMemory(MemoryProviderSingleton): + def __init__(self, cfg): + """ + Initializes the Redis memory provider. + + Args: + cfg: The config object. + + Returns: None + """ + redis_host = cfg.redis_host + redis_port = cfg.redis_port + redis_password = cfg.redis_password + self.dimension = 1536 + self.redis = redis.Redis( + host=redis_host, + port=redis_port, + password=redis_password, + db=0 # Cannot be changed + ) + self.cfg = cfg + if cfg.wipe_redis_on_start: + self.redis.flushall() + try: + self.redis.ft(f"{cfg.memory_index}").create_index( + fields=SCHEMA, + definition=IndexDefinition( + prefix=[f"{cfg.memory_index}:"], + index_type=IndexType.HASH + ) + ) + except Exception as e: + print("Error creating Redis search index: ", e) + existing_vec_num = self.redis.get(f'{cfg.memory_index}-vec_num') + self.vec_num = int(existing_vec_num.decode('utf-8')) if\ + existing_vec_num else 0 + + def add(self, data: str) -> str: + """ + Adds a data point to the memory. + + Args: + data: The data to add. + + Returns: Message indicating that the data has been added. + """ + if 'Command Error:' in data: + return "" + vector = get_ada_embedding(data) + vector = np.array(vector).astype(np.float32).tobytes() + data_dict = { + b"data": data, + "embedding": vector + } + pipe = self.redis.pipeline() + pipe.hset(f"{self.cfg.memory_index}:{self.vec_num}", mapping=data_dict) + _text = f"Inserting data into memory at index: {self.vec_num}:\n"\ + f"data: {data}" + self.vec_num += 1 + pipe.set(f'{self.cfg.memory_index}-vec_num', self.vec_num) + pipe.execute() + return _text + + def get(self, data: str) -> Optional[List[Any]]: + """ + Gets the data from the memory that is most relevant to the given data. + + Args: + data: The data to compare to. + + Returns: The most relevant data. + """ + return self.get_relevant(data, 1) + + def clear(self) -> str: + """ + Clears the redis server. + + Returns: A message indicating that the memory has been cleared. + """ + self.redis.flushall() + return "Obliviated" + + def get_relevant( + self, + data: str, + num_relevant: int = 5 + ) -> Optional[List[Any]]: + """ + Returns all the data in the memory that is relevant to the given data. + Args: + data: The data to compare to. + num_relevant: The number of relevant data to return. + + Returns: A list of the most relevant data. + """ + query_embedding = get_ada_embedding(data) + base_query = f"*=>[KNN {num_relevant} @embedding $vector AS vector_score]" + query = Query(base_query).return_fields( + "data", + "vector_score" + ).sort_by("vector_score").dialect(2) + query_vector = np.array(query_embedding).astype(np.float32).tobytes() + + try: + results = self.redis.ft(f"{self.cfg.memory_index}").search( + query, query_params={"vector": query_vector} + ) + except Exception as e: + print("Error calling Redis search: ", e) + return None + return [result.data for result in results.docs] + + def get_stats(self): + """ + Returns: The stats of the memory index. + """ + return self.redis.ft(f"{self.cfg.memory_index}").info() diff --git a/scripts/speak.py b/scripts/speak.py index f6242a37dd..13517d366f 100644 --- a/scripts/speak.py +++ b/scripts/speak.py @@ -42,7 +42,7 @@ def say_text(text, voice_index=0): if not cfg.elevenlabs_api_key: gtts_speech(text) else: - success = eleven_labs_speech(text) + success = eleven_labs_speech(text, voice_index) if not success: gtts_speech(text)