ContextWeaver AI

Project Description

ContextWeaver AI is a FastAPI-based project designed to build a context-based AI agent leveraging LangGraph and PydanticAI. The agent is capable of searching the web for information and answering queries based on the gathered data. It incorporates advanced features such as token budgeting, maintaining conversation context, and dynamic decision-making across subgraphs.

Key Features

1. Detailed Reporting: Provides a comprehensive report on any topic by searching the web.
2. Knowledge Base Creation: Creates a knowledge base from shared URLs using an efficient chunking strategy.
3. Context Preservation: Maintains the context of each conversation for seamless interactions.
4. QA Agent Integration: Uses a QA agent in each graph to validate answers and loops the process if the QA agent is not satisfied for each subgraph.
5. Dynamic Subgraph Switching: Dynamically decides which subgraph to activate based on the query and context.
6. Dockerized: The application is containerized using Docker for easy deployment and scalability.
7. streamlit frontend chatbot: Provides a user-friendly interface for interaction with the bot.

How to Run the Project

Prerequisites

Ensure you have the following installed on your system:

• Python 3.8 or higher
• pip (Python package manager)

Installation Steps

1. Clone the repository:

   git clone <repository-url>
   cd subgraph-web-bot

2. Create a virtual environment:

   python -m venv venv
   source venv/bin/activate  # On Windows: venvScriptsactivate

3. Install the required dependencies:

   pip install -r requirements.txt

4. Set up environment variables:

   • Copy the env.sample file to .env:

     cp env.sample .env

   • Update the .env file with your API keys and configuration.

Running the Application

1. Start the FastAPI server:

   uvicorn app.main:app --reload

2. Access the API documentation at:

   http://127.0.0.1:8000/docs
   

Example Usage

• Use the query endpoint to perform a web crawl and get detailed reports.
• Use the project-details endpoint to get project metadata.
• Use the health endpoint to check the application's health.

Future Enhancements

• Improved token budgeting strategies.
• Enhanced subgraph decision-making algorithms.
• Support for additional AI models and APIs.

🤝 Contributing

We welcome contributions from everyone! Please read our Contributing Guide to get started. ✨

License

This project is licensed under the MIT License.