CRAB Contextual Retrieval And Analysis Bot

CRAB: Contextual Retrieval and Analysis Bot

CRAB (Contextual Retrieval and Analysis Bot) is a sophisticated AI-powered conversational system that combines Retrieval-Augmented Generation (RAG), state machine-based conversation flow, and various AI tools to provide intelligent responses and perform file system operations.

Features

• RAG-based question answering using local documents and optional web search
• State machine-driven conversation flow
• File system operations (list directory, change directory, read files)
• Flexible mode switching between standard and web search
• Integration with ChromaDB for efficient document retrieval
• Sentence embedding for semantic search
• Logging and error handling
• Integration with LangSmith for tracing and monitoring

Installation

1. Clone the repository:

   git clone https://github.com/yourusername/CRAB.git
   cd CRAB

2. Install the required dependencies:

   pip install -r requirements.txt

3. Set up environment variables: Create a `.env` file in the project root and add your API keys:

   OPENAI_API_KEY=your_openai_api_key_here

Usage

To start CRAB, run:

python main.py

Once CRAB is running, you can:

• Ask questions and receive RAG-enhanced responses
• Use file system commands like "list files", "change directory", "read file"
• Switch modes between standard and web search using "set mode standard" or "set mode web_search"
• Type "exit" to end the chat session

Project Structure

• main.py: Main script containing the CRAB implementation
• requirements.txt: List of Python dependencies
• .env: Environment variables (create this file and add your API keys)

Customization

• Modify the RAG class to adjust retrieval and generation strategies
• Extend the GeneralInterpreter class to add new tools or conversation states
• Adjust the ChromaDBRetriever class to change document indexing and retrieval methods

Contributing

Contributions to CRAB