RAG (Retrieval Augmented Generation) System

Video Tutorial (Portuguese)

This repository is part of Bix Tech "Semana de Dados" a.k.a Data Week. For a further explanation aboutwhat is RAG and this tutorial, watch the video below:

🎥 Assista ao tutorial no YouTube

---

A Question-Answering system built using LangChain and ChromaDB that allows users to query their documents using natural language. The system uses OpenAI's language models to provide context-aware answers based on the content of the indexed documents.

Overview

This system allows you to:

• Index text documents
• Create and manage a vector store using ChromaDB
• Interact with your documents through a chat interface
• Get answers with source references
• View and manage the document store

Repository Structure

rag_system/
├── documents/         # Place your text files here
├── logs/             # System logs are stored here
├── db/               # Vector store database (created automatically)
├── core.py           # Core RAG system implementation
├── interface.py      # Interactive command-line interface
├── requirements.txt  # Project dependencies
└── .env             # Environment variables configuration

Setup Instructions

1. Create a Virtual Environment

For Windows:

# Create a virtual environment
python -m venv .venv

# Activate the virtual environment
.venv\Scripts\activate

For Linux/Mac:

# Create a virtual environment
python -m venv .venv

# Activate the virtual environment
source .venv/bin/activate

2. Install Dependencies

# Install all required packages
pip install -r requirements.txt

3. Environment Configuration

Create a .env file in the root directory with the following content:

OPENAI_API_KEY=your-api-key-here
MODEL_NAME=gpt-3.5-turbo
COLLECTION_NAME=my_documents
PERSIST_DIRECTORY=db

Replace your-api-key-here with your actual OpenAI API key.

4. Prepare Documents

1. Create a documents directory if it doesn't exist:

mkdir documents

2. Place your text files (.txt) in the documents directory. These are the documents that will be indexed and used for answering questions.

5. Running the Application

# Run the interactive interface
python interface.py

Using the System

Once running, the system provides the following options:

1. Index documents: Processes and indexes all text files in the documents directory
2. Check total number of documents: Shows how many documents are currently indexed
3. Delete document store: Removes all indexed documents
4. Start RAG chat: Begins an interactive Q&A session
5. Exit: Closes the application

Chat Commands

When in chat mode:

• Type your questions normally and press Enter
• Type 'sources' to see detailed source documents for the last answer
• Type 'quit', 'exit', or 'q' to return to the main menu

Requirements

• Python 3.8 or higher
• OpenAI API key
• Sufficient disk space for document storage
• Internet connection for API access

Dependencies

Main libraries used:

• langchain
• langchain-openai
• langchain-community
• langchain-chroma
• chromadb
• python-dotenv
• rich

Troubleshooting

If you encounter any issues:

1. Check the logs in the logs directory
2. Ensure your OpenAI API key is valid
3. Verify that your documents are text files (.txt)
4. Make sure all required directories exist
5. Check your internet connection

Additional Notes

• The system creates necessary directories automatically
• Logs are timestamped and stored in the logs directory
• The vector store is persistent and stored in the db directory
• All text files should be in UTF-8 or compatible encoding

For any other issues or questions, please refer to the logs or create an issue in the repository.