An intelligent AI-powered assistant that helps you understand and analyze data models through natural conversation. Built with Streamlit and powered by OpenAI's language models.

• PDF Document Processing: Automatically extracts and processes text from uploaded PDF files
• Conversational Interface: Natural dialogue with context-aware responses
• Vector Database: Uses ChromaDB with OpenAI embeddings for efficient storage/retrieval
• Streaming Responses: Real-time answer generation with typing indicator
• Session Persistence: Maintains conversation history and document context
• Debug Tools: Inspect retrieved document chunks during conversations

1. Clone Repository

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

2. Install Dependencies

   pip install -r requirements.txt

3. Configure Environment

   echo 'OPENAI_API_KEY="your-api-key-here"' > .env

4. Organize PDFs

   mkdir -p data  # Place PDF files here

5. Launch Application

   streamlit run app.py

• Python 3.10+
• OpenAI API key
• PDF documents in data/ directory

• Framework: Streamlit
• Language Model: GPT-4 via OpenAI API
• Embeddings: text-embedding-3-large
• Vector Database: ChromaDB
• PDF Processing: PyPDFDirectoryLoader

The application follows these key steps:

1. Document Ingestion:

   • Load PDFs from data/ directory
   • Split documents into 2000-character chunks
   • Generate vector embeddings for each chunk

2. Vector Storage:

   • Store embeddings in ChromaDB with persistent storage
   • Enable efficient similarity search

3. Conversation Flow:

   • Maintain chat history in session state
   • Combine retrieved documents with LLM context
   • Stream responses character-by-character

4. Error Handling:

   • Validate PDF directory existence
   • Catch and display processing errors
   • Graceful handling of missing credentials

Contributions are welcome! Please feel free to submit pull requests.

This project is licensed under the terms included in the LICENSE file.

• Ensure all dependencies are correctly installed
• Check that your OpenAI API key is valid and properly set in the .env file
• For PDF processing issues, ensure your PDFs are text-based and not scanned images
• If the vector store isn't loading, check file permissions and storage space