GeminiLLM_DocChat

Overview

GeminiLLM_DocChat is an AI-powered document chat system that extracts text from PDFs, generates vector embeddings, stores them in a vector database, and allows users to query information using natural language. It leverages LangChain, FAIS and google's embeddings to provide an efficient document retrieval and question-answering system.

Features

• 📄 Extracts text from PDFs, PPTs etc
• 🔍 Stores embeddings in a vector database (FAISS)
• 🤖 Uses AI embeddings for semantic search
• 💬 Allows users to ask natural language questions about the document
• 🚀 Fast and scalable retrieval-augmented generation (RAG)

Tech Stack

• Python (Core language)
• LangChain (Document processing and retrieval)
• FAISS (Vector database)
• google/models/text-embedding-004 (Embeddings model)
• Docling (PDF text extraction)
• Streamlit (API or UI for interaction)

Installation

1. Clone the repository:

   git clone https://github.com/SPYLoveC2/GeminiLLM_DocChat.git
   cd GeminiLLM_DocChat

2. Create a virtual environment and install dependencies:

conda env create -f environment.yml

Usage

1. Start the chatbot interface:

   streamlit run app.py  # If using Streamlit UI

2. Ask questions about the document in the UI or API.

Configuration

• Set your API keys (if using OpenAI embeddings) in an .env file:

GOOGLE_API_KEY=your_api_key GROQ=your_api_key

## Roadmap
- ✅ Basic PDF processing & embedding storage
- ✅ Q&A system with retrieval
- 🔜 Multi-document support
- 🔜 Fine-tuned LLM integration

## Contributing
Feel free to submit issues or pull requests to improve the project.