AI-Powered PDF Context Retrieval Chatbot (RAG) 🤖📄

Repository: Ashprogrammer29/AI-Powered-PDF-Context-Retrieval-Chatbot-RAG

Unlock intelligent context retrieval and querying from your PDFs using state-of-the-art AI! This Retrieval-Augmented Generation (RAG) chatbot leverages FastAPI, LangChain, Google Gemini, and powerful vector search for document Q&A

---

🗂️ Files & Structure

File/Folder	Description
PDF Context Retrieval Chatbot.ipynb	Main Jupyter notebook: code, API, model setup, PDF ingestion, and querying logic
requirements[1].txt	Python dependencies (FastAPI, LangChain, vector DB, Google GenAI, etc.)
LICENSE	Boost Software License v1.0 (see below)

Key Notebook Functions:

• PDF upload & text extraction (get_pdf_text)
• Text chunking & vector store creation (create_vectorstore)
• API endpoints (/process)
• Utility scripts: file/folder handling, model config, embedding setup
• Uses FastAPI for serving endpoints

---

⚡️ Quick Setup

1. Clone the repo

   git clone https://github.com/Ashprogrammer29/AI-Powered-PDF-Context-Retrieval-Chatbot-RAG.git
   cd AI-Powered-PDF-Context-Retrieval-Chatbot-RAG

2. Install dependencies

   python3 -m venv venv
   source venv/bin/activate
   pip install -r requirements[1].txt

3. Configure Model

   • Setup your Google Gemini API key (via notebook or environment)
   • Place model configs/weights as needed

---

🚦 Running the Chatbot

📝 Jupyter Notebook

• Open PDF Context Retrieval Chatbot.ipynb
• Step through: mount storage, set up API keys, upload PDFs, run cells for context retrieval & Q&A

🌐 API Server (FastAPI)

• Inside the notebook, the FastAPI app is initialized; you can run locally using:

  uvicorn main:app --reload

• Endpoints available for PDF upload & query

---

📡 API Endpoints

1. Upload PDF

• Endpoint: /process
• Method: POST
• Payload: JSON (see notebook's File model)

{
  "files": ["https://example.com/file1.pdf", ...],
  "rewrite": true
}

2. Query Context

• Use the vectorstore and Q&A logic in the notebook to ask questions about uploaded PDFs.

---

🧠 Model & Config Notes

• LLM: Google Gemini (via langchain-google-genai, API key required)
• Embeddings: GoogleGenerativeAIEmbeddings + FAISS for semantic search
• PDF Parsing: PyMuPDF (pymupdf)
• Text Splitting: RecursiveCharacterTextSplitter from LangChain
• API Models: Pydantic-based request bodies
• Configurable: Chunk size, rewrite mode, user/session IDs, etc.

---

📒 Example Usage

1. Upload PDFs via /process endpoint or notebook cell
2. Ask questions!: "What is the summary of the document?" or "Find the legal clause about termination."
3. Get answers with full context, citations, and semantic retrieval from your documents.

---

🎯 Use Cases

• Legal Document Q&A ⚖️
  Instantly find clauses, obligations, or summaries from contracts and agreements.
• Academic Research Assistant 🎓
  Extract findings, definitions, and references from research papers.
• Business Report Analysis 📊
  Query for revenue, trends, and executive summaries in reports.
• Technical Manual & FAQ Chatbot 🛠️
  Retrieve procedures and troubleshooting from manuals.
• Compliance & Policy Checking 🏢
  Automate policy and HR queries from company documents.
• Customer Support Automation 💬
  Answer product, feature, or troubleshooting questions from help docs.
• Onboarding & Training 👩‍💼
  Enable instant answers for new employee training materials.

---

🙌 Contributing

Pull requests, issues, and suggestions are welcome! 🎉

---

📜 License

This project is licensed under the Boost Software License - Version 1.0 - August 17th, 2003.

See the LICENSE file for details.

---

💡 Tech Stack

• FastAPI 🚀
• LangChain 🦜
• Google Gemini 🤖
• FAISS/Vector DB 🔍
• PyMuPDF 📄
• Pydantic 🛠️

---

🌟 Acknowledgements

Big thanks to the open-source AI/NLP community, LangChain, Google, and all contributors!

---

Made with ❤️ by Ashprogrammer29