Major Project: End-to-End PDF QA Chatbot

This project is an end-to-end system for querying PDF documents using a chatbot interface. It combines a modern Next.js frontend with Python microservices for document processing, retrieval, and LLM-based question answering. The system is designed for extensibility and performance, supporting scalable document ingestion and semantic search.

Techniques Used

• Server Components in Next.js: Efficient rendering and data fetching (MDN: Server Components).
• TypeScript for Type Safety: Ensures robust code and easier refactoring (MDN: TypeScript).
• Custom Hooks and Utility Functions: Modularizes logic for reusability.
• Python Microservices: Decouples document processing and QA logic for scalability.
• Chroma Vector Database: Enables fast semantic search over document embeddings (ChromaDB).
• PDF Parsing and Embedding: Processes and indexes PDF content for retrieval.
• LLM Integration: Uses language models for natural language question answering.

Notable Technologies & Libraries

• Next.js (React framework)
• ChromaDB (vector database)
• LangChain (LLM orchestration)
• FastAPI (Python web framework)
• PyPDF2 (PDF parsing)
• Tailwind CSS (utility-first CSS)
• Vercel (deployment platform)
• PostCSS (CSS processing)
• ESLint (code linting)
• TypeScript (typed JavaScript)
• React (UI library)

Project Structure

.
├── pdf-qa-chatbot/
│   ├── public/
│   ├── src/
│   │   ├── app/
│   │   ├── lib/
├── Python_Microservices_Be/
│   ├── chroma_db_legal/
│   ├── uploads/

• pdf-qa-chatbot/public/: Contains SVG assets for UI.
• pdf-qa-chatbot/src/app/: Next.js app directory, includes global styles and layout.
• pdf-qa-chatbot/src/lib/: Utility functions for frontend logic.
• Python_Microservices_Be/chroma_db_legal/: ChromaDB vector store and metadata.
• Python_Microservices_Be/uploads/: Uploaded PDF documents for processing.

Fonts

No custom fonts detected; uses system or default web fonts.

---

Next.js Frontend Setup

1. Install dependencies:

   cd pdf-qa-chatbot
   npm install

2. Run the development server:

   npm run dev

   The app will be available at http://localhost:3000.

---

Python Microservices Setup

1. Create a virtual environment:

   python -m venv venv

2. Activate the environment:
   • On Windows:

     venv\Scripts\activate

   • On macOS/Linux:

     source venv/bin/activate

3. Install dependencies:

   pip install -r Python_Microservices_Be/requirements.txt

---

Ollama Setup & Model Preparation

1. Install Ollama Desktop:

   • Download and install from Ollama Desktop.

2. Download GGUF Model from Hugging Face:

   • Visit Indian-LegalBot-Llama-3.1-8B-GGUF.
   • Download the recommended version: Q4_K_S or Q4_K_M.

3. Convert GGUF Model for Ollama Compatibility:

   • Create a Modfile in your model directory with the following content:

     FROM ./Indian-LegalBot-Llama-3.1-8B-Q4_K_S.gguf
     PARAMETER stop "<|eot_id|>"
     

   • Replace the filename with your downloaded GGUF file.
   • Build the model for Ollama:

     ollama create indian-legalbot -f Modfile

   • The model is now available for use with Ollama.

4. Update Python Microservices Configuration:

   • Open Python_Microservices_Be/config.py.
   • Set the model name:

     LLM_MODEL_NAME = "indian-legalbot"

   • This ensures your microservices use the correct Ollama model.

---

...

Next.js Frontend Setup

1. Navigate to the frontend directory:

   cd pdf-qa-chatbot

2. Install dependencies:

   npm install

3. Start the development server:

   npm run dev

   The application will be available at http://localhost:3000 ...

...

Starting Python Microservices

1. Activate your virtual environment:

   • On Windows:

     venv\Scripts\activate

   • On macOS/Linux:

     source venv/bin/activate

2. Start the microservices:

   flask run --port=5001

API Endpoints

1. POST /api/upload

BODY Example:

{
    "collection_name": "legal_case_a1b2c3d4e5",
    "message": "File 'sample_case.pdf' processed successfully."
}

2. POST /api/ask_rag

BODY Example:

{
  "question": "What is the main subject of this document?",
  "collection_name": "legal_case_a1b2c3d4e5"
}

3. POST api/ask_direct

BODY Example:

{
  "question": "Explain the concept of 'audi alteram partem' in Indian law."
}

...