RAG Advanced

A Retrieval-Augmented Generation (RAG) system that processes PDF documents and provides intelligent question-answering capabilities using embeddings, vector search, and large language models.

Features

• PDF Processing: Extract and chunk text from PDF documents using intelligent splitting strategies
• Vector Embeddings: Generate semantic embeddings using SentenceTransformers
• Vector Storage: Store and query embeddings using ChromaDB
• Query Augmentation: Enhance queries with hypothetical answers for improved retrieval
• LLM Integration: Generate responses using Mistral AI models
• Visualization: Plot embedding spaces with UMAP dimensionality reduction

Requirements

• Python >=3.13
• Mistral API key

Installation

1. Clone the repository:

git clone <repository-url>
cd rag-advanced

2. Install dependencies using uv (recommended):

uv sync

Or using pip:

pip install -e .

3. Set up environment variables:

cp .env.example .env
# Edit .env and add your MISTRAL_API_KEY

Usage

Basic Usage

Run the main RAG pipeline:

python main.py

This will:

1. Process the NVIDIA annual report PDF in data/
2. Split text into chunks and generate embeddings
3. Store embeddings in ChromaDB
4. Execute a sample query with query augmentation
5. Generate and display the answer
6. Create a visualization of the embedding space

Customizing Queries

Edit the original_query variable in main.py:112 to ask different questions:

original_query = "What was the total profit for the year, and how does it compare to the previous year?"

Adding Your Own Documents

1. Place PDF files in the data/ directory
2. Update the file path in main.py:42:

reader = PdfReader("data/your-document.pdf")

Project Structure

├── data/                           # PDF documents for processing
│   └── NVIDIA-2025-Annual-Report.pdf
├── main.py                         # Main RAG pipeline
├── helper_utils.py                 # Utility functions
├── .env.example                    # Environment variables template
├── .python-version                 # Python version specification
├── AGENTS.md                       # Guidelines for AI coding agents
├── pyproject.toml                  # Project dependencies
├── uv.lock                         # Dependency lock file
└── README.md                       # This file

How It Works

1. Document Processing: PDFs are loaded and text is extracted
2. Text Chunking: Documents are split into manageable chunks using recursive character splitting and token-based splitting
3. Embedding Generation: Each chunk is converted to vector embeddings using SentenceTransformers
4. Vector Storage: Embeddings are stored in ChromaDB for efficient similarity search
5. Query Augmentation: Original queries are enhanced with hypothetical answers to improve retrieval
6. Retrieval: Most relevant document chunks are retrieved using similarity search
7. Answer Generation: Mistral AI generates contextual answers based on retrieved chunks
8. Visualization: UMAP creates 2D projections of the embedding space for analysis

Environment Variables

• MISTRAL_API_KEY: Your Mistral AI API key (required)
• TOKENIZERS_PARALLELISM: Set to false to avoid tokenizer warnings

Troubleshooting

• Ensure your Mistral API key is valid and has sufficient credits
• Check that PDF files are readable and not encrypted
• Verify Python version compatibility (>=3.13 required)
• If you encounter import errors, run uv sync to install all dependencies
• For tokenizer warnings, set TOKENIZERS_PARALLELISM=false in your environment