Electric RAG - Advanced Retrieval-Augmented Generation System

📌 Overview

Electric RAG is a sophisticated Retrieval-Augmented Generation (RAG) system that combines document processing, semantic search, and multi-LLM orchestration to provide intelligent, context-aware responses. The system features a FastAPI backend with advanced document ingestion, embedding generation, and a React frontend for seamless user interaction.

UI

Chat UI

Document Ingestion

🚀 Key Features

🔍 Advanced RAG Pipeline: Complete document ingestion, chunking, embedding, and retrieval system
🧠 Multi-LLM Orchestration: Support for multiple language models with jury-based decision making
📄 Multi-Format Document Processing: PDF parsing with OCR support, text extraction, and JSON storage
🎯 Intelligent Chunking: Context-aware document segmentation with configurable overlap
🔗 Chain of Thought Processing: Enhanced reasoning capabilities with iterative thinking
🖼️ Image-Aware Processing: OCR-based image extraction and multi-modal content handling
📊 Vector Database Integration: Supabase vector database with analytics and monitoring
🌐 Modern Web Interface: React frontend with Tailwind CSS for optimal user experience
⚡ FastAPI Backend: High-performance API with CORS support and comprehensive endpoints

🛠️ Tech Stack

Backend

Framework: FastAPI with Uvicorn
Language Models: GPT-4, Claude-3, Gemini-Pro, Llama-2, Mistral-7B
Embeddings: Sentence Transformers (all-mpnet-base-v2)
Vector Database: Supabase Vector Database
Document Processing: PyMuPDF, Mistral OCR
Chunking: Advanced sentence and page-based algorithms
Reranking: TF-IDF, keyword matching, and vector similarity

Frontend

Framework: React 19 with modern hooks
Styling: Tailwind CSS
Icons: Lucide React
Build Tool: Create React App

📂 Project Structure

RAG/
├── backend/                          # FastAPI backend application
│   ├── chunking/                     # Document chunking algorithms
│   │   ├── chunker.py               # Main chunking interface
│   │   └── chunkers/                # Specific chunking implementations
│   ├── config/                       # Configuration settings
│   │   └── settings.py              # Global configuration parameters
│   ├── data/                         # Data storage and processing
│   │   ├── raw_document_data/       # Temporary document storage
│   │   └── raw_jsons/               # Processed JSON outputs
│   ├── database/                     # Database layer
│   │   ├── relational/               # Relational database components
│   │   └── vector/                   # Vector database integration
│   │       ├── vectorDB.py          # Vector database interface
│   │       └── vectorDBs/           # Specific vector DB implementations
│   ├── embedding/                    # Embedding generation
│   │   ├── encoder.py               # Main encoder interface
│   │   └── encoders/                # Specific encoder implementations
│   ├── generation/                   # LLM generation components
│   ├── helpers/                      # Utility helper modules
│   ├── ingestion/                    # Document ingestion pipeline
│   │   └── ingestion.py             # Main ingestion orchestrator
│   ├── llmservice/                   # LLM service layer
│   │   ├── multillmorchestrator.py  # Multi-LLM orchestration
│   │   ├── adaptiveJsonExtractor.py # JSON response extraction
│   │   ├── llmmodels.py             # LLM model definitions
│   │   └── prompts.py               # System prompts
│   ├── parsing/                      # Document parsing
│   │   ├── parser.py                # Main parser interface
│   │   └── parsers/                 # Specific parser implementations
│   ├── retriever/                    # Retrieval system
│   │   ├── retriever.py             # Main retrieval logic
│   │   └── reranker/                # Reranking algorithms
│   ├── utils/                        # Utility functions
│   │   └── logger.py                # Logging configuration
│   └── main.py                      # FastAPI application entry point
├── frontend/                         # React frontend application
│   └── electric-rag/                # Main frontend package
│       ├── src/                     # React source code
│       ├── public/                  # Static assets
│       └── package.json             # Frontend dependencies
├── requirements.txt                  # Python dependencies
└── README.md                        # Project documentation

✅ Core Functionality

Document Processing

Multi-format Support: PDF, text, and scanned documents
OCR Integration: Mistral OCR for image-based content extraction
Smart Chunking: Configurable chunk sizes with overlap optimization
Content Preservation: Maintains document structure and formatting

Retrieval System

Semantic Search: Vector similarity-based document retrieval
Multi-modal Retrieval: Text and image content matching
Reranking: Combines TF-IDF, keyword matching, and vector similarity
Configurable Top-K: Adjustable number of retrieved chunks

LLM Orchestration

Multi-LLM Support: Integration with multiple language model providers
Jury-based Decisions: Consensus-based response generation
Chain of Thought: Iterative reasoning for complex queries
Expert LLM Selection: Specialized model selection for specific tasks

API Endpoints

Chat Interface: /api/chat - Main conversation endpoint
Document Upload: /api/upload - File ingestion with optional training
Analytics: /api/analytics - Vector database statistics
Health Check: /api/health - System status monitoring
LLM Models: /api/llm-models - Available model information

🧪 Getting Started

Prerequisites

Python 3.8+
Node.js 16+ (for frontend)
Supabase account (for vector database)
API keys for LLM providers

Backend Setup

Clone the repository:

git clone <repository-url>
cd RAG/backend

Install Python dependencies:
```
pip install -r requirements.txt
```

Configure environment variables: Create a .env file with:

ENCODER_NAME=sentence_encoder
ENCODING_MODEL=all-mpnet-base-v2
PARSER=mistral_ocr
CHUNKER=group_embedding

Start the backend server:
```
python main.py
```
The API will be available at http://localhost:8000

Frontend Setup

Navigate to frontend directory:
```
cd ../frontend/electric-rag
```
Install dependencies:
```
npm install
```
Start the development server:
```
npm start
```
The frontend will be available at http://localhost:3000

🔧 Configuration

Backend Settings (`backend/config/settings.py`)

Chunking: DEFAULT_CHUNK_LENGTH, DEFAULT_OVER_LAP
Retrieval: TOP_K, SIMILARITY_TH
Processing: DEFAULT_TEMPERATURE, DEFAULT_THINKING_ITERATIONS
Storage: TEMP_FILES_DOC, TEMP_FILES_JSONS

Vector Database

Configure Supabase connection in backend/database/vector/vectorDBs/supabasevdb.py
Set up appropriate indexes for optimal search performance

LLM Providers

Configure API keys for supported providers
Adjust model selection in backend/llmservice/llmmodels.py

📊 API Usage

Chat Request

{
  "query": "What is the main topic of the document?",
  "multiLLM": false,
  "fetchChains": true,
  "noOfNeighbours": 5,
  "activeLLMs": [],
  "expertLLM": "GPT-4",
  "chainOfThought": true
}

Document Upload

curl -X POST "http://localhost:8000/api/upload" \
  -F "files=@document.pdf" \
  -F "train_model=false"

🚀 Deployment

Production Considerations

Use proper environment variable management
Implement authentication and authorization
Set up monitoring and logging
Configure CORS origins for production domains
Use production-grade vector database instances

Docker Support

Backend can be containerized with FastAPI
Frontend can be built and served statically
Use environment-specific configuration files

🤝 Contributing

Fork the repository
Create a feature branch (git checkout -b feature/amazing-feature)
Commit your changes (git commit -m 'Add amazing feature')
Push to the branch (git push origin feature/amazing-feature)
Open a Pull Request

Development Guidelines

Follow PEP 8 for Python code
Use TypeScript for frontend components
Add comprehensive tests for new features
Update documentation for API changes

📞 Support

Issues: Create GitHub issues for bugs or feature requests
Discussions: Use GitHub Discussions for questions and ideas
Documentation: Check the codebase for implementation details

📝 License

This project is licensed under the MIT License - see the LICENSE file for details.

🔮 Roadmap

Enhanced multi-modal support
Advanced reranking algorithms
Real-time collaboration features
Mobile application
Advanced analytics dashboard
Plugin system for custom processors

Name		Name	Last commit message	Last commit date
Latest commit History 24 Commits
.vscode		.vscode
backend		backend
frontend/electric-rag		frontend/electric-rag
.gitignore		.gitignore
FEATURE_ANALYSIS.md		FEATURE_ANALYSIS.md
README.md		README.md
USER_MANUAL.md		USER_MANUAL.md
requirements.txt		requirements.txt

Kathan-max/RAG-Enhanced-Chatbot-with-LoRA-Fine-Tuning

Folders and files

Latest commit

History

Repository files navigation