RC Data Extractor

A production-ready API service for extracting structured data from Registration Certificate (RC) images using OCR and LLM processing.

Features

• Image Processing: Extract text from RC front and back images using Tesseract.js (pure JavaScript OCR)
• LLM Integration: Process extracted text with Ollama LLM for structured data extraction
• RESTful API: Clean, well-documented API endpoints
• Production Ready: Error handling, logging, security headers, graceful shutdown
• TypeScript: Fully typed codebase for better development experience

Prerequisites

• Node.js (v18 or higher)
• npm or yarn
• Ollama running locally (for LLM processing)

Setting up Ollama

1. Install Ollama from https://ollama.ai
2. Pull the required model:

ollama pull llama3.1:8b

Installation

1. Clone the repository:

git clone https://gitlabsdev.abhibus.com/research/rc-text-extractor-llm.git
cd rc-data-extractor

2. Install dependencies:

npm install

3. Configure environment variables in .env (already included):

# Server Configuration
PORT=3000
NODE_ENV=development

# Ollama LLM Configuration
OLLAMA_URL=http://localhost:11434
MODEL_NAME=llama3.1:8b

# CORS Configuration
ALLOWED_ORIGINS=

Usage

Development

npm run dev

Production Build

npm run build
npm start

API Endpoints

Health Check

GET /health

Response:

{
  "status": "OK",
  "message": "RC Data Extractor API is running",
  "timestamp": "2025-08-21T10:30:00.000Z",
  "environment": "development"
}

Extract RC Data

POST /extract-rc

Request: Multipart form data with:

• front: RC front image file
• back: RC back image file

Response:

{
  "success": true,
  "data": {
    "registrationNumber": "MH01AB1234",
    "ownerName": "John Doe",
    "vehicleClass": "Motor Car",
    "fuelType": "Petrol",
    "engineNumber": "ENG123456",
    "chassisNumber": "CHS789012",
    "manufacturingDate": "2020-01-15",
    "registrationDate": "2020-02-01",
    "validUpto": "2035-01-31",
    "address": "123 Main Street, Mumbai, Maharashtra",
    "vehicleMake": "Maruti Suzuki",
    "vehicleModel": "Swift",
    "color": "White"
  },
  "extractionConfidence": 0.85,
  "processingTime": 3245
}

Project Structure

src/
├── controllers/
│   └── rcController.ts       # Main controller handling RC extraction
├── services/
│   ├── ocrService.ts        # Tesseract OCR integration
│   └── llmService.ts        # Ollama LLM integration
├── types/
│   └── index.ts             # TypeScript type definitions
├── utils/
│   └── imageProcessor.ts    # Image processing utilities
├── uploads/                 # Temporary upload directory
└── index.ts                 # Main application entry point

Configuration

Environment Variables

Variable	Description	Default
PORT	Server port	3000
NODE_ENV	Environment mode	development
OLLAMA_URL	Ollama server URL	http://localhost:11434
MODEL_NAME	LLM model name	llama3.1:8b
ALLOWED_ORIGINS	CORS allowed origins	Empty (allows all in dev)

Security Features

• CORS configuration with environment-based origins
• Security headers (X-Content-Type-Options, X-Frame-Options, X-XSS-Protection)
• File upload validation (image files only, size limits)
• Temporary file cleanup after processing
• Error message filtering in production

Performance

• Concurrent OCR processing for front and back images
• Automatic cleanup of uploaded files
• Processing time monitoring and reporting
• Confidence scoring for extraction quality

Error Handling

• Global error handler with environment-aware error messages
• Graceful shutdown on SIGTERM/SIGINT
• Comprehensive error logging
• Proper HTTP status codes