Document Intelligence System

AI-powered document extraction system for Indian documents with support for multiple document types including Aadhaar, PAN, Voter ID, Driving License, Passport, and Invoices.

Features

• Multi-Document Support: Aadhaar, PAN, Voter ID, Driving License, Passport, Invoice, eKYC
• Hybrid Extraction: Gemini AI + OCR fallback for maximum accuracy
• Async Processing: Concurrent document processing with worker pool
• Smart Caching: File hash-based caching to avoid reprocessing
• Type-Safe Schemas: Pydantic models with validation and review flags
• REST API: FastAPI-based API with async task management
• Web Interface: Simple frontend for document upload and results

Architecture

┌─────────────┐
│   Frontend  │
└──────┬──────┘
       │
┌──────▼──────────────────────────────────────┐
│              FastAPI Server                  │
│  ┌────────────┐  ┌──────────────┐           │
│  │   Upload   │  │  Task Queue  │           │
│  │   Router   │  │              │           │
│  └─────┬──────┘  └──────┬───────┘           │
│        │                │                    │
│  ┌─────▼────────────────▼─────┐             │
│  │     Worker Pool              │            │
│  │  (Concurrent Processing)     │            │
│  └─────┬────────────────────────┘            │
│        │                                     │
│  ┌─────▼──────────┐  ┌──────────────┐       │
│  │  LLM Service   │  │ OCR Service  │       │
│  │  (Gemini/Groq) │  │ (Tesseract)  │       │
│  └────────────────┘  └──────────────┘       │
│                                              │
│  ┌──────────────┐  ┌──────────────┐         │
│  │Cache Manager │  │ Normalizer   │         │
│  └──────────────┘  └──────────────┘         │
└──────────────────────────────────────────────┘

Installation

Prerequisites

• Python 3.11+
• Tesseract OCR
• Google Gemini API key or Groq API key

Setup

1. Clone the repository

git clone <repository-url>
cd project-text-extraction

2. Install dependencies

pip install -r requirements.txt

3. Install Tesseract OCR

• Windows: Download from https://github.com/UB-Mannheim/tesseract/wiki
• Linux: sudo apt-get install tesseract-ocr
• Mac: brew install tesseract

4. Configure environment variables

cp .env.example .env
# Edit .env with your API keys

Required environment variables:

GEMINI_API_KEY=your_gemini_api_key
GROQ_API_KEY=your_groq_api_key (optional)
AI_PROVIDER=gemini
OCR_ENGINE=tesseract

Usage

Start the Server

python -m uvicorn main:app --reload --host 0.0.0.0 --port 8000

Access the Application

• Web Interface: http://localhost:8000
• API Documentation: http://localhost:8000/docs
• Health Check: http://localhost:8000/health

API Endpoints

Upload Document

POST /api/upload
Content-Type: multipart/form-data

Response:
{
  "task_id": "abc123",
  "status": "pending",
  "message": "Task created successfully"
}

Check Task Status

GET /api/tasks/status/{task_id}

Response:
{
  "task_id": "abc123",
  "filename": "document.jpg",
  "status": "completed",
  "progress": 100.0
}

Get Task Result

GET /api/tasks/result/{task_id}

Response:
{
  "task_id": "abc123",
  "filename": "document.jpg",
  "status": "completed",
  "extracted_data": {
    "document_type": "aadhaar",
    "name": "John Doe",
    "aadhaar_number_masked": "XXXX XXXX 1234",
    ...
  }
}

Document Types

Supported Documents

1. Aadhaar Card

   • Name, DOB, Gender, Address
   • Masked Aadhaar number
   • Front/Back side detection

2. PAN Card

   • Name, Father's name, DOB
   • PAN number

3. Voter ID

   • Name, Father's name, DOB, Gender
   • Voter ID number, Address
   • Polling station details

4. Driving License

   • Name, DOB, Address
   • License number, Issue/Expiry dates
   • Vehicle classes

5. Passport

   • Name, DOB, Gender, Nationality
   • Passport number, Issue/Expiry dates
   • MRZ parsing

6. Invoice

   • Vendor/Buyer details
   • Invoice number, date
   • Line items, amounts

Configuration

Environment Variables

Variable	Default	Description
AI_PROVIDER	gemini	AI provider (gemini/groq)
GEMINI_API_KEY	-	Google Gemini API key
GROQ_API_KEY	-	Groq API key (optional)
OCR_ENGINE	tesseract	OCR engine
OCR_LANG	eng+hin	OCR languages
UPLOAD_DIR	uploads	Upload directory
OUTPUT_DIR	outputs	Output directory
CACHE_DIR	cache	Cache directory

Worker Pool

Configure concurrent processing:

• Default: 8 workers
• Adjust based on CPU cores and memory

Cache

• File hash-based caching
• Automatic cache invalidation
• Persistent storage in JSON format

Development

Project Structure

project/
├── main.py                 # FastAPI application
├── requirements.txt        # Python dependencies
├── .env                    # Environment variables
├── models/                 # Pydantic schemas
│   ├── schemas.py          # Document schemas
│   └── task_schemas.py     # Task schemas
├── routes/                 # API routes
│   ├── upload.py           # Upload endpoint
│   ├── tasks.py            # Task management
│   └── batch.py            # Batch processing
├── services/               # Business logic
│   ├── document_service.py # Main processing
│   ├── llm_service.py      # AI extraction
│   ├── ocr_service.py      # OCR processing
│   ├── cache_manager.py    # Caching
│   ├── task_queue.py       # Task management
│   └── worker_pool.py      # Concurrent processing
├── frontend/               # Web interface
│   └── index.html          # UI
├── tests/                  # Test suite
└── logs/                   # Application logs

Running Tests

python -m pytest tests/

Performance

• Processing Time: 5-10 seconds per document
• Cache Hit Rate: >50% for repeated documents
• Concurrent Processing: Up to 8 documents simultaneously
• Accuracy: >90% for clear images

Security

• Automatic Aadhaar number masking
• No storage of sensitive data in logs
• File hash-based deduplication
• Input validation and sanitization

License

MIT License

Support

For issues and questions, please open an issue on the repository.