CrimeLinkAnalyzer ML Services

This repository contains Python-based machine learning microservices for the CrimeLinkAnalyzer application.

Services

1. Call Analysis Service (Port 5001)

Analyzes call records from PDF files to identify patterns, build network graphs, and match against criminal database.

Features:

• PDF parsing with multiple format support
• Network graph generation using NetworkX
• Call pattern analysis (frequency, time patterns, common contacts)
• Criminal database matching
• Risk score calculation

Endpoints:

• GET /health - Health check
• POST /analyze - Upload PDF and analyze call records
• GET /results/<analysis_id> - Get analysis results
• GET /results - List all results

2. Facial Recognition Service (Port 5002)

Coming soon - Will integrate with custom Asian face recognition model

Installation

Prerequisites

• Python 3.9+
• pip
• PostgreSQL database access

Setup

1. Install dependencies:

cd call_analysis_service
pip install -r requirements.txt

2. Configure environment:

cp .env.example .env
# Edit .env with your database credentials

3. Run the service:

python app.py

The service will start on http://localhost:5001

Testing

Test Call Analysis

curl -X POST http://localhost:5001/analyze \
  -F "file=@sample_call_records.pdf"

Check Health

curl http://localhost:5001/health

PDF Format Support

The parser supports multiple call record formats:

Format 1:

2024-01-15 10:30:45 | +94771234567 | Outgoing | 00:05:23

Format 2:

15/01/2024 10:30 +94771234567 OUT 5m 23s

Format 3:

15-01-2024 10:30, +94771234567, Outgoing, 00:05:23

Integration with Spring Boot

The Spring Boot backend calls these services via REST API:

RestTemplate restTemplate = new RestTemplate();
String pythonServiceUrl = "http://localhost:5001/analyze";
ResponseEntity<Map> response = restTemplate.postForEntity(
    pythonServiceUrl, 
    fileData, 
    Map.class
);

Database Schema

The service interacts with these tables:

criminals

• id (VARCHAR PRIMARY KEY)
• name (VARCHAR)
• nic (VARCHAR)
• contact_number (VARCHAR)
• secondary_contact (VARCHAR)

call_analysis_results

• id (VARCHAR PRIMARY KEY)
• file_name (VARCHAR)
• analysis_data (JSONB)
• risk_score (INTEGER)
• created_at (TIMESTAMP)

Development

Project Structure

call_analysis_service/
├── app.py                  # Flask application
├── requirements.txt        # Python dependencies
├── .env.example           # Environment configuration
├── uploads/               # Uploaded PDF files
├── results/               # Analysis results
└── utils/
    ├── pdf_parser.py      # PDF parsing logic
    ├── network_analyzer.py # Network graph analysis
    └── database.py        # Database operations

Deployment

Using Docker (Recommended)

docker build -t crimelink-call-analysis .
docker run -p 5001:5001 --env-file .env crimelink-call-analysis

Production Considerations

• Use Redis for result caching instead of in-memory storage
• Implement rate limiting
• Add authentication/API keys
• Use Celery for async processing of large files
• Set up monitoring and logging
• Use environment-specific configurations

License

Proprietary - Crime Link Analyzer System