license-analyzer 0.0.3

A robust Python module for analyzing and comparing software licenses

License Analyzer

A robust Python module for analyzing and comparing software licenses using multiple matching strategies including SHA256, fingerprinting, and semantic embeddings.

Features

• Multiple Matching Methods: SHA256 exact matching, canonical fingerprinting, and semantic embeddings
• Lazy Loading: Embedding models are only loaded when needed for performance
• Efficient Caching: Automatically manages license database with incremental updates
• Batch Processing: Analyze multiple license files in a single operation
• Flexible API: Both object-oriented and functional interfaces
• Command Line Interface: Easy-to-use CLI for quick analysis
• Comprehensive Testing: Full test suite with mocking for reliable operation

Installation

# Install from PyPI (when published)
pip install license-analyzer

# Or install from source
git clone https://github.com/yourusername/license-analyzer.git
cd license-analyzer
pip install -e .

# For development
pip install -e .[dev]

Requirements

• Python 3.8+
• numpy
• sentence-transformers (for semantic analysis)
• torch (dependency of sentence-transformers)

Quick Start

As a Library

from license_analyzer import LicenseAnalyzer

# Initialize analyzer
analyzer = LicenseAnalyzer()

# Analyze a single file
matches = analyzer.analyze_file("LICENSE.txt", top_n=5)
for match in matches:
    print(f"{match.name}: {match.score:.4f} ({match.method.value})")

# Analyze text directly
license_text = """MIT License

Copyright (c) 2024 Your Name

Permission is hereby granted..."""

matches = analyzer.analyze_text(license_text)
print(f"Best match: {matches[0].name} (score: {matches[0].score:.4f})")

# Analyze multiple files
results = analyzer.analyze_multiple_files([
    "LICENSE1.txt", 
    "LICENSE2.txt", 
    "LICENSE3.txt"
])

for file_path, matches in results.items():
    print(f"\n{file_path}:")
    for match in matches[:3]:  # Top 3 matches
        print(f"  {match.name}: {match.score:.4f}")

Command Line Interface

# Analyze a single file
license-analyzer LICENSE.txt

# Analyze multiple files
license-analyzer LICENSE1.txt LICENSE2.txt LICENSE3.txt

# Get top 10 matches in JSON format
license-analyzer --format json --top-n 10 LICENSE.txt

# Set minimum score threshold
license-analyzer --min-score 0.8 LICENSE.txt

# Use custom SPDX directory
license-analyzer --spdx-dir /custom/path/spdx LICENSE.txt

# Verbose output with database stats
license-analyzer --verbose LICENSE.txt

Architecture

Core Components

1. LicenseAnalyzer: Main interface class
2. LicenseDatabase: Manages license data with lazy loading
3. LicenseMatch: Represents analysis results
4. DatabaseEntry: Internal license database entry
5. MatchMethod: Enumeration of matching strategies

Matching Strategies

1. SHA256: Exact byte-for-byte matching (fastest, most accurate)
2. Fingerprint: Canonical token-based matching (handles formatting differences)
3. Embedding: Semantic similarity using sentence transformers (most flexible)

Performance Optimization

• Lazy Loading: Embedding models are only loaded when needed
• Incremental Updates: Database only processes changed files
• Smart Matching: Perfect matches skip expensive embedding computation
• Efficient Caching: JSON-based database with SHA