codemetry/laravel

Laravel adapter for Codemetry analysis pipeline

v1.4.0 2026-02-08 20:56 UTC

Requires

• php: ^8.2
• codemetry/core: v1.4.0
• illuminate/console: ^11.0|^12.0
• illuminate/support: ^11.0|^12.0

Requires (Dev)

• orchestra/testbench: ^9.0|^10.0
• pestphp/pest: ^3.0

MIT

README

Laravel adapter for Codemetry - provides Artisan commands and service provider integration for Git repository analysis.

This package depends on codemetry/core.

Documentation | Getting Started

Requirements

• PHP 8.2+
• Laravel 11.x or 12.x
• Git

Installation

composer require codemetry/laravel

This automatically installs codemetry/core as a dependency.

Publish the configuration file:

php artisan vendor:publish --tag=codemetry-config

Upgrading

Both codemetry/laravel and codemetry/core must be updated together to ensure compatibility:

composer update "codemetry/*"

Or explicitly:

composer update codemetry/laravel codemetry/core

Note: Running composer update codemetry/laravel alone may not work because both packages are versioned together and require matching versions.

Usage

# Table output (one row per day)
php artisan codemetry:analyze --days=7

# JSON output
php artisan codemetry:analyze --days=7 --format=json

# Filter by author or branch
php artisan codemetry:analyze --days=7 --author="Jane Doe" --branch=main

# Specify date range
php artisan codemetry:analyze --since=2024-01-01 --until=2024-01-31

# Enable AI-powered explanations (requires API keys in config)
php artisan codemetry:analyze --days=7 --ai=1

Configuration

After publishing, edit config/codemetry.php:

return [
    // Number of days to analyze by default
    'days' => 7,

    // Baseline period for normalization (in days)
    'baseline_days' => 56,

    // Follow-up fix detection horizon (in days)
    'follow_up_horizon_days' => 3,

    // AI configuration (optional)
    'ai' => [
        'enabled' => false,
        'engine' => 'openai', // openai, anthropic, deepseek, google
        'api_key' => env('CODEMETRY_AI_API_KEY'),
    ],
];

AI Integration

AI integration is opt-in and metrics-only. Engines never receive raw code or diffs.

Supported engines:

• OpenAI
• Anthropic (Claude)
• DeepSeek
• Google

When AI is requested but unavailable (missing keys, API failure), analysis continues normally with a warning.

Privacy

• The AI engine receives aggregated metrics only - never raw source code, full diffs, or file contents.
• All analysis runs locally via Git commands against your repository.
• No data is sent to external services unless AI engines are explicitly enabled.

License

MIT