AI Quality CI 🤖

AI-powered code quality analysis tool with support for multiple LLM providers and GitHub integration.

Features ✨

• 🧠 Multiple AI Providers

  • OpenAI (GPT-4, GPT-3.5-turbo)
  • Azure OpenAI
  • Claude (claude-3-opus, claude-3-sonnet)
  • DeepSeek (deepseek-coder)

• 🔍 Code Analysis

  • Static analysis with Pylint
  • AI-powered improvement suggestions
  • Automatic fixes
  • Code formatting

• 🌟 GitHub Integration

  • Pull Request analysis
  • Automatic comments
  • Commit creation
  • Code suggestions

• 💻 User-Friendly Interface

  • Colored output
  • Git-style diffs
  • Interactive fix application
  • Multi-language support

Installation 📦

pip install ai-quality-ci

Configuration ⚙️

1. Environment Variables

Set up the required environment variables:

# OpenAI
export OPENAI_API_KEY="your-key"

# Azure OpenAI
export AZURE_OPENAI_KEY="your-key"
export AZURE_OPENAI_ENDPOINT="your-endpoint"

# GitHub
export GITHUB_TOKEN="your-token"

2. Configuration File (Optional)

Create a config.yaml file to customize settings:

provider:
  name: openai  # or azure, claude, deepseek
  model: gpt-4  # provider-specific model
  
analysis:
  ignore_patterns:
    - "test_*.py"
    - "setup.py"
  pylint_config: "path/to/pylintrc"

output:
  language: en  # or pt-BR, es
  human_readable: true

Usage 🚀

File Analysis

# Analyze a file
ai-quality-ci review-files file.py

# Analyze a directory
ai-quality-ci review-files src/

# Ignore specific files
ai-quality-ci review-files . --ignore "test_*.py"

# Human-readable output
ai-quality-ci review-files src/ --human-readable

Pull Request Analysis

# Analyze PR
ai-quality-ci review-pr username/repo 123

# Analyze PR with specific model
ai-quality-ci review-pr username/repo 123 --provider azure --model gpt-4

Options 🎛️

Option	Description	Default
--provider	AI provider	openai
--model	AI model	gpt-4
--language	Output language (en, pt-BR)	en
--auto-apply	Apply fixes automatically	false
--show-fixes	Show suggested fix details	false
--human-readable	Use rich formatting for better readability	false
--recursive	Search files recursively	true
--ignore	Patterns to ignore (can use multiple)	-
--config	Configuration file	-

Suggested Fixes 🛠️

When using --show-fixes, the tool shows corrections in git-diff format:

Fix #1: Add parameter typing
--- a/file.py
+++ b/file.py
@@ -1,5 +1,5 @@
-def process_data(data):
-    result = []
+def process_data(data: List[Dict]) -> List[Dict]:
+    result: List[Dict] = []
     for item in data:
         # processing
     return result

For each fix you can:

1. View title and description
2. View proposed changes
3. Choose to apply or not
4. Generate commit automatically

Display Modes 👀

1. Standard Mode: Plain text output
2. Human-Readable Mode (--human-readable):
   • Colors and icons
   • Rich formatting
   • Colored diffs
   • Friendly interaction

Applying Fixes ✅

1. Manual (--show-fixes):

   • Individual choice
   • Separate commits
   • Preview changes

2. Automatic (--auto-apply --show-fixes):

   • All fixes
   • Prior confirmation
   • Automatic commits

⚠️ Note: --auto-apply requires --show-fixes

Development 🔧

Environment

# Clone repository
git clone https://github.com/username/ai-quality-ci
cd ai-quality-ci

# Install in development mode
pip install -e ".[dev]"

Tests

# Run tests
pytest

# With coverage
pytest --cov=ai_quality_ci

License 📜

MIT License - See LICENSE for details.