DeepReview - AI-Powered Code Review Tool

DeepReview is an intelligent code review assistant that uses local LLMs (Large Language Models) to analyze git branch differences and provide comprehensive code reviews. It helps developers identify potential issues, maintain code quality, and ensure best practices across their codebase.

🚀 Features

• Automated code review using local LLMs through Ollama
• Git branch difference analysis
• Detailed feedback on code quality, security, and performance
• Customizable review focus areas
• Support for multiple programming languages
• Markdown and YAML report generation

🛠️ Installation

1. Install the package:

pip install .

2. Install Ollama from ollama.ai

3. Pull the required model:

ollama pull codellama

📖 Usage

Run from your git repository:

deepreview

Command Line Parameters

DeepReview supports several command line parameters to customize its behavior:

Parameter	Description	Example
--qa	Enables interactive Q&A mode with the LLM about the analysis results	deepreview --qa
--output	Uses an existing analysis file for Q&A instead of generating a new one	deepreview --qa --output ./analysis_results/review_2024_11_23.md

Examples:

Generate a new analysis and start Q&A session:

deepreview --qa

Use existing analysis file for Q&A:

deepreview --qa --output ./analysis_results/previous_review.md

⚙️ Configuration Guide

Create a config.yaml file in:

• Your project directory: ./config.yaml

Configuration Parameters Explained

Configuration Reference

LLM Configuration

Parameter	Description	Default
llm_model	The Ollama model to use	codellama
main_branch	Your main branch name	main

File Exclusions

Pattern	Description	Default
ignore_files	List of files to ignore	[".lock", "package-lock.json", "yarn.lock", ".gitignore", "*.md", "*.log", "node_modules/**", "venv/**"]

Analysis Configuration

Parameter	Description	Default
max_diff_size	Maximum diff size to analyze (bytes)	100000
min_changes	Minimum changes needed for analysis	1

Review Priority Weights

Weight	Value	Description
security	0.3	Security considerations
performance	0.25	Performance implications
code_quality	0.25	Code quality and style
architecture	0.2	Architectural considerations

Language-Specific Settings

Python

Setting	Description	Default
check_typing	Check type hints	true
pep8_compliance	Check PEP 8 style guide	true

JavaScript

Setting	Description	Default
check_eslint	Check ESLint rules	true
modern_features	Check modern JS features	true

TypeScript

Setting	Description	Default
strict_mode	Check TypeScript strict mode	true
interface_check	Check interface implementations	true

Output Configuration

Parameter	Description	Default
format	Output format (markdown or yaml)	markdown
include_stats	Include change statistics	true
include_suggestions	Include improvement suggestions	true

📊 Output

DeepReview generates analysis reports covering:

1. Impact Analysis

   • Component/functionality effects
   • Breaking change detection

2. Code Quality Review

   • Style and best practices
   • Potential bugs
   • Design patterns

3. Security Assessment

   • Vulnerability detection
   • Auth/authorization review
   • Data handling analysis

4. Performance Review

   • Performance implications
   • Resource usage
   • Scalability concerns

5. Testing Recommendations

   • Required test types
   • Edge cases
   • Integration scenarios

Reports are saved in ./analysis_results by default.

📋 Requirements

• Python 3.7+
• Git
• Ollama
• Local LLM (default: codellama)

🤝 Contributing

Contributions are welcome! Please feel free to submit issues and pull requests.

📄 License

This project is licensed under the MIT License. This README provides: A clear introduction to the project Detailed installation steps Usage instructions A comprehensive configuration guide with explanations for all parameters Description of the analysis output Project requirements Information about contributing The configuration section is particularly detailed, explaining each parameter in the config.yaml file to help users customize the tool for their needs.