🚀 Unified MCP Code Review Assistant

A comprehensive, intelligent code generation and review system with dual modes: Simple for quick tasks and Advanced for complex projects with iterative improvement.

✨ Key Features

🎯 Dual Mode Architecture

• Simple Mode: Fast, straightforward code generation and review
• Advanced Mode: LangChain + LangGraph with iterative improvement
• Seamless Switching: Change modes without restarting

🔧 Simple Mode Features

• ✅ Quick code generation with Gemini 2.5
• ✅ Comprehensive code reviews
• ✅ Session-based history
• ✅ Fast response times
• ✅ Perfect for quick tasks and learning

⚡ Advanced Mode Features

• ✅ Iterative code improvement (up to 3 cycles)
• ✅ Quality metrics with numerical scores (1-10)
• ✅ LangGraph workflow orchestration
• ✅ State management and memory checkpointing
• ✅ Conditional routing based on code quality
• ✅ Structured reviews with priority issues
• ✅ Perfect for complex projects

📁 Unified Project Structure

mcp_code_review_assistant/
├── unified_main.py              # 🎯 Single entry point
├── config.py                    # ⚙️ Unified configuration
├── utils.py                     # 🛠️ Common utilities
├── unified_requirements.txt     # 📦 All dependencies
├── test_unified.py             # 🧪 Comprehensive tests
├── UNIFIED_README.md           # 📖 This documentation
│
├── # Core Components (Simple Mode)
├── orchestrator.py             # Simple workflow orchestrator
├── schemas.py                  # Basic MCP message schemas
├── agents/
│   ├── coder_agent.py         # Basic code generation
│   └── reviewer_agent.py      # Basic code review
│
├── # Enhanced Components (Advanced Mode)
├── langgraph_orchestrator.py   # LangGraph workflow
├── enhanced_schemas.py         # Advanced schemas with metrics
├── langchain_agents/
│   ├── enhanced_coder_agent.py # LangChain-powered coder
│   └── enhanced_reviewer_agent.py # Structured reviewer
│
└── # Configuration & Environment
    ├── .env                    # API key configuration
    └── .gitignore             # Git ignore rules

🚀 Quick Start

1. Installation

# Clone or download the project
cd mcp_code_review_assistant

# Install core dependencies (required)
pip install google-generativeai pydantic python-dotenv

# Install advanced dependencies (optional, for advanced mode)
pip install langchain langgraph langchain-google-genai langchain-core

2. Configuration

Create or edit .env file:

GEMINI_API_KEY=your_actual_api_key_here

Get your API key from: Google AI Studio

3. Run the Application

python unified_main.py

🎮 Usage Guide

Mode Selection

When you start the application, you'll see:

🎯 Choose Your Experience Mode:
1. 🚀 Simple Mode - Quick code generation and review
2. ⚡ Advanced Mode - LangChain + LangGraph with iterative improvement
3. ℹ️  Show feature comparison
4. 🚪 Exit

Simple Mode Commands

📝 Simple Mode Commands:
• Enter any code request
• 'history' - View workflow history
• 'clear' - Clear message history
• 'mode' - Switch to advanced mode
• 'quit' - Exit application

Advanced Mode Commands

📝 Advanced Mode Commands:
• Enter any code request for iterative improvement
• 'workflow' - Show LangGraph workflow visualization
• 'history' - Show current thread history
• 'clear' - Clear thread memory
• 'mode' - Switch to simple mode
• 'quit' - Exit application

📊 Feature Comparison

Feature	Simple Mode	Advanced Mode
Code Generation	✅ Single pass	✅ Iterative (up to 3x)
Code Review	✅ Text-based	✅ Structured metrics
Quality Scores	❌ No	✅ 1-10 scale
Auto Improvement	❌ No	✅ Yes
Workflow Control	✅ Linear	✅ Conditional routing
State Management	✅ Basic	✅ Advanced (LangGraph)
Memory	✅ Session	✅ Persistent threads
Error Handling	✅ Basic	✅ Robust with fallbacks
Speed	✅ Fast	⚡ Moderate (thorough)
Best For	📝 Quick tasks	🎯 Complex projects

🧪 Testing

Run comprehensive tests:

python test_unified.py

Expected output:

✅ Core Imports: PASSED
✅ Simple Mode Imports: PASSED
✅ Advanced Mode Imports: PASSED
✅ Configuration: PASSED
✅ Utilities: PASSED
✅ Simple Mode Functionality: PASSED
✅ Advanced Mode Functionality: PASSED
✅ Unified Main Module: PASSED

Tests passed: 8/8
🎉 All tests passed!

🔧 Configuration Options

Edit config.py to customize:

# API Configuration
GEMINI_MODEL = "gemini-2.5-flash-lite-preview-06-17"

# Simple Mode Settings
SIMPLE_MODE = {
    'temperature': 0.3,
    'max_tokens': 2000,
    'timeout': 30
}

# Advanced Mode Settings
ADVANCED_MODE = {
    'temperature': 0.3,
    'max_tokens': 2000,
    'timeout': 30,
    'max_iterations': 3,
    'quality_threshold': 8.0
}

🎯 Example Usage

Simple Mode Example:

Your request: Create a function to calculate factorial

🔧 Generated Code:
def factorial(n):
    if n == 0 or n == 1:
        return 1
    return n * factorial(n - 1)

🔍 Code Review:
The function correctly implements factorial calculation using recursion...

Advanced Mode Example:

Your request: Create a function to calculate factorial

🔧 Generating code (iteration 1)...
🔍 Reviewing code (iteration 1)...
🔄 Code needs improvement (iteration 1)
⚡ Improving code (iteration 1)...
🔍 Reviewing code (iteration 2)...
🎯 Code quality is satisfactory, finalizing...

📊 Code Quality Metrics:
   📖 Readability: 9.0/10
   🔧 Maintainability: 8.5/10
   ⚡ Performance: 8.0/10
   🔒 Security: 9.0/10
   🎯 Overall Score: 8.6/10

📈 Workflow Summary:
   🔄 Total iterations: 2
   💬 Messages processed: 6
   📍 Final step: completed

🛠️ Troubleshooting

Common Issues:

1. "GEMINI_API_KEY not set"

   • Solution: Add your API key to the .env file

2. "Advanced mode import error"

   • Solution: Install advanced dependencies:

     pip install langchain langgraph langchain-google-genai

3. "Configuration issue"

   • Solution: Check your .env file and API key validity

Getting Help:

1. Run tests: python test_unified.py
2. Check configuration: Verify .env file
3. Check dependencies: Ensure all packages are installed

🚀 Advanced Features

LangGraph Workflow Visualization:

[START] → Generate Code → Review Code → Decision Point
                                            ↓
                                      Needs Improvement?
                                            ↓
                              YES → Improve Code ──┐
                               ↑                   │
                               └───────────────────┘
                                            ↓
                              NO → Finalize → [END]

Quality Metrics:

• Readability: Code clarity and documentation
• Maintainability: Ease of modification and extension
• Performance: Efficiency and optimization
• Security: Vulnerability assessment
• Overall: Comprehensive quality score

State Management:

• Persistent conversation threads
• Memory checkpointing
• Session management
• History tracking

📝 Dependencies

Core (Required):

• google-generativeai - Gemini API integration
• pydantic - Data validation and schemas
• python-dotenv - Environment variable management

Advanced (Optional):

• langchain - LLM application framework
• langchain-google-genai - Gemini integration for LangChain
• langgraph - State management and workflow orchestration
• langchain-core - Core LangChain functionality

🎉 Benefits of Unified Architecture

1. Single Entry Point: One command to rule them all
2. Mode Flexibility: Choose the right tool for the task
3. Seamless Switching: Change modes without restart
4. Consistent Interface: Same commands across modes
5. Progressive Enhancement: Start simple, upgrade when needed
6. Comprehensive Testing: Full test coverage
7. Clean Configuration: Centralized settings management

🔮 Future Enhancements

• Web interface for better user experience
• Support for multiple programming languages
• Integration with popular IDEs
• Team collaboration features
• Custom workflow templates
• Performance analytics and insights

---

Ready to start coding smarter? 🚀

python unified_main.py