https://github.com/erishen/ai-analyze
MCP + DeepSeek powered code analysis — Serena AST parsing, AI quality assessment, Docker auto-generation, ¥0.01-0.03 per run
https://github.com/erishen/ai-analyze
ai ast code-analysis deepseek docker mcp serena static-analysis
Last synced: 15 days ago
JSON representation
MCP + DeepSeek powered code analysis — Serena AST parsing, AI quality assessment, Docker auto-generation, ¥0.01-0.03 per run
- Host: GitHub
- URL: https://github.com/erishen/ai-analyze
- Owner: erishen
- Created: 2026-01-22T14:15:09.000Z (6 months ago)
- Default Branch: main
- Last Pushed: 2026-06-16T12:08:10.000Z (21 days ago)
- Last Synced: 2026-06-16T14:10:27.703Z (21 days ago)
- Topics: ai, ast, code-analysis, deepseek, docker, mcp, serena, static-analysis
- Language: Python
- Homepage:
- Size: 602 KB
- Stars: 0
- Watchers: 0
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# AI-Powered Code Analysis Tool
An intelligent code analysis toolkit based on **MCP (Model Context Protocol)** and **DeepSeek AI**, providing Serena code structure analysis, AI-driven quality assessment, Docker auto-generation, and end-to-end solutions.
## ✨ Key Features
- 🔍 **Serena Code Structure Analysis** — Multi-language support (20+ languages), intelligent symbol parsing
- 🤖 **AI-Enhanced Analysis** — DeepSeek-powered code quality & architecture evaluation
- 🔌 **MCP Server** — Expose analysis capabilities as MCP Tools for AI agents (Claude, Cursor, Trae, etc.)
- 🐳 **Docker Auto-Generation** — AI-driven, intelligent containerization recommendations
- 💰 **Ultra-Low Cost** — ~¥0.01-0.03 per full analysis
- ⚡ **One-Click Automation** — Makefile-driven, 2-3 minutes for complete workflow
- 📊 **Multi-Format Reports** — Markdown + JSON dual output
- 🎯 **Smart Optimization** — 15+ optimization dimensions based on best practices
## 📦 Project Structure
```
.
├── src/ # Source code
│ ├── serena_client.py # Serena MCP client
│ ├── serena_stdio_client.py # Stdio communication client
│ ├── unified_analyzer.py # Unified analysis engine
│ ├── analysis_api.py # Analysis REST API
│ ├── analysis_integration.py # Analysis integration layer
│ ├── ast_analyzer.py # AST-based code analyzer
│ ├── ast_rules.py # AST analysis rules
│ ├── ast_visualizer.py # AST visualization
│ ├── benchmark.py # Performance benchmarking
│ ├── cache_warmer.py # Cache pre-warming
│ ├── config.py # Configuration management
│ ├── data_store.py # Data persistence
│ ├── dependency_graph.py # Dependency graph builder
│ ├── exceptions.py # Custom exception hierarchy
│ ├── incremental_analyzer.py # Incremental analysis
│ ├── logger.py # Structured logging
│ ├── memory.py # Memory management
│ ├── multi_level_cache.py # Multi-level cache system
│ ├── performance_analyzer.py # Performance analysis
│ ├── plugin_system.py # Plugin architecture
│ ├── progress.py # Progress tracking
│ ├── quality_score.py # Code quality scoring
│ ├── report_system.py # Report generation
│ ├── retry.py # Retry with backoff
│ ├── security_scanner.py # Security vulnerability scanner
│ ├── similarity.py # Code similarity detection
│ ├── tech_debt.py # Technical debt tracker
│ └── __init__.py
├── tools/ # Tool scripts
│ ├── full_analyzer.py # One-click analysis tool
│ ├── analyze_project_multilang.py # Multi-language analysis
│ ├── docker_generator.py # Docker generator
│ ├── smart_docker_config.py # Smart Docker configuration
│ ├── ai_enhanced_analyzer.py # AI-enhanced analyzer (with framework upgrade)
│ ├── analyze_with_ai.py # AI analysis wrapper
│ ├── ast_analyzer_tool.py # AST analyzer CLI tool
│ └── clean_generated_files.py # Cleanup tool
├── tests/ # Test suite (30+ test files)
│ ├── test_serena_client.py
│ ├── test_unified_analyzer.py
│ ├── test_ast_analyzer.py
│ ├── test_security_scanner.py
│ ├── test_quality_score.py
│ ├── test_dependency_graph.py
│ └── ...
├── examples/ # Example code
│ └── serena_example.py # Serena usage example
├── docs/ # Documentation
│ ├── FEATURES.md # Feature overview
│ ├── INDEX.md # Documentation index
│ ├── OPTIMIZATION_HISTORY.md # Optimization history
│ └── QUICK_START.md # Quick start guide
├── scripts/ # CI/CD scripts
│ └── ci-verify.sh # CI verification
├── .github/workflows/ # GitHub Actions
│ ├── ci.yml # CI pipeline
│ └── publish.yml # Publish pipeline
├── reports/ # Analysis report output
├── pyproject.toml # Project config
├── Makefile # Build commands (35+)
├── .env.example # Environment template
└── README.md # This file
```
## 🚀 Quick Start
### Prerequisites
⚠️ **Two steps required before use**:
1. **Download Serena MCP**
```bash
git clone git@github.com:oraios/serena.git
cd serena
python -m venv venv
source venv/bin/activate # Windows: venv\Scripts\activate
pip install -e .
python -m serena --version
```
2. **Configure DeepSeek API Key**
```bash
cp .env.example .env
# Edit .env:
OPENAI_API_KEY=sk-your-deepseek-api-key-here
OPENAI_BASE_URL=https://api.deepseek.com
OPENAI_MODEL=deepseek-chat
PROJECT_PATH=/path/to/your/project
```
### 1. Install Dependencies
#### Option 1: uv (recommended)
```bash
make uv-install
# Or manually:
uv sync
uv sync --extra dev
uv shell
```
#### Option 2: pip
```bash
make install
# Or:
pip install -e .
```
**uv vs pip performance**:
| Operation | pip | uv | Improvement |
|-----------|-----|-----|-------------|
| Install | ~30s | ~2s | 15x |
| Resolve | ~10s | ~0.2s | 50x |
| Venv create | ~5s | ~0.1s | 50x |
### 2. Configure Environment
```bash
cp .env.example .env
# Configure:
# - PROJECT_PATH: target project path
# - OPENAI_API_KEY: DeepSeek API key
# - OPENAI_BASE_URL: https://api.deepseek.com
# - OPENAI_MODEL: deepseek-chat
```
### 3. One-Click Full Analysis
```bash
# Full: Serena + AI + Docker
make analyze-full
# Serena only
make analyze-full-serena
# Skip AI, run Serena + Docker
make analyze-full-skip-ai
# Skip Docker, run Serena + AI
make analyze-full-skip-docker
# Force overwrite existing Docker config
make analyze-full-force
```
### 4. Individual Features
```bash
# AI enhancement only (with existing Serena report)
make analyze-ai-only REPORT=reports/your_report.json
# Docker config only
make docker-generate
# Build and run Docker
make docker-build
make docker-run
```
## 📋 Available Commands (35+)
Run `make help` to see all commands.
### 🔍 One-Click Analysis
| Command | Description | AI Cost |
|---------|-------------|---------|
| `make analyze-full` | Full: Serena + AI + Docker | ~¥0.017 |
| `make analyze-full-serena` | Serena only | ¥0 |
| `make analyze-full-skip-ai` | Serena + Docker | ¥0 |
| `make analyze-full-skip-docker` | Serena + AI | ~¥0.017 |
| `make analyze-full-force` | Force overwrite Docker config | ~¥0.017 |
### 🔍 Core Analysis
| Command | Description | AI Cost |
|---------|-------------|---------|
| `make analyze-ai` | AI-enhanced analysis (full flow) | ~¥0.017 |
| `make analyze-skip-ai` | Serena only | ¥0 |
| `make analyze-ai-only` | AI enhancement on existing report | ~¥0.017 |
| `make analyze` | Markdown report (multi-language) | ¥0 |
| `make analyze-json` | JSON report (multi-language) | ¥0 |
### 🐳 Docker Commands
| Command | Description |
|---------|-------------|
| `make docker-check` | Check Docker config |
| `make docker-generate` | Generate Docker config (AI) |
| `make docker-build` | Build Docker image |
| `make docker-run` | Run Docker container |
| `make docker-compose-up` | Start docker-compose |
| `make docker-compose-down` | Stop docker-compose |
### 🧹 Cleanup
| Command | Description |
|---------|-------------|
| `make clean` | Clean cache and temp files |
| `make clean-reports` | Clean old reports |
| `make clean-all` | Clean all generated files |
### 🛠️ Development
| Command | Description |
|---------|-------------|
| `make install` | Install dependencies |
| `make test` | Run tests |
| `make lint` | Lint code |
| `make format` | Format code |
| `make debug` | Show debug info |
## 📖 Documentation
- [Quick Start Guide](docs/QUICK_START.md)
- [Feature Overview](docs/FEATURES.md)
- [Optimization History](docs/OPTIMIZATION_HISTORY.md)
- [Documentation Index](docs/INDEX.md)
## 🔌 MCP Server
ai-analyze can run as an MCP Server, exposing analysis capabilities as tools for AI agents (Claude, Cursor, Trae, etc.).
### Available MCP Tools
| Tool | Description |
|------|-------------|
| `analyze_project` | Full project analysis (security, quality, dependency, AST) |
| `scan_security` | Security vulnerability scanning |
| `analyze_quality` | Code quality scoring (0-100, A-F grade) |
| `analyze_ast` | Single file AST analysis (complexity, code smells) |
| `detect_similarities` | Duplicate and similar code detection |
| `analyze_dependencies` | Module dependency graph analysis |
### Start MCP Server
```bash
# Via CLI entry point
ai-analyze-mcp
# Or directly
python -m src.mcp_server
```
### Configure in AI Agents
Add to your MCP client configuration (e.g., Claude Desktop, Cursor, Trae):
```json
{
"mcpServers": {
"ai-analyze": {
"command": "python",
"args": ["-m", "src.mcp_server"],
"cwd": "/path/to/ai-analyze"
}
}
}
```
Or with `uv`:
```json
{
"mcpServers": {
"ai-analyze": {
"command": "uv",
"args": ["run", "python", "-m", "src.mcp_server"],
"cwd": "/path/to/ai-analyze"
}
}
}
```
## 🎯 Feature Details
### 1. Serena Multi-Language Code Analysis
- **Supported Languages**: Python, JavaScript, TypeScript, Java, Go, Rust, C/C++, PHP, Ruby, Swift, and 20+ more
- **Analysis Scope**:
- Directory structure and file distribution
- Code symbols (classes, functions, variables)
- Language usage ratios
- Project size statistics
- **Output**: Markdown + JSON dual-format reports
### 2. AI-Enhanced Deep Analysis (DeepSeek)
- **Cost**: ~¥0.01-0.03 per analysis
- **Dimensions**:
- Project architecture evaluation
- Tech stack modernization
- Code quality score (1-10)
- Potential issue identification
- Improvement suggestions (8-15 items)
- Scalability and maintainability analysis
- **Docker strategy recommendations**
- **Framework upgrade suggestions**
- **Output**: AI-enhanced Markdown reports
### 3. AI-Driven Docker Auto-Generation
- **Smart Detection**:
- Project type (Next.js, FastAPI, Django, React, Vue, Go, Java, etc.)
- Database requirements (PostgreSQL, MySQL, MongoDB, Redis)
- Dependencies and config files
- Smart npm/pnpm choice based on lockfile
- **AI Optimization**:
- Base image selection (alpine, slim, distroless)
- Multi-stage build strategy
- Port and security optimization
- Health check configuration
- CI/CD integration
- **Generated Files**:
- `Dockerfile`, `docker-compose.yml`, `.dockerignore`
- `docker-build.sh`, `docker-run.sh`, `DOCKER_USAGE.md`
### 4. Report Management
```
reports/
├── project_analysis_20260125.json # Serena raw data
├── project_analysis_20260125.md # Serena Markdown report
└── project_analysis_20260125-ai.md # AI-enhanced report
```
## 💰 Cost Analysis
### Token Usage
| Analysis Type | Tokens | Cost |
|--------------|--------|------|
| Serena Structure | 0 | ¥0 |
| AI Code Quality | ~2,500 | ~¥0.005 |
| AI Docker Strategy | ~5,100 | ~¥0.007 |
| AI Framework Upgrade | ~2,500 | ~¥0.005 |
| **Full Analysis** | **~10,100** | **~¥0.017** |
### Cost Comparison
| Approach | Cost | Time | Quality |
|----------|------|------|---------|
| Python Script | **¥0.018/run** | 2-3 min | Professional |
| IDE + AI | Free (limited) | 15-30 min | Prompt-dependent |
| Manual | ¥50-200 | 30-60 min | Inconsistent |
## 🎬 Use Cases
### Quick Project Onboarding
```bash
export PROJECT_PATH=/path/to/new/project
make analyze-full
```
### Batch Analysis
```bash
for dir in services/*/; do
PROJECT_PATH="$dir" make analyze-full
done
```
### CI/CD Integration
```yaml
- name: Analyze code
run: make analyze-full
- name: Build and test
run: |
make docker-build
make docker-run
```
### Code Review Assistant
```bash
make analyze-full-skip-docker
# Review AI suggestions
```
### Containerize Legacy Projects
```bash
export PROJECT_PATH=/path/to/legacy/project
make analyze-full-skip-ai
make docker-build && make docker-run
```
## 🔧 Requirements
- **Python**: 3.9+
- **Serena MCP**: Latest
- **AI API**: DeepSeek API Key (or OpenAI-compatible)
- **OS**: macOS, Linux, Windows (WSL)
## 🤝 Contributing
Issues and pull requests are welcome!
## 📄 License
MIT
---
**💡 Tip**: Detailed documentation is available in the `docs/` directory.
**💰 Cost Reminder**: AI analysis requires an API key — each run costs only ~¥0.01-0.03.