🚀 Karate Feature Graph Analyzer (MCP-Powered)

A powerful Model Context Protocol (MCP) tool designed for AI agents (Claude, Antigravity, Cursor) to analyze Karate Framework feature files, manage multi-project dependencies, and generate interactive graphs.

✨ Why use this with AI?

Instead of running scripts manually, you can just chat with your codebase:

• 🔍 Impact Analysis: Ask AI "What happens if I change this API?" and see the exact lines affected.
• 📈 Ecosystem View: Merge multiple projects into one global graph via chat.
• 📝 Source Snippets: View relevant code snippets around dependencies directly in AI search results.
• 🎫 Ticket tracking: Automatically link Jira tags to code components.
• 🩺 Health Check: Get AI-generated reports on cycles and redundant components.

🚀 Quick Start

1. Installation

Option A: Standard (Requires Internet)

git clone <repository-url>
cd karate-feature-graph-analyzer
pip install -e .

Option B: Offline / Vendor Mode (No Internet Required) If your environment blocks external downloads, use the pre-packaged libraries in python_libs:

git clone <repository-url>
cd karate-feature-graph-analyzer
# Install all dependencies from local vendor folder
pip install --no-index --find-links=python_libs -r requirements.txt
# Install the project in editable mode
pip install -e . --no-deps

2. Configure MCP Client

Add the following to your AI client configuration (e.g., claude_desktop_config.json):

{
  "mcpServers": {
    "karate-analyzer": {
      "command": "python",
      "args": ["C:/path/to/repo/src/karate_graph_analyzer/mcp_server.py"],
      "env": {
        "PYTHONPATH": "C:/path/to/repo/src",
        "PYTHONIOENCODING": "utf-8"
      }
    }
  }
}

[!IMPORTANT] Replace C:/path/to/repo with the actual absolute path of this repository.

3. Start Analyzing via Chat

Once connected, just ask your AI:

• "Register the project at E:/my-karate-fw and analyze it."
• "Show me all test cases affected by a change in the OrderManagement page."
• "Scan all my registered projects and give me a health summary."

🤖 Available AI Tools

📊 Visualizing Results

The analyzer generates interactive HTML graphs in the output/ directory using a standardized Visual Identity System:

• Shapes & Meanings:

  • ⭐ Test Case (Star): Main entry points / End-to-End scenarios.
  • 🟦 Workflow/Common (Square): Reusable logic blocks and service files.
  • ⬢ API Group (Hexagon): Domain or path-based API groupings.
  • ⬥ API Endpoint (Diamond): Individual API calls.
  • ⬭ Page (Ellipse): UI Page Objects.
  • 📐 Action (Triangle): Specific UI interactions/locators.
  • 🛢️ Database (Cylinder): Database feature files.
  • ⚪ Query (Dot): Specific DB query scenarios.
  • 📦 Data (Box): External JSON/CSV data files.

• Interactive Features:

  • Focus Mode: Double-click any node to isolate its dependency chain.
  • AI Insights: Click a node to see "Expert Analysis" and "Fix Suggestions" generated by AI.
  • Status Coloring: Nodes turn 🔴 Red (Failed), 🟢 Green (Passed), or 🟠 Orange (Modified in Diff) based on context.

🛠️ Advanced Usage (CLI)

If you still prefer the terminal, you can use the built-in scanners:

Scan a single project

python scan_project.py <path_to_project> <output_name>

Scan multiple projects at once

python scan_all_projects.py <path1> <path2> <parent_folder>

📁 Project Structure

karate-feature-graph-analyzer/
├── src/karate_graph_analyzer/
│   ├── mcp_server.py                # MCP Server Entry Point
│   ├── mcp_interface/               # MCP Tool Logic
│   ├── parser/                      # Karate/Gherkin Parser
│   ├── graph/                       # Graph Construction
│   └── visualization/               # HTML/Vis.js Generator
├── python_libs/                     # [NEW] Offline python dependencies (.whl)
├── requirements.txt                 # Dependency list for installation
├── output/                          # Generated HTML/JSON files
└── tests/                           # 300+ Unit & Integration tests

📝 License

This project is licensed under the MIT License.

Built with ❤️ for the Karate Community.