Nutest Cortex - Automated Test Bug Fixing System

Automated system that monitors Jira tickets with "nutest_cortex" label, uses AI (Cursor Premium) to analyze test failures, fixes issues in the nutest-py3-tests repository, and automatically creates Gerrit code reviews.

Features

• 🤖 AI-Powered Fix Generation: Uses Cursor Premium to automatically generate test fixes
• 📊 Learning System: Learns from code review feedback and human fixes to improve over time
• 🔄 Iterative Fixes: Responds to CR comments with improved fixes
• 📝 Comprehensive Commit Messages: Auto-generates detailed commit messages with fix summaries
• 🔍 CR Monitoring: Continuously monitors Gerrit CRs for comments, votes, and updates
• 🎯 Branch Mapping: Automatically maps Jira fixversions to Git branches

Architecture

The system consists of:

1. Jira Webhook Listener - Monitors for tickets with "nutest_cortex" label
2. Ticket Parser - Extracts ticket details, logs from description and attachments
3. AI Fix Engine - Uses Cursor Premium to analyze failures and generate fixes
4. Git Automation - Clones repo, applies fixes, commits, and pushes to Gerrit
5. Branch Mapper - Maps Jira fixversion to Git branch using JSON config
6. Status Reporter - Updates Jira ticket with progress and CR link
7. Gerrit CR Monitor - Monitors CR comments and review votes (+1, +2, -1)
8. Feedback Learning System - Learns from comments, reviews, and human fixes
9. Iterative Fix Engine - Responds to CR comments with improved fixes

Prerequisites

• Python 3.9+
• Git
• Cursor CLI installed (for AI fix generation)
• Docker and Docker Compose (for Jira container, optional)
• Access to:
  • Jira instance (or use Docker container)
  • Gerrit server
  • nutest-py3-tests repository

Installation

1. Clone Repository

git clone <repository_url>
cd nutest_cortex

2. Create Virtual Environment

python3 -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

3. Install Dependencies

pip install -r requirements.txt

4. Configure Environment

Copy .env.example to .env and fill in your credentials:

cp .env.example .env

Edit .env with your values:

# Optional - will auto-detect if not set
CURSOR_CLI_PATH=/usr/local/bin/cursor
CURSOR_AGENT_PATH=/usr/local/bin/cursor-agent

# Required
GERRIT_USERNAME=your_username
GERRIT_PASSWORD=your_password
JIRA_API_TOKEN=your_jira_token
CURSOR_API_KEY=your_cursor_api_key

# Optional - defaults to /tmp/nutest_cortex
WORKSPACE_DIR=/tmp/nutest_cortex
LEARNING_DB_PATH=data/learning_db.sqlite

5. Configure Branch Mapping

Edit config/branch_mapping.json to map your fixversions to branches:

{
  "1.0.0": "master",
  "1.1.0": "release/1.1",
  "2.0.0": "release/2.0"
}

6. Configure Application Settings

Edit config/config.yaml with your Jira and Gerrit URLs.

Usage

Start the System

python main.py

The system will:

1. Validate startup requirements (Cursor CLI, Git, etc.)
2. Start the webhook server on port 8000
3. Begin monitoring for Jira tickets with "nutest_cortex" label
4. Start monitoring active Gerrit CRs

Using with Docker Jira (Optional)

If using the Docker Compose setup for Jira:

cd docker
docker-compose up -d

This will start:

• PostgreSQL database for Jira
• Jira Software instance on http://localhost:8080

Webhook Configuration

Configure Jira webhook to point to:

http://your-vm-ip:8000/webhook/jira

Webhook should trigger on:

• Issue created
• Issue updated

How It Works

1. Ticket Detection: System receives webhook when ticket with "nutest_cortex" label is created/updated
2. Ticket Parsing: Extracts test logs, error messages, and test file paths
3. AI Fix Generation: Uses Cursor Premium to analyze failure and generate fix
4. Fix Application: Clones repo, applies fix, generates commit message
5. CR Creation: Pushes fix to Gerrit and creates code review
6. Monitoring: Continuously monitors CR for comments and votes
7. Learning: Stores feedback and learns patterns for future fixes
8. Iterative Improvement: Responds to comments with improved fixes

Learning System

The system learns from:

• CR Comments: Extracts actionable feedback
• Review Votes: Tracks +1, +2, -1 patterns
• Human Patches: Compares AI fixes vs human fixes when someone else fixes the CR
• Jira Comments: Extracts feedback from ticket comments
• CR Outcomes: Tracks which fixes get merged vs abandoned

Learned patterns are stored in SQLite database and used to improve future fixes.

Configuration

Branch Mapping

Edit config/branch_mapping.json to map Jira fixversions to Git branches.

Monitoring Settings

In config/config.yaml:

monitoring:
  poll_interval: 30  # seconds between polling Gerrit
  enable_learning: true
  max_iterative_fixes: 3  # max patch sets per CR

Learning Settings

learning:
  db_path: "data/learning_db.sqlite"
  enable_human_patch_learning: true
  enable_comment_learning: true
  enable_jira_comment_learning: true

Project Structure

nutest_cortex/
├── src/
│   ├── __init__.py
│   ├── jira_webhook.py      # Webhook listener
│   ├── ticket_parser.py     # Parse ticket details
│   ├── ai_fixer.py          # AI fix generation
│   ├── git_automation.py    # Git operations
│   ├── branch_mapper.py     # Branch mapping
│   ├── status_reporter.py   # Jira status updates
│   ├── gerrit_monitor.py    # CR monitoring
│   ├── feedback_learner.py  # Learning system
│   ├── fix_summarizer.py    # Commit message generation
│   └── logging_config.py    # Logging setup
├── data/
│   └── learning_db.sqlite    # Learning database
├── config/
│   ├── branch_mapping.json  # Fixversion mappings
│   └── config.yaml          # Configuration
├── docker/
│   └── docker-compose.yml   # Jira setup
├── requirements.txt
├── main.py                  # Entry point
└── README.md

Troubleshooting

Cursor CLI Not Found

The system auto-detects Cursor CLI. If not found:

1. Install Cursor CLI on your VM
2. Set CURSOR_CLI_PATH and CURSOR_AGENT_PATH in .env

Git Authentication Issues

Ensure Gerrit credentials are correct in .env:

• GERRIT_USERNAME
• GERRIT_PASSWORD

Webhook Not Receiving Events

1. Check Jira webhook configuration
2. Verify webhook URL is accessible from Jira
3. Check firewall settings
4. Review logs for errors

Development

Running Tests

# Add test files and run
pytest tests/

Logging

Logs are written to:

• Console (stdout)
• logs/nutest_cortex.log (rotating, 10MB max, 5 backups)

Log level can be configured in main.py or via environment variable.

License

[Your License Here]

Contributing

[Contributing Guidelines]