Kabrino - CML Manager for VS Code

Kabrino is a Visual Studio Code extension that integrates with Cloudera Machine Learning (CML) to manage projects, jobs, experiments, and AutoML workflows directly from your editor.

Features

• 🔌 Connect to CML: Configure your CML instance with base URL and API key
• 📁 Browse Projects: View all accessible CML projects in a tree view
• 🎯 Select Active Project: Choose which project to work with
• 📂 File Management: Browse, upload, download, and manage project files
• 🚀 Run Python Files: Execute Python files directly in CML jobs
• 📊 Monitor Jobs: View jobs and their execution history
• 🧪 Experiments: Create, view, and manage ML experiments with runs tracking
• 🤖 AutoML Projects: Create complete AutoML projects with pre-configured workflows
• ⚡ Real-time Status: Track job execution with live status updates
• 📝 DSN Management: Switch between different CML instances seamlessly

Getting Started

Prerequisites

1. Access to a Cloudera Machine Learning instance
2. A valid CML API key

Installation

1. Install the extension from VS Code Marketplace (or build from source)
2. Install required dependencies:

cd kabrino
npm install

Configuration

1. Click on the Kabrino icon in the Activity Bar (🔥)
2. Click the gear icon (⚙️) to configure connection
3. Enter your CML credentials:
   • Base URL: Your CML instance URL (e.g., https://ml-xxx.cloudera.site)
   • API Key: Your CML API key

How to Get Your API Key

1. Log in to your CML instance
2. Click on your user profile (top right corner)
3. Navigate to "Settings" or "API Keys"
4. Generate a new API key
5. Copy and paste it into the Kabrino configuration

Usage

Selecting a Project

1. Open the "CML Projects" view in the Kabrino sidebar
2. Click on any project to select it as the active project
3. All other views (Files, Jobs, Experiments) will update to show content for the selected project

Managing Files

The "CML Files" view shows all files in your selected project:

• Browse Files: Navigate through project directories
• Upload Files: Right-click on a folder → "Upload File to CML"
• Download Files: Right-click on a file → "Download File from CML"
• Delete Files: Right-click on a file → "Delete File from CML"
• Refresh: Click the refresh icon to reload the file tree

Running Python Files

There are two ways to run a Python file in CML:

Method 1: Context Menu

1. Right-click on any .py file in the editor
2. Select "Run Python File in CML"

Method 2: Editor Toolbar

1. Open any .py file
2. Click the play icon (▶️) in the editor toolbar

The extension will:

1. Upload the file to your CML project
2. Create or find a suitable job
3. Execute the job
4. Monitor the execution
5. Notify you when complete

Managing Jobs

• View all jobs in the "CML Jobs" panel
• Expand a job to see its execution history (up to 10 recent runs)
• See real-time status with emoji indicators:
  • ✅ Success
  • ❌ Failed
  • 🔄 Running
  • ⚪ Pending/Unknown
• Right-click on a job to:
  • Stop a running job
  • View job details

Working with Experiments

The "CML Experiments" view lets you manage ML experiments:

• View Experiments: See all experiments in the selected project
• View Runs: Expand an experiment to see its runs
• Run Status: Track experiment runs with status indicators
• Create Experiments: Use the API to programmatically create experiments

Creating AutoML Projects

Quickly set up a complete AutoML workflow:

1. Right-click on "CML Projects" view
2. Select "Create AutoML Project"
3. Kabrino will:
   • Create a new CML project
   • Install required AutoML packages
   • Upload helper scripts (feature selection, hyperparameter tuning, model selection)
   • Create the Driver AutoML job
   • Upload AutoML configuration YAML
4. Your AutoML project is ready to use!

Managing Multiple CML Instances

Switch between different CML environments:

1. Click the gear icon (⚙️) to configure connection
2. Enter a DSN name for the connection
3. Configure base URL and API key
4. Use the dropdown in the Projects view to switch between saved DSNs

Commands

All commands are available through the Command Palette (Ctrl+Shift+P / Cmd+Shift+P):

Configuration

• Kabrino: Configure CML Connection - Set up your CML credentials

Projects

• Kabrino: Refresh Projects - Reload the projects list
• Kabrino: Select Project - Choose a project to work with
• Kabrino: Create AutoML Project - Set up a complete AutoML workflow

Files

• Kabrino: Refresh Files - Reload the file tree
• Kabrino: Upload File to CML - Upload a file to the selected project
• Kabrino: Download File from CML - Download a file from CML
• Kabrino: Delete File from CML - Remove a file from the project

Jobs

• Kabrino: Refresh Jobs - Reload the jobs list
• Kabrino: Run Active Python File in CML - Execute the currently open Python file
• Kabrino: Stop Job - Terminate a running job

Experiments

• Kabrino: Refresh Experiments - Reload the experiments list
• Kabrino: View Experiment Runs - Show all runs for an experiment

Extension Settings

This extension contributes the following settings:

• kabrino.baseUrl: CML instance base URL
• kabrino.apiKey: CML API key (stored securely)
• kabrino.dsn: Data Source Name for the current connection
• kabrino.connections: Saved CML connections (multiple instances support)

Known Issues

• Large Python files may take time to upload
• Job execution depends on CML cluster resources
• Experiment runs with large datasets may take time to load (timeout set to 10 minutes)
• File operations on very large files may be slow

Development

Building from Source

git clone <repository>
cd kabrino
npm install

Running in Development

1. Open the project in VS Code
2. Press F5 to start debugging
3. A new Extension Development Host window will open

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

License

See LICENSE file for details.

Enjoy using Kabrino! 🔥

For support or questions, please visit the GitHub repository.