MLSysOps CLI (mls)

The official command-line interface for the MLSysOps Framework. This tool empowers you to manage applications, infrastructure resources, and orchestration agents seamlessly across the device-edge-cloud continuum.

📋 Table of Contents

• ✨ Features
• ✅ Prerequisites
• 📦 Installation
• 🔧 Configuration
• 🚀 Quick Start
• 📚 Command Reference
  • Application Commands
  • Infrastructure Commands
  • Management Commands
  • Framework Commands
• 💡 Pro Tip: Enable Tab Completion
• 📄 License

✨ Features

• 📱 Application Management: Deploy, list, and remove applications with simple commands.
• 🏗️ Infrastructure Insight: Query and register infrastructure components like clusters and nodes.
• ⚙️ System Control: Ping system agents and switch operational modes Heuristic or ML.
• 🤖 Agent Deployment: Easily deploy orchestration agents to nodes, clusters, or an entire continuum.
• 🌐 Framework installation: Install the required components of the Framework.

✅ Prerequisites

Before you begin, ensure you have the following installed:

• Python 3.7+
• pip (Python package installer)

📦 Installation

You can install the MLSysOps CLI using one of the following methods.

Option 1: Install from PyPI (Recommended)

pip install mlsysops-cli

After installation, you can use the mls command directly in your terminal.

Option 2: Install from source (For Development)

If you plan to contribute to the CLI or need the latest unreleased features, install it from the source repository.

# 1. Clone the repository (CLI branch)
git clone -b CLI [https://github.com/mlsysops-eu/mlsysops-framework.git](https://github.com/mlsysops-eu/mlsysops-framework.git)

# 2. Navigate to the CLI directory
cd mlsysops-framework/mlsysops-cli

# 3. Install in editable mode
pip install -e .

This will also expose the mls command in your terminal, and any changes you make to the code will be reflected immediately.

🔧 Configuration

Make sure you have your environment variables or .env file set up with:

# Configuration with the framework API
export MLS_API_IP=<MLS API host ip>
export MLS_API_PORT=8000

# Deployment
export KARMADA_HOST_KUBECONFIG=<path to karmada host kubeconfig>
export KARMADA_API_KUBECONFIG=<path to karmada api kubeconfig>
export KARMADA_HOST_IP=<karmada host ip>

🚀 Quick Start

Get an overview of all available commands and options with the --help flag.

mls --help

Each major component of the framework has its own command group:

• mls apps – Manage application deployments
• mls infra – Query and register infrastructure
• mls manage – System control (ping, mode switch)
• mls framework – Deploy the core framework components.

📚 Command Reference

🧹 Application Commands

Manage the lifecycle of your applications

• Deploy an application

mls apps deploy-app --path ./my_app.yaml

• List all running applications

mls apps list-all

• Remove an application

mls apps remove-app <application-id>

🏗️ Infrastructure Commands

Query information about your registered infrastructure.

• List Infrastructure by type (e.g., Cluster, Node)

mls infra list-infra --type Cluster

⚙️ Management Commands

Perform system-level administrative tasks.

-Ping a system agent to check its status:

mls manage ping-agent

-Set the system's operational mode: 0 for Heuristic 1 for ML

mls manage set-mode --mode 1

🌐 Framework Commands

Deploy core components of the MLSysOps framework and orchestration agents.

mls framework deploy-all
mls framework deploy-cluster
mls framework deploy-continuum
mls framework deploy-node
mls framework deploy-services

Optional path argument: Use the --path flag to specify the system descriptions folder.

mls framework deploy-all/cluster/continuum/node --path ./descriptions

The descriptions folder should contain subfolders like node, cluster, or continuum for proper agent configuration.

Optional inventory argument: Use the --inventory flag to specify the inventory YAML file used during the K3s installation.

mls framework add-system-agents --inventory ./inventory.yaml

Note: Only one of --path or --inventory can be specified at a time. If both options are provided, the command will throw an error.

💡 Pro Tip: Enable Tab Completion

Enable tab-completion for the mls CLI in your terminal to quickly discover available commands and options: This will help you auto-complete commands and options by simply pressing the [TAB] key.

echo 'eval "$(_MLS_COMPLETE=bash_source mls)"' >> ~/.bashrc
source ~/.bashrc

Now you can type mls [TAB][TAB] to see all available commands. It's a game-changer! 🎉

📄 License

This project is licensed by MLSysOps.

Copyright © 2025 MLSysOps.