CloudRobo Client

Local data ingestion is available through cloudrobo ingestion start/status/stop. See ingestion setup, configuration and recovery. The full bundle requires Python >=3.10; individual legacy packages retain their own requirements.

CloudRobo command-line tool and Python SDK for Huawei Cloud's Embodied Intelligence Platform (CloudRobo) - resource management and task orchestration.

中文文档

Features

• Asset Management: Manage lifecycle of models, datasets, and algorithms
• Data Processing: Create and monitor data processing tasks (cleaning, format conversion, etc.)
• Model Training: Submit and monitor training tasks (pre-training, fine-tuning, simulation reinforcement learning)
• Inference Service: Deploy and manage model inference services
• Robot Management: Register robots and export device certificates
• Task Dispatch: Execute embodied intelligence tasks on robots
• Workspace Management: Manage resource-isolated workspaces
• Resource Management: Query resource quotas and resource pools
• Data Plane SDK: Robot-side Zenoh data plane client (optional)

Installation

For Users (Recommended)

Install from PyPI:

pip install hw-cloudrobo-client

After installation, the cloudrobo CLI tool is ready to use.

For Developers

Install from source (requires git):

git clone <repository-url>
cd cloudrobo-client
pip install -r requirements-dev-editable.txt

This installs all sub-packages and the repository-only ingestion mock tool in editable mode. Code changes take effect immediately.

Quick Start

1. Configure Authentication

# Configure Huawei Cloud AK/SK (automatically encrypted)
cloudrobo config set ak your-access-key sk your-secret-key

# Configure region
cloudrobo config set region cn-southwest-2

Or use environment variables:

export HUAWEI_CLOUD_AK="your-access-key"
export HUAWEI_CLOUD_SK="your-secret-key"
export CLOUDROBO_REGION="cn-southwest-2"

2. View Available Commands

cloudrobo --help

3. Example: List Workspaces

cloudrobo workspace list

Documentation

Each functional module's documentation is maintained by its respective package:

Project Structure

This project uses a Monorepo architecture where each functional module is an independent package:

cloudrobo-client/
├── packages/                    # Independent functional packages
│   ├── cloudrobo-core/          # Core SDK + CLI framework
│   ├── cloudrobo-asset/         # Asset management
│   ├── cloudrobo-dataset/       # Dataset processing
│   ├── cloudrobo-train/         # Model training
│   ├── cloudrobo-infer/         # Inference service
│   ├── cloudrobo-robot/         # Robot management
│   ├── cloudrobo-dispatch/      # Task dispatch
│   ├── cloudrobo-workspace/     # Workspace management
│   ├── cloudrobo-resource/      # Resource management
│   └── cloudrobo-r2c/           # Data Plane SDK
├── tools/ingestion-mock/         # Developer-only foreground tests (not published)
├── docs/                        # Project-level documentation
└── pyproject.toml               # Aggregate installation configuration

Testing

# Run all tests
python -m pytest tests/ packages/ tools/ingestion-mock/tests/ -v

# Run tests for a single package
python -m pytest packages/cloudrobo-asset/tests/ -v

License

Apache License 2.0