Installation

Install OpenMOA and set up your environment for streaming machine learning.

Prerequisites

Note: OpenMOA uses the MOA (Massive Online Analysis) Java framework under the hood, bridged via JPype1. This is why Java is required.

Quick Install (User)

# Step 1: Verify Java environment
java -version

# Step 2: (Optional) Create virtual environment
conda create -n openmoa python=3.11
conda activate openmoa

# Step 3: (Optional) Install PyTorch CPU version
pip install torch torchvision --index-url https://download.pytorch.org/whl/cpu

# Step 4: Install OpenMOA
pip install openmoa

# Step 5: Verify installation
python -c "import openmoa; print(openmoa.__version__)"

During installation, two build scripts run automatically:

• Download MOA JAR (moa.jar) — The core MOA Java engine
• Generate Python stubs — For IDE autocompletion support

Developer Install

# Step 1: Install Java and PyTorch (same as above)

# Step 2: Install Pandoc (required for documentation)
# Ubuntu:
sudo apt-get install -y pandoc
# macOS:
brew install pandoc
# conda:
conda install -c conda-forge pandoc

# Step 3: Clone repository
git clone https://github.com/ZW-SIYUAN/OpenMOA.git
cd OpenMOA

# Step 4: Editable install with dev dependencies
pip install --editable ".[dev,doc]"

Available invoke commands for developers:

Docker Install

cd docker
docker compose build
docker compose up
# Access Jupyter at http://localhost:8888

The Docker image is based on jupyter/base-notebook and includes OpenJDK 8, PyTorch CPU, and OpenMOA pre-installed.

Environment Variables

Verify Installation

import openmoa
print(openmoa.__version__)      # Print version

# Quick test: load built-in dataset and read one instance
from openmoa.datasets import ElectricityTiny
stream = ElectricityTiny()
instance = stream.next_instance()
print(f"Features: {instance.x}")
print(f"Label: {instance.y_label}")

If the above code runs without errors, installation is successful.

Core Dependencies

Troubleshooting