Skip to content

Installation Guide

ONAD supports Python 3.10+ and can be installed via pip. This guide covers various installation options and requirements.

Prerequisites

  • Python: 3.10 or higher
  • Operating System: Windows, macOS, Linux
  • Memory: At least 512 MB RAM (more for larger datasets)

Basic Installation

The simplest way to install ONAD is using pip:

pip install onad

This installs the core package with essential dependencies:

  • numpy>=2.2.3 - Numerical computing
  • faiss-cpu>=1.10.0 - Similarity search engine
  • tqdm>=4.67.1 - Progress bars
  • scipy>=1.15.2 - Scientific computing

Installation Options

ONAD provides several optional dependency groups for different use cases:

Evaluation Tools

For model evaluation and benchmarking:

pip install onad[eval]

Includes: - scikit-learn - Machine learning utilities and metrics

Deep Learning Support

For neural network-based models:

pip install onad[dl]

Includes: - torch - PyTorch deep learning framework

Benchmarking Tools

For comparing with other streaming libraries:

pip install onad[benchmark]

Includes: - river - Online machine learning library

Development Tools

For contributing to ONAD:

pip install onad[dev]

Includes: - ruff - Fast Python linter and formatter - pre-commit - Git hooks for code quality - pytest - Testing framework - pytest-cov - Coverage reporting - mypy - Static type checking

Documentation Tools

For building documentation:

pip install onad[docs]

Includes: - mkdocs - Documentation generator - mkdocs-material - Material theme for MkDocs - mkdocstrings - API documentation from docstrings

Complete Installation

To install all optional dependencies:

pip install onad[all]

Virtual Environment Setup

We recommend using a virtual environment to avoid dependency conflicts:

conda create -n onad python=3.10
conda activate onad
pip install onad[all]

python -m venv onad-env
source onad-env/bin/activate  # On Windows: onad-env\Scripts\activate
pip install onad[all]

virtualenv onad-env
source onad-env/bin/activate  # On Windows: onad-env\Scripts\activate
pip install onad[all]

Development Installation

If you want to contribute to ONAD or use the latest development version:

git clone https://github.com/OliverHennhoefer/onad.git
cd onad
pip install -e .[dev]

The -e flag installs in "editable" mode, so changes to the code take effect immediately.

Verify Installation

Test your installation:

import onad

print(f"ONAD version: {onad.__version__}")

# Test a simple model
from onad.model.iforest import OnlineIsolationForest

model = OnlineIsolationForest()
print("✅ Installation successful!")

Common Issues

ImportError: No module named 'faiss'

Problem: FAISS is not installed or incompatible with your system.

Solution: 

# Try installing FAISS separately
pip install faiss-cpu
# Or for GPU support (if available)
pip install faiss-gpu

Memory Issues with Large Datasets

Problem: Out of memory errors when processing large streams.

Solution: Configure memory limits in your models: 

model = OnlineIsolationForest(window_size=1000)  # Limit window size

Type Checking Errors

Problem: MyPy reports type errors in your code.

Solution: Ensure you're using Python 3.10+ and install type stubs: 

pip install types-tqdm

System Requirements

Minimum Requirements

  • CPU: Any modern processor
  • RAM: 512 MB
  • Disk: 100 MB free space
  • Python: 3.10+
  • CPU: Multi-core processor for parallel processing
  • RAM: 2+ GB for large-scale streaming
  • Disk: 1+ GB for datasets and models
  • Python: 3.11+ for better performance

Next Steps

After installation:

  1. Quick Start: Get started with your first anomaly detection model
  2. User Guide: Learn about ONAD's features in detail
  3. Examples: Explore real-world use cases

Need Help?

If you encounter installation issues, please:

  1. Check the troubleshooting section
  2. Search existing issues
  3. Create a new issue with your system details