Metadata-Version: 2.4
Name: ssad
Version: 0.1.2
Summary: Framework for self-supervised training of reconstruction-based autoencoder models for anomaly detection.
Author-email: Samuel Berlemont <samuel.berlemont@orange.com>
Maintainer: Mohammed Achraf El Khamlichi
Maintainer-email: Julien Cumin <julien1.cumin@orange.com>
License: MIT
Project-URL: Homepage, https://github.com/Orange-OpenSource/SSAD
Project-URL: Issues, https://github.com/Orange-OpenSource/SSAD/issues
Project-URL: Repository, https://github.com/Orange-OpenSource/SSAD
Classifier: Programming Language :: Python :: 3
Classifier: Operating System :: OS Independent
Requires-Python: <3.15,>=3.10
Description-Content-Type: text/markdown
License-File: LICENSE
Requires-Dist: category_encoders<3.0,>=2.8
Requires-Dist: numpy<3.0,>=1.26
Requires-Dist: pandas<3.0,>=2.2
Requires-Dist: scipy<2.0,>=1.15
Requires-Dist: scikit-learn<2.0,>=1.9
Requires-Dist: torch<3.0,>=2.10
Requires-Dist: lightning<3.0,>=2.5
Requires-Dist: mlflow<4.0,>=3.12
Requires-Dist: psutil<8.0,>=7.0
Requires-Dist: nvidia-ml-py<14.0,>=13.610
Provides-Extra: dev
Requires-Dist: pandas-stubs==2.2.2.240603; extra == "dev"
Requires-Dist: torchvision<0.29,>=0.25; extra == "dev"
Requires-Dist: types-pytz>=2026.2; extra == "dev"
Requires-Dist: mypy<2.0,>=1.11; extra == "dev"
Requires-Dist: pylint<4.0,>=3.2; extra == "dev"
Dynamic: license-file

# SSAD — Self-Supervised Anomaly Detection Library

A Python library for autoencoder-based **anomaly detection** with self-supervised training and dynamic per-sample **confidence** updates.

## Key Features

- Compute per-sample anomaly scores
- Estimate confidence from score distributions
- Recalibrate confidence intervals during training
- Apply confidence-aware losses (normal / abnormal / uncertain)
- Track experiments and artifacts with **MLflow**

## Installation

```bash
pip install ssad
```

For development setup:

```bash
pip install -e .[dev]
```

## Quick Links

- **Repository**: https://github.com/Orange-OpenSource/SSAD
- **Examples**: https://github.com/Orange-OpenSource/SSAD/tree/main/examples
- **Issues**: https://github.com/Orange-OpenSource/SSAD/issues


## References

1. N. Najari, S. Berlemont, G. Lefebvre, S. Duffner, C. Garcia,  
   *Robust Variational Autoencoders and Normalizing Flows for Unsupervised Network Anomaly Detection*,  
   AINA 2022, doi: 10.1007/978-3-030-99587-4_24

2. N. Najari, S. Berlemont, G. Lefebvre, S. Duffner, C. Garcia,  
   *RADON: Robust Autoencoder for Unsupervised Anomaly Detection*,  
   SIN 2021, doi: 10.1109/SIN54109.2021.9699174
