ANOMX PACKAGE

From data
to understanding.

Python building blocks for time series, forecasting, and anomaly detection. In your notebook, your pipeline, or your own system.

v0.2.92Python >=3.11Apache 2.0124 modules

Start with a few lines.

Install Anomx in a virtual environment. The core install includes native time series, models, scorers, and detectors.

python3 -m venv .venv
source .venv/bin/activate
python -m pip install --upgrade anomx

Windows: activate the environment with .venv\Scripts\Activate.ps1 in PowerShell.

Your first forecast.

Create a time series, split it chronologically, and evaluate a seasonal baseline on held-out data.

Python

import numpy as np
from anomx import TimeSeries
from anomx.models import NaiveSeasonal
from anomx.metrics import mae

# A regular, deterministic signal with a 12-step cycle.
series = TimeSeries.from_values(
    np.sin(np.arange(120) * 2 * np.pi / 12)
)
train, test = series.split_before(0.8)

model = NaiveSeasonal(K=12).fit(train)
forecast = model.predict(len(test))
print(mae(test, forecast))

TimeSeries ↗NaiveSeasonal ↗mae ↗

Explore the API.

From the data container to the final decision, the reference follows the package’s modules.

Data & time series

Load, structure, and prepare observations for analysis.

TimeSeries · Dataset · Transforms 02

Models

Learn normal behavior, forecast values, and discover structure.

Forecasting · Reconstruction · Representation 03

Anomaly detection

Measure deviations and turn continuous scores into decisions.

Scorers · Detectors · Pipelines 04

Metrics

Evaluate forecasts and results with reproducible metrics.

MAE · RMSE · Classification 05

Execution

Compose local workflows and track their results.

Jobs · Runs · WorkContext 06

Integrations

Connect optional model libraries and external data sources.

Darts · Platform · Workspaces

Add what you need.

Install optional dependencies as extras. Darts, Torch, and ONNX are separate from the core install.

python -m pip install "anomx[visualization]"
python -m pip install "anomx[torch]"
python -m pip install "anomx[darts]"

anomx[visualization]anomx[calendar]anomx[interop]anomx[pyod]anomx[storage]anomx[darts]anomx[torch]anomx[onnx]anomx[ml]anomx[platform]

The API reference is generated from the Python source. Signatures, types, inheritance, and docstrings reflect the documented package version. Reference text retains the language of the source.