Retiring anomaly_detection_toolkit (asset_health)

The Anomaly Detection Toolkit repository (folder asset_health, import name anomaly_detection_toolkit) has been fully superseded by anomsmith. After upgrading, you can archive or delete that repository.

Install and imports

pip install anomsmith[deep,wavelet,plots,stats]

# Old
# import anomaly_detection_toolkit as adt

# New (examples)
import anomsmith
from anomsmith import IsolationForestDetector, VotingEnsemble, FeatureExtractor

Module map

API differences

• Labels: anomsmith uses ``1`` = anomaly, ``0`` = normal in LabelView. The old toolkit used ``-1`` / ``1`` in raw numpy arrays.

• Fit signature: primitives use fit(y, X=None) with time-series-first semantics, not only fit(X).

• Scores / predictions: use .score(y) → ScoreView and .predict(y) → LabelView instead of score_samples / predict returning bare arrays.

prepare_pm_features renamed the boolean ``add_degradation_rates`` to ``include_degradation_rates`` to avoid shadowing the helper function of the same name.