Source code for

AND Aggregator

Aggregator that identifies a time step as anomalous if all the components
are flagged as anomalous (logical AND).

from typing import Sequence

from darts import TimeSeries
from import NonFittableAggregator

[docs]class AndAggregator(NonFittableAggregator): def __init__(self) -> None: super().__init__() def __str__(self): return "AndAggregator" def _predict_core(self, series: Sequence[TimeSeries]) -> Sequence[TimeSeries]: return [ s.sum(axis=1).map(lambda x: (x >= s.width).astype(s.dtype)) for s in series ]