captest.filters.Outliers

class captest.filters.Outliers(*, envelope_kwargs, custom_name, name)

Remove statistical outliers in the (poa, power) plane via sklearn EllipticEnvelope.

Reads capdata.floc[["poa", "power"]] (the regression-mapped columns), drops any NaN rows by delegating to capdata.filter_missing (which records its own step), fits sklearn.covariance.EllipticEnvelope on the cleaned 2-D matrix, and keeps the rows whose predict is 1.

envelope_kwargs carries user overrides for EllipticEnvelope; defaults (support_fraction=0.9, contamination=0.04) are merged in at run time. The merged dict is exposed on envelope_kwargs_resolved for display.

__init__(**params)

Methods

__init__(**params)
from_config(config)	Build an instance from a to_config() dict.
run(capdata)	Execute the step, record runtime state, and append self to filters.
to_config()	Serialize this step to a config dict (every param, defaults included; the param-system name is omitted).

Attributes

args_repr	Render EllipticEnvelope(k=v, ...) using the resolved kwargs.
custom_name
envelope_kwargs
explanation	Human-readable description of the step's effect (read after run()).
name
param