pyeeg.features.FeaturePipeline

class pyeeg.features.FeaturePipeline(config: PipelineConfig)

Pipeline for extracting and aligning multiple features from stimuli.

Runs the extractors declared in the configuration’s feature_specs over an input text, then either returns the word-level feature arrays directly (by word position) or aligns them to a neural signal using the configured AlignmentHandler and an optional Praat TextGrid.

Parameters:

config (PipelineConfig) – Pipeline configuration listing the feature specs, alignment options, and normalization method.

config

Pipeline configuration.

Type:

PipelineConfig

_extractors

Mapping of extractor name (from FeatureSpec.name) to the instantiated extractor object.

Type:

dict of str -> object

_alignment_handler

Alignment handler constructed from config.alignment_config, or None if no alignment was configured.

Type:

AlignmentHandler or None

Methods

FeaturePipeline.extract(text[, textgrid, ...])	Extract features from text and optionally align to signal.
FeaturePipeline.extract_audio(signal, srate)	Extract acoustic features from an audio signal.
FeaturePipeline.normalize_features(features)	Normalize features.