pmrf.extract_features

pmrf.extract_features(source, frequency, features, sparam_kind='all', dtype=<class 'jax.numpy.complex128'>)[source]

Extract frequency-dependent features from a model or network into a unified matrix.

Features are combined column-by-column, resulting in a matrix where rows correspond to frequency points and columns to the requested features. This supports extracting scalars (e.g., ‘s11_mag’), complex parameters, or custom functions from both simulation models and measured networks.

Parameters:

  • source (Model, skrf.Network, or NetworkCollection) – The source object. If a list of networks is provided, it is treated as a single stacked network with isolated ports. Missing networks in a collection should be specified using integers representing the port count at that index.

  • frequency (pmrf.Frequency or skrf.Frequency, optional) –

    The frequency points for extraction.

    • Required if source is a Model.

    • Optional if source is a Network; defaults to the Network’s native frequency points. If provided, the Network is interpolated to these points.

  • features (str, list, dict, or tuple) – The features to extract. See Notes for formatting syntax.

  • sparam_kind ({'transmission', 'reflection', 'all'} or None, optional) – Filters port expansion when generic features are requested (e.g., ‘s_re’ without specific indices).

  • dtype (data-type, optional) – The desired data-type for the output feature matrix.

Returns:

feature_matrix – The extracted feature matrix of shape (M, N), where M is the number of frequency points and N is the total number of features.

Return type:

ndarray

Notes

Feature Syntax

Features can be specified using string aliases, explicit tuples, or dictionaries for sub-components:

  • String Aliases: Convenient shorthands for standard parameters.

    • Example: 's11_db', 'a21_deg', 's11_mag'.

    • Custom: Any model attribute dependent on frequency can be accessed by name (e.g., 'my_custom_gain').

  • Explicit Tuples: Defined as (path, parameter, index).

    • Example: ('', 's_db', (0, 0)) is equivalent to 's11_db'.

    • The empty string '' denotes the base model.

  • Dictionaries: Used to target submodels (via getattr) or specific networks.

    • Example: {'lna': 's21_db'} extracts S21 from the ‘lna’ submodel.

    • Nested: {'front_end.lna': 's11'} accesses deep submodels.

    • Measurement: Maps to the label of a specific network in a collection.

  • Lists: A list containing any combination of the above will be flattened into columns.