sample_parameter_space API Reference

geoloop.utils.RunManager.sample_parameter_space

sample_parameter_space(
    config: SingleRunConfig,
) -> tuple[dict, list[str]]

Sample the model parameter space based on definitions provided in config.

Parameters:

Name	Type	Description	Default
config	SingleRunConfig	Object containing parameter definitions. Must include: n_samples : int Number of Monte Carlo samples to generate. Remaining keys should define either: * A fixed numerical value, or * A distribution definition list: [dist_type, p1, p2, (optional p3)] where dist_type ∈ {"normal", "uniform", "lognormal", "triangular"}.	required

Returns:

Name	Type	Description
sample_space	dict of array-like	Dictionary mapping parameter names → sampled values with shape (n_samples,).
varying_parameters	list of str	List of parameter names that were sampled from statistical distributions.

Notes

Locked parameters (non-varying) are broadcast to all samples.