mix_draws

params.mix_draws(*sources, n=None, seed=None)

Combine draw matrices from different sources into one matrix.

Each source is resampled to n rows by drawing whole rows, so the joint structure within a source (for example, a calibrated posterior’s correlation) is preserved while sources stay independent of each other.

Parameters

Name	Type	Description	Default
sources	pd.DataFrame	Two or more draw matrices with disjoint column names. Valid inputs are ParameterSet.sample output, heormodel.calibrate.CalibrationResult.posterior, or any external draw matrix indexed by iteration.	()
n	int | None	Rows in the mixed matrix. Defaults to the shortest source’s length. Sources shorter than n are resampled with replacement; sources at least as long are subsampled without replacement.	None
seed	int | np.random.Generator | None	Integer seed or numpy Generator for the resampling.	None

Returns

Name	Type	Description
	pd.DataFrame	DataFrame with a fresh RangeIndex named "iteration" and every
	pd.DataFrame	column from every source.

Raises

Name	Type	Description
	ValueError	If no sources are given, a source is empty, n is not positive, or column names collide across sources.

Example

import pandas as pd from heormodel.params import mix_draws lit = pd.DataFrame({“u”: [0.6, 0.7, 0.8]}) post = pd.DataFrame({“beta”: [0.1, 0.2]}) mixed = mix_draws(lit, post, n=4, seed=0) list(mixed.columns) [‘u’, ‘beta’] mixed.index.name ‘iteration’ len(mixed) 4