The `KDSource` Component

Distributional source, based on a MCPL-format particle list, on which the Kernel Density Estimation (KDE) method is applied.

Identification

Site:
Author: O.I.Abbate, N.S.Schmidt, updates for McStas by T.Kittelmann and P.Willendrup
Origin: Instituto Balseiro
Date: Dec 2022 / May 2026

Description

Distributional source, based on a MCPL-format particle list, on which
the Kernel Density Estimation (KDE) method is applied.

It allows sampling more particles than the number present in a virtual,
previously generated virtual source, without repeating samples, controlled via the nloop
input parameter.

To function, this component requires a KDSource installation v.2.0.2 or later, as
distributed on conda-forge and pypi and included with (conda-based) McStas 3.7.0 or later.

As inputs the component needs access to all of:

The original MPCL file
Outputs of a KDSource-analyzed / optimized KDE source:

An XML parameter-file containing the needed configuration
A "bandwidth" file (source_bws)


For information on performing KDSource analysis, please refer to the KDSource example
notebooks and the KDSource online documentation (links below).

Input parameters

Parameters in boldface are required; the others are optional.

Name	Unit	Description	Default
filename	str	Name of the XML parameters file containing KDSource definition.	0
Emin	meV	Lower energy bound. Particles found in the MCPL-file below the limit are skipped.	0
Emax	meV	Upper energy bound. Particles found in the MCPL-file above the limit are skipped.	FLT_MAX
nloop	int	Number of times to loop through the file.	1

AT (	,	,		) RELATIVE
ROTATED (	,	,		) RELATIVE

The KDSource Component

Identification

Description

Input parameters

Links

The `KDSource` Component