PlotPeakByLogValue v1

../_images/PlotPeakByLogValue-v1_dlg.png

PlotPeakByLogValue dialog.

Summary

Fits a number of spectra with the same function.

Properties

Name Direction Type Default Description
Input Input string Mandatory A list of sources of data to fit. Sources can be either workspace names or file names followed optionally by a list of spectra/workspace-indices or values using the notation described in the description section of the help page.
Spectrum Input number 1 Set a spectrum to fit. However, if spectra lists (or workspace-indices/values lists) are specified in the Input parameter string these take precedence.
WorkspaceIndex Input number 0 Set a workspace-index to fit (alternative option to Spectrum). However, if spectra lists (or workspace-indices/values lists) are specified in the Input parameter string, or the Spectrum parameter integer, these take precedence.
OutputWorkspace Output TableWorkspace Mandatory The output TableWorkspace
Function Input string Mandatory The fitting function, common for all workspaces in the input WorkspaceGroup
LogValue Input string   Name of the log value to plot the parameters against. Default: use spectra numbers.
StartX Input number Optional A value of x in, or on the low x boundary of, the first bin to include in the fit (default lowest value of x)
EndX Input number Optional A value in, or on the high x boundary of, the last bin the fitting range (default the highest value of x)
FitType Input string Sequential Defines the way of setting initial values. If set to ‘Sequential’ every next fit starts with parameters returned by the previous fit. If set to ‘Individual’ each fit starts with the same initial values defined in the Function property. Allowed values: [‘Sequential’, ‘Individual’]
PassWSIndexToFunction Input boolean False For each spectrum in Input pass its workspace index to all functions thathave attribute WorkspaceIndex.
Minimizer Input string Levenberg-Marquardt Minimizer to use for fitting. Minimizers available are ‘Levenberg-Marquardt’, ‘Simplex’, ‘FABADA’, ‘Conjugate gradient (Fletcher-Reeves imp.)’, ‘Conjugate gradient (Polak-Ribiere imp.)’ and ‘BFGS’
CostFunction InOut string Least squares Cost functions to use for fitting. Cost functions available are ‘Least squares’ and ‘Ignore positive peaks’. Allowed values: [‘Least squares’, ‘Poisson’, ‘Rwp’, ‘Unweighted least squares’]
MaxIterations Input number 500 Stop after this number of iterations if a good fit is not found
PeakRadius Input number 0 A value of the peak radius the peak functions should use. A peak radius defines an interval on the x axis around the centre of the peak where its values are calculated. Values outside the interval are not calculated and assumed zeros.Numerically the radius is a whole number of peak widths (FWHM) that fit into the interval on each side from the centre. The default value of 0 means the whole x axis.
CreateOutput Input boolean False Set to true to create output workspaces with the results of the fit(default is false).
OutputCompositeMembers Input boolean False If true and CreateOutput is true then the value of each member of a Composite Function is also output.
ConvolveMembers Input boolean False If true and OutputCompositeMembers is true members of any Convolution are output convolved with corresponding resolution
EvaluationType Input string CentrePoint The way the function is evaluated: CentrePoint or Histogram. Allowed values: [‘CentrePoint’, ‘Histogram’]
Exclude Input dbl list   A list of pairs of real numbers, defining the regions to exclude from the fit.
IgnoreInvalidData Input boolean False Flag to ignore infinities, NaNs and data with zero errors.

Description

This algorithm fits a series of spectra with the same function. Each spectrum is fit independently and the result is a table of fitting parameters unique for each spectrum. The sources for the spectra are defined in the Input property. The Input property expects a list of spectra identifiers separated by semicolons (;). An identifier is itself a comma-separated list of values. The first value is the name of the source. It can be either a workspace name or a name of a file (RAW or Nexus). If it is a name of a WorkspaceGroup all its members will be included in the fit. The second value selects a spectrum within the workspace or file. It is an integer number with a prefix defining the meaning of the number: “sp” for a spectrum number, “i” for a workspace index, or “v” for a range of values on the numeric axis associated with the workspace index. For example, sp12, i125, v0.5:2.3. If the data source is a file only the spectrum number option is accepted. The third value of the spectrum identifier is optional period number. It is used if the input file contains multiperiod data. In case of workspaces this third parameter is ignored. This are examples of Input property

 "test1,i2; MUSR00015189.nxs,sp3; MUSR00015190.nxs,sp3; MUSR00015191.nxs,sp3"
 "test2,v1.1:3.2"
 "test3,v" - fit all spectra in workspace test3

Internally PlotPeakByLogValue uses Fit v1 algorithm to perform fitting and the following properties have the same meaning as in Fit v1: Function, StartX, EndX, Minimizer, CostFunction. Property FitType defines the way of setting initial values. If it is set to “Sequential” every next fit starts with parameters returned by the previous fit. If set to “Individual” each fit starts with the same initial values defined in the Function property.

LogValue property specifies a log value to be included into the output. If this property is empty the values of axis 1 will be used instead. Setting this property to “SourceName” makes the first column of the output table contain the names of the data sources (files or workspaces).

Output workspace format

PlotPeakByLogValue_Output.png

PlotPeakByLogValue_Output.png

In this example a group of three Matrix workspaces were fitted with a Gaussian on a linear background.

The output workspace is a table in which rows correspond to the spectra in the order they (spectra) appear in the Input property. The first column of the table has the log values. It is followed by pairs of columns with parameter values and fitting errors. If a parameter was fixed or tied the error will be zero. Here is an example of the output workspace:

Minimizer setup

It is possible to supply a fully configured minimizer via the Minimizer property, this can also include several flags that are used to indicate the workspace name and workspace index that has been fitted (useful for fitting using a minimizer that output workspaces (i.e. FABADA)).

The possible flags are:

$wsname
The name of the workspace being fitted
$wsindex
The index of the spectrum being fitted
$basename
A convenience flag for $wsname_$wsindex (the same format the fit result and parameter output takes)
$outputname
The output workspace name (i.e. the value of the OutputWorkspace property).

Usage

Example - fitting a single spectrum of in a workspace:

ws = CreateSampleWorkspace()
function = "name=Gaussian,Height=10.0041,PeakCentre=10098.6,Sigma=48.8581;name=FlatBackground,A0=0.3"
peaks = PlotPeakByLogValue(ws, function, Spectrum=1)

Example - sequentially fitting a workspace:

import numpy as np

ws = CreateSampleWorkspace()
function = "name=Gaussian,Height=10.0041,PeakCentre=10098.6,Sigma=48.8581;name=FlatBackground,A0=0.3"

#create string of workspaces to fit (ws,i0; ws,i1, ws,i2 ...)
workspaces = [ws.name() + ',i%d' % i for i in range(ws.getNumberHistograms())]
workspaces = ';'.join(workspaces)

peaks = PlotPeakByLogValue(workspaces, function, Spectrum=1)

#get peak centres for comparison
peak_centres = peaks.column('f0.PeakCentre')
ref = np.empty(len(peak_centres))
ref.fill(10098.6)

print(np.allclose(ref, peak_centres, 1e-3))

Output:

True

Categories: AlgorithmIndex | Optimization

Source

C++ source: PlotPeakByLogValue.cpp (last modified: 2019-07-17)

C++ header: PlotPeakByLogValue.h (last modified: 2018-11-21)