grapa.parse.curve_subclasses_utils.FitHandler

class grapa.parse.curve_subclasses_utils.FitHandler(filename)

Bases: object

This class should work as a minimum version, although it has no fit function implemented - see below shild class FitHandlerBasicFunc to see how to do that.

The configuration file is a grapa-style text file, with following attributes (lines of parameters). See example grapa/datatypes/XRF_fitparameters.txt. Each column is a fit “preset”, intended to be ready to be used. Required attributes:

  • label: used for referencing the preset

  • roi: list [valuemin, valuemax], range of interest, in units of the data

  • function: str, name of the fit function, that must be impolemented in a child class. Function names must start with func_

  • p0: a list, default initial guess for fit parameters. Presumably numeric. str are also possible. For example ‘1/350’, see function p0_interpret_value; or “Background” that will return a interp1d interpolation function over the data stored in preset curve with label “Background”.

  • fixed: a list of booleans, True if parameter is fixed by default

  • showingui if the preset is shown in the gui, by default

  • sum_amp_gaussian: not needed, specific to CurveMCA.

Then, create a subclass containing the fit functions:

  • fit functions must be named starting with func_

  • also: overrides needed for additional behaviors (e.g. amplitude in CurveMCA)

  • also: overrides to add e.g. interpretation of text into numerical values

  • other needed behaviors

For implementation into Curve subclass, take care of notably:

  • class variable - FitHandler, or rather subclass of if containing the fit functions of interest

    FITHANDLER = FitHandler(os.path.join(_PATHXRD, FITPARAMETERS_FILE))

  • __init__:

    # in __init__, as class itself not defined at instantiation of class variable
self.FITHANDLER.set_typecurvefit(CurveXRD)  # to adapt with own Curve type

  • funcListGUI

    # fit-related elements
out += self.FITHANDLER.funcListGUI(self, **kwargs)

  • updateFitParam, to provide fit function to Curve.updateFitParam, and in case textual keywords are used in fit parameters for e.g. “Background”.

    def updateFitParam(self, *param):
    # override default behavior, additional things to do
    param, revert, func = self.FITHANDLER.updateFitParam_before(self, *param)
    # call base function, including parameter func
    super().updateFitParam(*param, func=func)
    # revert callable parameter to its initial string value
    self.FITHANDLER.updateFitParam_after(self, revert)
    return True

  • fit_explicit, so curve action can point curve.fit_explicit and not more complex dependency, and to host the docstring to report in printHelp()

    def fit_explicit(self, roi, *args, funcname="", preset_label="", p0_raw=None,
                 **kwargs):
    # docstring to write here between triple double-quotes
    return self.FITHANDLER.fit_explicit(self, roi, *args, funcname=funcname,
                           preset_label=preset_label, p0_raw=p0_raw, **kwargs)

  • printHelp

__init__(filename)

Methods

__init__(filename)

fit(curve, func, roi, p0, fixed)

fit_explicit(curve, roi, *_args[, funcname, ...])

Fit fit_explicit: fit data, for example a peak of a CurveMCA data.

fit_preset(curve, preset_label[, roi, p0, fixed])

Fit fit_preset: fit a Curve, for example a peak of a CurveMCA data.

fit_resamplex(deltax, curve)

Change the base of datapoints a fit result is displayed on

funcListGUI(curve, **kwargs)

List of options for GUI

funcListGUI_ifnewline(counter, _lbl, _func)

by dfault, new line every 4 parameters

funcListGUI_isafit(curve, **_kwargs)

If is a fit: updateFitParam, resamplex

funcListGUI_notafit_choicepresets(curve, ...)

if not a fit, show fit options

funcListGUI_notafit_preset(curve, preset, ...)

Fit using preset settings

funcListGUI_notafit_whichones(curve, **_kwargs)

Chooses which preset to show

get_fitfunction(funcname)

fit functions must have names starting with func_, and be implemented in a child class

get_preset(label)

Returns first preset which label matches label

get_presets(label[, startswith, fitdefined])

tolerant input, useful for GUI Curve Actions Returns a list of (idx, preset Curve)

p0_interpret(p0[, curve, interp])

Returns a list of parameters, after interpreting textual values e.g. contaiing +, * or /, or background curves.

set_guishowfitfunc(which, fitimmediately[, ...])

To make GUI show specific fit functions

set_typecurvefit(typecurvefit)

set the type of Curve to be fitted

updateFitParam_after(curve, revert)

Small bit of code performed after updateFitParam

updateFitParam_before(curve, *popt)

Small bit of code performed before updateFitParam
fit_explicit(curve, roi, *_args, funcname='', preset_label='', p0_raw=None, **kwargs)

Fit fit_explicit: fit data, for example a peak of a CurveMCA data. Pre-set fit parameters are configured in file indicated at instantiation.

Parameters:

  • curve – a Curve object containing the data to fit.

  • roi – range of interest, in unit of channel.

  • funcname – name of the fit function. Mandatory.

  • preset_label – label of the fit preset in the configuration file, if relevant.

  • p0_raw – p0 values of the fit preset in configuration file, if relevant.

  • kwargs –

    to provide custom initial guess p0. Weird syntax for grapa GUI.

    Syntax 1: keywords “p0” and “fixed” are provided. Instances of list.

    Syntax 2: to meet limitations of grapa GUI. “0p0”: 1.0, “0fi”: True (i.e. is fixed fit parameter), “1p0”: 2.1, etc. These are later compiled into lists: p0 and fixed

Returns:

a CurveMCA fit curve

fit_preset(curve, preset_label, roi='auto', p0='auto', fixed='auto')

Fit fit_preset: fit a Curve, for example a peak of a CurveMCA data. Pre-set fit parameters are configured in file XRF_fitparameters.txt.

Parameters:

  • curve – a Curve object contaiing the data to fit

  • preset_label – see row “label” in config file. e.g. “Cu Ka1,2”, “Ag Ka1,2”

  • roi – range of interest, in unit of channel. Overwrite preset.

  • p0 – list, to provide custom initial guess p0. Overwrite preset.

  • fixed – list of boolean, to provide custom initial guess p0. Overwrite reset.

Returns:

a CurveMCA fit curve

fit_resamplex(deltax, curve)

Change the base of datapoints a fit result is displayed on

funcListGUI(curve, **kwargs)

List of options for GUI

funcListGUI_ifnewline(counter, _lbl, _func)

by dfault, new line every 4 parameters

funcListGUI_isafit(curve, **_kwargs)

If is a fit: updateFitParam, resamplex

funcListGUI_notafit_choicepresets(curve, **_kwargs)

if not a fit, show fit options

funcListGUI_notafit_preset(curve, preset, **_kwargs)

Fit using preset settings

funcListGUI_notafit_whichones(curve, **_kwargs)

Chooses which preset to show

get_fitfunction(funcname)

fit functions must have names starting with func_, and be implemented in a child class

get_preset(label)

Returns first preset which label matches label

get_presets(label, startswith=True, fitdefined=True)

tolerant input, useful for GUI Curve Actions Returns a list of (idx, preset Curve)

p0_interpret(p0, curve=None, interp=True)

Returns a list of parameters, after interpreting textual values e.g. contaiing +, * or /, or background curves

set_guishowfitfunc(which, fitimmediately, curve: Curve = None)

To make GUI show specific fit functions

Parameters:

  • which – label of the function to show

  • fitimmediately – if True, fit immediately

  • curve – must be provided

set_typecurvefit(typecurvefit)

set the type of Curve to be fitted

updateFitParam_after(curve, revert)

Small bit of code performed after updateFitParam

updateFitParam_before(curve, *popt)

Small bit of code performed before updateFitParam