Deglitch module

araucaria.xas.deglitch.deglitch(group, e_window='xas', sg_window_length=9, sg_polyorder=3, alpha=0.025, max_glitches='default', max_glitch_length=4, update=False)[source]

Algorithm to deglitch a XAFS spectrum.

Parameters

  • group (Group) – Group containing the spectrum to deglitch.

  • e_window (Union[str, list]) – Energy window to seach for outliers. Oprions are ‘xas’, ‘xanes’ and ‘exafs’. Alternatively a list with 2 floats can be provided for the start and end energy for the search. See the Notes for further details. The default is ‘xas’.

  • sg_window_length (int) – Windows length for the Savitzky-Golay filter on the normalized spectrum. Must be an odd value. The default is 7.

  • sg_polyorder (int) – Polynomial order for the Savitzky-Golay filter on the normalized spectrum. The default is 3.

  • alpha (float) – Significance level for generalized ESD test for outliers. The default is 0.025.

  • max_glitches (Union[int, str]) – Maximum number of outliers to remove. The default is the floor division of the array length by 10.

  • max_glitch_length (int) – Maximum length of glitch in energy points. The default is 4.

  • update (bool) – Indicates if the group should be updated with the autobk attributes. The default is False.

Return type

dict

Returns

Dictionary with the following arguments:

  • index_glitches : indices of glitches in the original energy array.

  • energy_glitches: glitches in the original energy array.

  • energy : deglitched energy array.

  • mu : deglitched array. Returned if group.get_mode() = 'mu'.

  • fluo : deglitched array. Returned if group.get_mode() = 'fluo'.

  • mu_ref : deglitched array. Returned if group.get_mode() = 'mu_ref'.

  • deglitch_pars : dictionary with deglitch parameters.

Raises

  • TypeError – If group is not a valid Group instance.

  • AttributeError – If attribute energy does not exist in group.

Warning

Running deglitch() with update=True will overwrite the energy and the absorption attribute of group.

Notes

This function deglitches a XAFS spectrum through a two-step fitting with Savitzky-Golay filter and outlier identification with a generalized extreme Studentized deviate (ESD) test 1.

  • e_window='xas' considers the full spectrum for deglitching.

  • e_window='xanes' considers the beginning of the energy array up to 150 eV above \(E_0\).

  • e_window='exafs' considers from 150eV above \(E_0\) to the end of the energy array

  • e_window=[float,float] provides start and end energies in eV.

References

1

Wallace, S. M., Alsina, M. A., & Gaillard, J. F. (2021) “An algorithm for the automatic deglitching of x-ray absorption spectroscopy data”. J. Synchrotron Rad. 28, https://doi.org/10.1107/S1600577521003611

Example

>>> from numpy import allclose
>>> from araucaria.testdata import get_testpath
>>> from araucaria import Group
>>> from araucaria.io import read_dnd
>>> from araucaria.xas import deglitch, pre_edge, autobk
>>> from araucaria.utils import check_objattrs
>>> fpath  = get_testpath('dnd_glitchfile.dat')
>>> group  = read_dnd(fpath, scan='fluo')  # extracting fluo and mu_ref scans
>>> cgroup = group.copy()
>>> degli  = deglitch(cgroup, update=True)
>>> attrs  = ['index_glitches', 'energy_glitches', 'deglitch_pars']
>>> check_objattrs(cgroup, Group, attrs)
[True, True, True]
>>> allclose(cgroup.energy_glitches, group.energy[cgroup.index_glitches])
True
>>> print(cgroup.energy_glitches)
[7552.2789 7548.1747 7390.512  7387.2613]

>>> # plotting original and deglitched spectrum
>>> from araucaria.plot import fig_xas_template
>>> import matplotlib.pyplot as plt
>>> for g in [group, cgroup]:
...     pre   = pre_edge(g, update=True)
...     autbk = autobk(g, update=True)
>>> fig, ax = fig_xas_template(panels='xe')
>>> line = ax[0].plot(group.energy,  group.norm,  label='original', color='tab:red')
>>> line = ax[0].plot(cgroup.energy, cgroup.norm, label ='degliched', color='k')
>>> line = ax[1].plot(group.k, group.k**2 * group.chi, color='tab:red')
>>> line = ax[1].plot(cgroup.k, cgroup.k**2 * cgroup.chi, color='k')
>>> leg  = ax[0].legend()
>>> fig.tight_layout()
>>> plt.show(block=False)

(Source code, png, hires.png, pdf)

../_images/xas_deglitch-1.png