Utils

Utils for the ESPM package

espm.utils.approx_density(atomic_fraction=False, *, elements_dict={})[source]: Wrapper to the density_of_mixture function of hyperspy. Takes a dict of chemical composition expressed in atomic weight fractions. Returns an approximated density.

espm.utils.arg_helper(params, d_params, replace=True)[source]

Check if all parameter of d_params are in params. If not, they are added to params with the default value.

Parameters:

paramsdict: Dictionary of parameters to be checked.
d_paramsdict: Dictionary of default parameters.

Returns:

paramsdict: Dictionary of parameters with the default parameters added if not present.

espm.utils.atomic_to_weight_dict(*, elements_dict={})[source]: Wrapper to the atomic_to_weight function of hyperspy. Takes a dict of chemical composition expressed in atomic fractions. Returns a dict of chemical composition expressed in atomic weight fratiom.

espm.utils.bin_spim(data, n, m)[source]: Take a 3D array of size (x,y,k) [px, py, e] Returns a 3D array of size (n,m,k) [new_px, new_py, e]

espm.utils.check_keys(params, d_params, upperkeys='', toprint=True, replace=True)[source]

Check if all parameter of d_params are in params. If not, they are added to params with the default value.

Parameters:

paramsdict: Dictionary of parameters to be checked.
d_paramsdict: Dictionary of default parameters.
upperkeysstr: String of the upper keys.
toprintbool: If True, print the warning.

Returns:

paramsdict: Dictionary of parameters with the default parameters added if not present.

Examples

>>> params = {'a':1,'b':2}
>>> d_params = {'a':1,'b':2,'c':3}
>>> check_keys(params,d_params)
>>> params
{'a': 1, 'b': 2, 'c': 3}

espm.utils.close_all()[source]: Close all opened windows.

espm.utils.cluster_analysis_concentration_report(s, cluster_source=None, print_std=False)[source]

Performs quantification in atomic % each cluster. s must have been quantified using .elemental_mapping() method.

Parameters:

shs.signals.Signal1D: Quantified Spectrum Image. Must have valid s.quantification_signal_1d.
cluster_source: None, hs.signals.Signal1D or Numpy Array: Object that contains the labeled coordinates where to average the quantification. If None, s.get_cluster_labels() is called.
print_stdBool: If True, the method also prints the standard deviation of quantifications for each element, for each cluster.

Returns:

quantificationnp.ndarray: An array containing the atomic % of each element averaged for each cluster.
stdnp.ndarray: An array containing the std of the atomic % for each element in each cluster.

espm.utils.create_laplacian_matrix(nx, ny=None)[source]

Helper method to create the laplacian matrix for the laplacian regularization

Parameters:

:param nx: height of the original image
:param ny: width of the original image

Returns:

rtype:: scipy.sparse.csr_matrix ..

:return:the n x n laplacian matrix, where n = nx*ny

espm.utils.erode_masks(masks, erosion_radius=1, footprint=array([[1., 1., 1.], [1., 1., 1.], [1., 1., 1.]]), extra_safe=False)[source]

Erodes masks so as to prevent phase overlap regions

Parameters:

masks: np.array: Tipically labels from cluster analysis
erosion_radius: int: The higher the more the masks are eroded
footprint: See skimage.morphology.binary erosion
extra_safe: bool: Perfroms erosion before opening/closing. In practice it creats a harsher erosion.

espm.utils.fancy_cluster_plot(s)[source]

Performs Cluster label plot with fancy colors.

Parameters:

shs.signals.Signal1D: s should be a signal for which you have performed cluster analysis.

espm.utils.get_explained_intensity_W(G, W, H)[source]

Compute the explained intensity of each element of W.

Parameters:

G (np.array 2D) – G matrix of the ESpM-NMF decomposition
W (np.array 2D) – W matrix of the ESpM-NMF decomposition
H (np.array 2D) – H matrix of the ESpM-NMF decomposition

Returns:

np.array 2D

espm.utils.is_number(i)[source]

Return True if i is a number

Parameters:: i (any) – variable to check
Returns:: True if i is a number
Return type:: bool

espm.utils.is_symbol(i)[source]

Return True if i is a chemical symbol

Parameters:: i (any) – variable to check
Returns:: True if i is a chemical symbol
Return type:: bool

espm.utils.isdict(p)[source]

Return True if the variable a dictionary.

Parameters:: p (any) – variable to check
Returns:: True if p is a dictionary
Return type:: bool

espm.utils.num_to_symbol(num)[source]

Converts number to atomic symbol.

Parameters:

numstr: Number to be converted to atomic symbol. E. g. “1” return “H”

Returns:

elementstr: Corresponding atomic symbol.

espm.utils.number_to_symbol_dict(func)[source]: Decorator Takes a dict of elements (a.k.a chemical composition) with atomic numbers as keys (e.g. 26 for Fe) returns a dict of elements with symbols as keys (e.g. Fe for iron)

espm.utils.number_to_symbol_list(func)[source]: Decorator Takes a dict of elements (a.k.a chemical composition) with symbols as keys (e.g. Fe for iron) returns a dict of elements with atomic numbers as keys (e.g. 26 for iron)

espm.utils.process_losses(losses)[source]

Process the losses to be plotted

Parameters:

losses: np.ndarray: Array of losses (output of espm.estimators.NMFEstimator.get_losses method)

Returns:

values: np.ndarray: Array of values
names: list: List of names

espm.utils.quant_profile_linear_fit(pf, **linfit_kwargs)[source]

Performs linear fit on profile data.

Parameters:

pflist of hs.signals.Signal1D: Output of s.quantification_profile
linfit_kwargs: are passed to sklearn.linear_model.LinearRegression

espm.utils.quant_spectrum(s1, skip_elements=[])[source]

Performs quantification in atomic % for a single spectrum. Elements from the metadata of another spectrum can be passed. The quantification is done using SmoothNMF with one component so as to only do the fitting.

Parameters:

s1hs.signals.Signal1D: Spectrum to be quantified. Elements to be quantified should be defined in the metadata.

Returns:

quantificationdict: A dictionary containing the atomic % of each element.
s1hs.signals.EDS_espm: The spectrum after doing espm quantification. The object contains all espm related information.

espm.utils.radial_profile(data, center='middle')[source]

espm.utils.rescaled_DH(D, H)[source]

Rescale the matrices D and H such that the columns of H sums approximately to one.

Parameters:

D (np.array 2D) – n x k matrix
H (np.array 2D) – k x m matrix

Returns:

D_rescale, H_rescale

Return type:

np.array 2D, np.array 2D

espm.utils.symbol_list()[source]

espm.utils.symbol_to_number_dict(func)[source]: Decorator Takes a dict of elements (a.k.a chemical composition) with symbols as keys (e.g. Fe for iron) returns a dict of elements with atomic numbers as keys (e.g. 26 for iron)

espm.utils.symbol_to_number_list(func)[source]: Decorator Takes a dict of elements (a.k.a chemical composition) with symbols as keys (e.g. Fe for iron) returns a dict of elements with atomic numbers as keys (e.g. 26 for iron)