mdigest.core.plots package

Submodules

mdigest.core.plots.dim_reduction_plots module

mdigest.core.plots.dim_reduction_plots.plot_pca_space(pcfit, labelsarr, names_list, out, **kwargs)

Plot pca space

Parameters:

  • pcfit (np.ndarray,) – array of fitted transformed data from [pca_call = sklearn.PCA() –> pcfit = pca_call.fit_transform(data)]

  • labelsarr (np.ndarray,) –

    array of labels

    example: if the data is the concatenation of two simulations then data = np.concatenate([sim_1, sim_2]) and labelsarr = np.asarray(names_list[0]*len(sim_2) + names_list[1]*len(sim_2)); the order has to match the data.

  • names_list (list of strings) – list containing the title name of each simulation

  • out (str,) – plot name ('/path/to/plot/plotname.pdf')

mdigest.core.plots.general_plots module

mdigest.core.plots.general_plots.plot_eigenvalues_distribution(**kwargs)

Plot the eigenvalues of a given correlation matrix.

Parameters:

kwargs (dict, possible options are:) –

  • matrix, np.ndarray correlation matrix

  • eigenvalues, np.ndarray of eigenvalues

  • dim, int specifying number of eigenvalues to display

  • color, sns.palette or array of colors

  • ax, mpl.ax object

  • matrix_type, str, title for the legend

if ‘matrix’ is provided the latter will be diagonalized and the first n eigenvalues shown as a barplot; alternatively, use ‘eigenvalues’.

mdigest.core.plots.plot_correlation module

class mdigest.core.plots.plot_correlation.Plots(MDSIMA, MDSIMB, matrix_type='gcc_lmi', compute_centrality=True, **kwargs)

Bases: object

load_parameters(params)

Load parameters

Parameters:

params (dict,) – dictionary containing parameters for plotting example: params={'labels':['APO', 'HOLO'], 'fig': plt.subplots(self.num_replicas, 3, figsize=(3*self.num_replicas+10,1*self.num_replicas+3.5))}

plot_eigcent_per_replica()

Plot eigenvector centrality for each replica

plot_gcc_per_replica()

Plot difference heatmaps of matrices for each replica

Module contents

MDiGest v.0.1.0

__version__ = 0.1.0 __author__ = Federica Maschietto <federica.maschietto@gmail.com>, Brandon Allen <bcallen95@gmail.com>