visualization module

visualization module#

Organization: InsightSolver Solutions Inc.
Project Name: InsightSolver
Module Name: insightsolver
File Name: visualization.py
Authors: Noé Aubin-Cadot <noe.aubin-cadot@insightsolver.com>,
Arthur Albo <arthur.albo@insightsolver.com>

Description#

This file contains some visualization functions, some of which are integrated as a method of the InsightSolver class.

Naming conventions of the visualization functions#

draw_: Draws the content on a specified plt.Axes.
make_: Makes the plt.Figure with the content on it and return the figure.
plot_: Plots the content and show it to the user.

Utilities#

classify_variable_as_continuous_or_categorical: Classifies a variable as continuous or categorical.
compute_feature_label: Computes the label of a feature.
truncate_label: Truncates a label.
p_value_to_p_text: Converts the p-value to a text.
svg_to_pil: Converts a SVG to a PIL image.
wrap_text_with_word_boundary: Wrap text on multiple lines.
save_figs_in_pdf: Stack multiple figures vertically on a page and save to PDF.

Mutual information#

draw_mutual_information: Draws the mutual information on a given plt.Axes.
make_mutual_information: Makes the mutual information plt.Figure.
plot_mutual_information: Plots the mutual information and show it to the user.

Legend#

make_legend_img: Makes the legend image.
plot_legend_img: Plots the legend image and show it to the user.
make_legend_fig: Makes the legend figure.
plot_legend_fig: Plots the legend figure and show it to the user.

Feature contributions#

draw_feature_contributions_for_i: Draws the feature contributions for the rule at index i on a given plt.Axes.
make_feature_contributions_for_i: Makes the feature contributions for the rule at index i figures as a plt.Figure.
plot_feature_contributions_for_i: Plots the feature contributions for the rule at index i and show them to the user.
plot_feature_contributions_for_all: Plots the feature contributions for all rules.

Feature distribution for feature#

draw_feature_distribution_for_feature: Draws the distribution for a specified feature on a given plt.Axes.
make_feature_distribution_for_feature: Makes the distribution for a specified feature as a plt.Figure.
plot_feature_distribution_for_feature: Plots the distribution for a specified feature and show it to the user.

Feature distributions for S#

draw_feature_distributions_for_S: Draws the distributions of all features in a rule S on a given plt.Axes.
make_feature_distributions_for_S: Makes the distributions of all features in a rule S as a plt.Figure.
plot_feature_distributions_for_S: Plots the distributions of all features in a rule S and show it to the user.

Mosaic of rule vs complement for a feature for the rule i#

draw_mosaic_rule_vs_comp_for_feature_for_i: Draws the mosaic for a specified feature for the rule at index i on a given plt.Axes.
make_mosaic_rule_vs_comp_for_feature_for_i: Makes the mosaic for a specified feature for the rule at index i as a plt.Figure.
plot_mosaic_rule_vs_comp_for_feature_for_i: Plots the mosaic for a specified feature for the rule at index i and show it to the user.

Mosaics of rule vs complement for the rule i#

draw_mosaics_rule_vs_comp_for_i: Draws the mosaics for the full rule at index i and for restricted rules on a given plt.Axes.
make_mosaics_rule_vs_comp_for_i: Makes the mosaics for the full rule at index i and for restricted rules on its features as a plt.Figure.
plot_mosaics_rule_vs_comp_for_i: Plots the mosaics for the full rule at index i and for restricted rules on its features and show it to the user.

Mosaic of rule vs pop vs complement for the rule i#

draw_mosaic_rule_vs_pop_for_i: Draws the mosaic plot comparing rule vs population on a given plt.Axes.
make_mosaic_rule_vs_pop_for_i: Makes the mosaic plot comparing rule vs population as a plt.Figure.
plot_mosaic_rule_vs_pop_for_i: Plots the mosaic plot comparing rule vs population and show it to the user.

Mosaics of rule vs pop vs complement#

draw_mosaics_rule_vs_pop: Draw the mosaics for all rules in the ruleset on a given plt.Axes.
make_mosaics_rule_vs_pop: Makes the mosaics for all rules in the ruleset as a plt.Figure.
plot_mosaics_rule_vs_pop: Plots the mosaics for all rules in the ruleset and show it to the user.

Complete plot#

make_all: Makes all visualization figures and returns them as a list of tuples (name, figure).
plot_all: Plots all visualization figures and show them to the user.

Export to PDF#

make_pdf: Exports a PDF of all the figures.

Export to ZIP#

make_zip: Exports a ZIP of all the content related to the solver.

License#

Exclusive Use License - see LICENSE for details.

insightsolver.visualization.classify_variable_as_continuous_or_categorical(s: Series, unique_ratio_threshold: float = 0.1, max_categories: int = 20) → str#

Classify a pandas Series as ‘continuous’ or ‘categorical’.

Heuristic#

If dtype is object/string/bool → categorical
If all values are equal → categorical
If all values are integers: - Few unique values (<= max_categories) → categorical - Low unique ratio (<= unique_ratio_threshold) → categorical
Otherwise → continuous

Parameters#

spd.Series: Input series.
unique_ratio_thresholdfloat, optional: Threshold for ratio (#unique / #non-missing) to treat integers as categorical.
max_categoriesint, optional: Absolute cap for number of unique categories to treat as categorical.

Returns#

str: “categorical” or “continuous”

insightsolver.visualization.compute_feature_label(solver, feature_name: str, S: dict) → [<class 'str'>, <class 'str'>]#

This function computes the label of a feature in a rule S.

Parameters#

solver: InsightSolver: The solver.
feature_name: str: The name of the feature.
S: dict: The rule S.

Returns#

feature_label: str: The label of the feature.
feature_relationship: str: The relationship of the feature to the constraints.

insightsolver.visualization.truncate_label(label, max_length=30, asterisk=False)#

This function truncates a string if it exceeds a specified length, adding an ellipsis.

Parameters#

label: string: the feature rule’s modalities.
max_length: int: the maximum number of character accepted.
asterisk: bool: whether we want an asterisk to appear after the truncation.

Returns#

truncated_label: str: The truncated label.

insightsolver.visualization.p_value_to_p_text(p_value, precision_p_values: str) → str#

This function converts the p-value to a string.

Parameters#

p_value: float or mpmath.mpf: The p-value to convert.
precision_p_values: str: The precision of the p-values.

Returns#

p_text: str: The p_value formatted as a string.

insightsolver.visualization.svg_to_pil(svg_filename: str, assets_package: str = 'insightsolver.assets', subfolder: str = 'google_fonts_icons', size: tuple[int, int] = (80, 80))#

Convert an SVG file resource into a PIL (Pillow) Image object with a specified size.

Parameters#

svg_filename: str: The filename of the SVG icon located in the assets subfolder.
assets_package: str: The Python package name where the assets are located (e.g., ‘insightsolver.assets’).
subfolder: str: The subfolder within the assets package where the SVG file resides.
size: tuple(int, int): The target size (width, height) in pixels for the output PIL Image.

Returns#

imgPIL.Image.Image: The converted image as a PIL Image object, typically in RGBA format.

insightsolver.visualization.wrap_text_with_word_boundary(text: str, max_line_length: int = 150) → str#

Wraps a text string into multiple lines by inserting line breaks around a target character width, while preserving word boundaries whenever possible.

If the next word would cause the line to exceed max_line_length, a line break is inserted before that word.
If a single word is longer than max_line_length, the word is split with a hyphen followed by a line break.

Parameters#

textstr: The input text to wrap.
max_line_lengthint, optional: The maximum allowed line length before wrapping occurs (default is 150).

Returns#

str: The wrapped string, with line breaks (and occasional hyphenation) inserted at appropriate positions.

insightsolver.visualization.save_figs_in_pdf(figs: list, pdf, do_padding: bool = True)#

Stack multiple figures vertically on a page and save to PDF.

Parameters#

figslist of matplotlib.figure.Figure: The figures to stack.
pdfPdfPages: The PdfPages object to save to.
do_padding: bool, default True: If a padding should be present in the pdf.

insightsolver.visualization.draw_mutual_information(ax: Axes, s_mi: Series, kind: str = 'barh') → Axes#

Draws the mutual information bar plot on the given axes.

Parameters#

axplt.Axes: The axes to draw on.
s_mipd.Series: The mutual information series.
kindstr: Kind of plot (‘bar’ or ‘barh’).

Returns#

axplt.Axes: The axes with the plot.

insightsolver.visualization.make_mutual_information(solver, n_samples: int | None = 1000, n_cols: int | None = 20, kind: str = 'barh', fig_width: float = 12) → Figure#

Creates a figure showing the mutual information.

Parameters#

solverInsightSolver: The solver object.
n_samplesint: Number of samples.
n_colsint: Max number of columns.
kindstr: Kind of plot.
fig_widthfloat: Width of the figure.

Returns#

figplt.Figure: The created figure.

insightsolver.visualization.plot_mutual_information(solver, n_samples: int | None = 1000, n_cols: int | None = 20, kind: str = 'barh', fig_width: float = 12) → None#: Displays the mutual information plot.

insightsolver.visualization.make_banner_img_for_i(solver, i: int, loss: float = None, fig_width: float = 12, dpi: int = 200, icon_size: tuple[int, int] = (80, 80))#

Generate a dynamic InsightSolver banner composed of SVG icons and text.

Parameters#

solverInsightSolver: The solver containing the rules.
iint: Index of the rule to display.
lossfloat, optional: Optional loss value to display.
fig_widthfloat: Width of the banner (in inches).
dpiint: DPI resolution (pixels per inch).
icon_sizetuple: Icon size in pixels (width, height).

Returns#

PIL.Image: The generated banner.

insightsolver.visualization.plot_banner_img_for_i(solver, i: int, loss: float = None, fig_width: float = 12, dpi: int = 200, icon_size: tuple[int, int] = (80, 80)) → None#

Displays the banner image of the statistics of the rule at index i.

Parameters#

solverInsightSolver: The fitted solver.
iint: The index of the rule to be displayed in the banner.
lossfloat: An optional loss value to display in the banner.
fig_widthfloat: Width of the banner in inches.
dpiint: DPI resolution (pixels per inch).
icon_sizetuple[int, int]: Icon size in pixels (width, height).

insightsolver.visualization.make_banner_fig_for_i(solver, i: int, loss: float = None, fig_width: float = 12, dpi: int = 200, icon_size: tuple[int, int] = (80, 80)) → Figure#

This function generates a figure of the banner of the statistics of the rule at index i.

Parameters#

solver: InsightSolver: The fitted solver.
i: int: The index of the rule to be displayed in the banner.
loss: float: An optional loss value to display in the banner.
fig_width: float: Width of the figure in inches. Height is automatically adjusted.
dpi: int: Resolution (dots per inch) of the figure.
icon_size: tuple[int, int]: Size of the icons used in the banner image (width, height in pixels).

Returns#

fig_bannermatplotlib.figure.Figure: The Matplotlib Figure object containing the rule banner.

insightsolver.visualization.plot_banner_fig_for_i(solver, i: int, loss: float = None, fig_width: float = 12, dpi: int = 200, icon_size: tuple[int, int] = (80, 80)) → None#

Displays the banner of the statistics of the rule at index i.

Parameters#

solverInsightSolver: The fitted solver.
iint: The index of the rule to be displayed in the banner.
lossfloat: An optional loss value to display in the banner.
fig_widthfloat: Width of the figure in inches. Height is automatically adjusted.
dpiint: Resolution (dots per inch) of the figure.
icon_sizetuple[int, int]: Size of the icons used in the banner image (width, height in pixels).

insightsolver.visualization.make_legend_img(do_plot_loss: bool = True, fig_width: float = 12, dpi: int = 200, icon_size: tuple[int, int] = (80, 80), language: str = 'en', verbose: bool = False)#

Generate a legend that explains what the icons of the legend represent.

Parameters#

do_plot_lossbool: If we want to describe the symbol for the loss in the legend.
fig_widthfloat: Width of the legend (in inches).
dpiint: DPI resolution (pixels per inch).
icon_sizetuple: Icon size in pixels (width, height).
verbose: bool: Verbosity.

Returns#

PIL.Image: The generated legend.

insightsolver.visualization.plot_legend_img(do_plot_loss: bool = True, fig_width: float = 12, dpi: int = 200, icon_size: tuple[int, int] = (80, 80), language: str = 'en', verbose: bool = False) → None#

Displays the legend image.

Parameters#

do_plot_lossbool: If True, describe the symbol for the loss in the legend.
fig_widthfloat: Width of the legend (in inches).
dpiint: DPI resolution (pixels per inch).
icon_sizetuple: Icon size in pixels (width, height).
languagestr: Language of the text labels (‘fr’ or ‘en’).
verbosebool: If True, display verbose output.

insightsolver.visualization.make_legend_fig(do_plot_loss: bool = False, fig_width: float = 12, dpi: int = 200, icon_size: tuple[int, int] = (80, 80), language: str = 'en', verbose: bool = False) → Figure#

This function generates a figure of the legend which explains the meaning of the icons in the banner.

Parameters#

do_plot_loss: bool: If True, describe the loss icon in the legend.
fig_width: float: The width of the figure to generate.
dpi: int: The dots per inch (resolution) for the figure.
icon_size: tuple[int, int]: The size (width, height) in pixels for the icons used within the legend image.
language: str: Language of the text labels in the figure legend (‘fr’ or ‘en’).
verbose: bool: If True, display verbose output during the legend image generation process.

Returns#

fig_legendmatplotlib.figure.Figure: The Matplotlib Figure object containing the legend image.

insightsolver.visualization.plot_legend_fig(do_plot_loss: bool = False, fig_width: float = 12, dpi: int = 200, icon_size: tuple[int, int] = (80, 80), language: str = 'en', verbose: bool = False) → None#

Displays the legend.

Parameters#

do_plot_loss: bool: If True, describe the loss icon in the legend.
fig_width: float: The width of the figure to generate.
dpi: int: The dots per inch (resolution) for the figure.
icon_size: tuple[int, int]: The size (width, height) in pixels for the icons used within the legend image.
language: str: Language of the text labels in the figure legend (‘fr’ or ‘en’).
verbose: bool: If True, display verbose output during the legend image generation process.

insightsolver.visualization.draw_feature_contributions_for_i(ax: Axes, df_feature_contributions_S: DataFrame, language: str = 'en', do_grid: bool = True, do_title: bool = False, i: int | None = None, rule_i: dict | None = None, precision_p_values: str = 'float64', bar_annotations: str = 'p_value_ratio') → Axes#

Draws the feature contributions bar plot on the given axes.

Parameters#

axplt.Axes: The axes to draw on.
df_feature_contributions_Spd.DataFrame: The DataFrame containing feature contributions.
languagestr: Language of the plot (‘en’ or ‘fr’).
do_gridbool: If True, show the grid.
do_titlebool: If True, show the title.
iint, optional: Index of the rule.
rule_idict, optional: The rule dictionary.
precision_p_valuesstr: Precision of p-values (‘float64’ or ‘mpmath’).
bar_annotationsstr: Type of annotations on bars (‘p_value_ratio’, ‘p_value_contribution’, or None).

Returns#

axplt.Axes: The axes with the plot.

insightsolver.visualization.make_feature_contributions_for_i(solver, i: int, a: float = 0.5, b: float = 1, fig_width: float = 12, language: str = 'en', do_grid: bool = True, do_title: bool = False, do_banner: bool = True, bar_annotations: str = 'p_value_ratio', loss: float | None = None) → List[Figure]#

Creates the feature contributions figures (banner, plot, details).

Parameters#

solverInsightSolver: The solver object.
iint: Index of the rule to show.
afloat: Height per bar.
bfloat: Height for the margins and other elements.
fig_widthfloat: Width of the figure in inches.
languagestr: Language of the figure.
do_gridbool: If True, show a vertical grid.
do_titlebool: If True, show a title automatically generated.
do_bannerbool: If True, show the banner.
bar_annotationsstr: Type of values to show at the end of the bars.
lossfloat, optional: If we want to show a loss.

Returns#

figsList[plt.Figure]: A list of figures (banner, plot, details).

insightsolver.visualization.plot_feature_contributions_for_i(solver, i: int, a: float = 0.5, b: float = 1, fig_width: float = 12, language: str = 'en', do_grid: bool = True, do_title: bool = False, do_banner: bool = True, bar_annotations: str = 'p_value_ratio', loss: float | None = None) → None#

Displays the feature contributions figures.

Parameters#

solverInsightSolver: The solver object.
iint: Index of the rule to show.
afloat: Height per bar.
bfloat: Height for the margins and other elements.
fig_widthfloat: Width of the figure in inches.
languagestr: Language of the figure.
do_gridbool: If True, show a vertical grid.
do_titlebool: If True, show a title automatically generated.
do_bannerbool: If True, show the banner.
bar_annotationsstr: Type of values to show at the end of the bars.
lossfloat, optional: If we want to show a loss.

insightsolver.visualization.plot_feature_contributions_for_all(solver, a: float = 0.5, b: float = 1, fig_width: float = 12, language: str = 'en', do_grid: bool = True, do_title: bool = False, do_banner: bool = True, bar_annotations: str = 'p_value_ratio') → None#

This function generates a horizontal bar plot of the feature contributions for each rule found in a solver.

Parameters#

solverInsightSolver: The solver object.
afloat: Height per bar.
bfloat: Height for the margin and other elements.
fig_widthfloat: Width of the figure in inches.
languagestr: Language of the figure.
do_gridbool: If True, show a grid.
do_titlebool: If True, show a title which is automatically generated.
do_bannerbool: If True, show the banner.
bar_annotationsstr: Type of values to show at the end of the bars.

insightsolver.visualization.draw_feature_distribution_for_feature(ax: Axes, solver, df_filtered: DataFrame, S: dict, feature_name: str, missing_value: bool = False, language: str = 'en', padding_y: int = 5, do_plot_kde: bool = False, do_plot_vertical_lines: bool = False, verbose: bool = False) → Axes#

Draws the distribution of a feature on the given axes.

Parameters#

axplt.Axes: The axes to draw on.
solverInsightSolver: The solver object.
df_filteredpd.DataFrame: The filtered DataFrame.
Sdict: The rule S.
feature_namestr: The name of the feature.
missing_valuebool: If True, plot the missing values.
languagestr: Language of the plot (‘en’ or ‘fr’).
padding_yint: Padding for the y-axis.
do_plot_kdebool: If True, show the KDE plot.
do_plot_vertical_linesbool: If True, show vertical lines for the rule boundaries.
verbosebool: If True, print verbose output.

Returns#

axplt.Axes: The axes with the plot.

insightsolver.visualization.make_feature_distribution_for_feature(solver, df_filtered: DataFrame, S: dict, feature_name: str, missing_value: bool = False, language: str = 'en', padding_y: int = 5, do_plot_kde: bool = False, do_plot_vertical_lines: bool = False, fig_width: float = 12, verbose: bool = False) → Figure#

Creates a figure showing the distribution of a feature.

Parameters#

solverInsightSolver: The solver object.
df_filteredpd.DataFrame: The filtered DataFrame.
Sdict: The rule S.
feature_namestr: The name of the feature.
missing_valuebool: If True, plot the missing values.
languagestr: Language of the plot (‘en’ or ‘fr’).
padding_yint: Padding for the y-axis.
do_plot_kdebool: If True, show the KDE plot.
do_plot_vertical_linesbool: If True, show vertical lines for the rule boundaries.
fig_widthfloat: Width of the figure.
verbosebool: If True, print verbose output.

Returns#

figplt.Figure: The created figure.

insightsolver.visualization.plot_feature_distribution_for_feature(solver, df_filtered: DataFrame, S: dict, feature_name: str, missing_value: bool = False, language: str = 'en', padding_y: int = 5, do_plot_kde: bool = False, do_plot_vertical_lines: bool = False, fig_width: float = 12, verbose: bool = False) → None#

Displays the distribution of a feature.

Parameters#

solverInsightSolver: The solver object.
df_filteredpd.DataFrame: The filtered DataFrame.
Sdict: The rule S.
feature_namestr: The name of the feature.
missing_valuebool: If True, plot the missing values.
languagestr: Language of the plot (‘en’ or ‘fr’).
padding_yint: Padding for the y-axis.
do_plot_kdebool: If True, show the KDE plot.
do_plot_vertical_linesbool: If True, show vertical lines for the rule boundaries.
fig_widthfloat: Width of the figure.
verbosebool: If True, print verbose output.

insightsolver.visualization.draw_feature_distributions_for_S(axes_list: List[Axes], solver, S: dict, language: str = 'en', padding_y: int = 5, do_plot_kde: bool = False, do_plot_vertical_lines: bool = False) → List[Axes]#

Draws the distributions of all features in rule S on the given list of axes.

Parameters#

axes_listList[plt.Axes]: List of axes to draw on (one or two axes per feature depending on missing values).
solverInsightSolver: The solver object.
Sdict: The rule S.
languagestr: Language of the plot (‘en’ or ‘fr’).
padding_yint: Padding for the y-axis.
do_plot_kdebool: If True, show the KDE plot.
do_plot_vertical_linesbool: If True, show vertical lines for the rule boundaries.

Returns#

axes_listList[plt.Axes]: The list of axes with the plots.

insightsolver.visualization.make_feature_distributions_for_S(solver, S: dict, language: str = 'en', padding_y: int = 5, do_plot_kde: bool = False, do_plot_vertical_lines: bool = False, fig_width: float = 12) → List[Figure]#

Creates figures showing the distributions of all features in rule S.

Parameters#

solverInsightSolver: The solver object.
Sdict: The rule S.
languagestr: Language of the plot (‘en’ or ‘fr’).
padding_yint: Padding for the y-axis.
do_plot_kdebool: If True, show the KDE plot.
do_plot_vertical_linesbool: If True, show vertical lines for the rule boundaries.
fig_widthfloat: Width of the figure in inches.

Returns#

figsList[plt.Figure]: A list of figures, one per feature.

insightsolver.visualization.plot_feature_distributions_for_S(solver, S: dict, language: str = 'en', padding_y: int = 5, do_plot_kde: bool = False, do_plot_vertical_lines: bool = False, fig_width: float = 12) → None#

This function generates bar plots of the distributions of the points in the specified rule S.

Parameters#

solverInsightSolver: The solver object.
Sdict: The rule S.
languagestr: Language of the plot (‘en’ or ‘fr’).
padding_yint: Padding for the y-axis.
do_plot_kdebool: If True, show the KDE plot.
do_plot_vertical_linesbool: If True, show vertical lines for the rule boundaries.
fig_widthfloat: Width of the figure in inches.

insightsolver.visualization.draw_mosaic_rule_vs_comp_for_feature_for_i(ax: Axes, solver, i: int, feature_name: str | None = None, verbose: bool = False) → Axes#

Draws the mosaic plot for the rule i on the given axes.

Parameters#

axplt.Axes: The axes to draw on.
solverInsightSolver: The solver object.
iint: Index of the rule.
feature_namestr, optional: Name of the feature to filter by.
verbose: bool: Verbosity

Returns#

axplt.Axes: The axes with the plot.

insightsolver.visualization.make_mosaic_rule_vs_comp_for_feature_for_i(solver, i: int, feature_name: str | None = None) → Figure#

Creates a figure showing the mosaic plot for the rule i.

Parameters#

solverInsightSolver: The solver object.
iint: Index of the rule.
feature_namestr, optional: Name of the feature to filter by.

Returns#

figplt.Figure: The created figure.

insightsolver.visualization.plot_mosaic_rule_vs_comp_for_feature_for_i(solver, i: int, feature_name: str | None = None) → None#

Displays the mosaic plot for the rule i.

Parameters#

solverInsightSolver: The solver object.
iint: Index of the rule.
feature_namestr, optional: Name of the feature to filter by.

insightsolver.visualization.draw_mosaics_rule_vs_comp_for_i(axes: list[Axes], solver, i: int, ncols: int = 3) → list[Axes]#

Draws mosaics of the rule vs complement for the whole rule and each feature.

Parameters#

axeslist of plt.Axes: Axes where to draw the plots.
solverInsightSolver: The solver object.
iint: Index of the rule.
ncolsint: Number of columns in the subplot grid.

Returns#

axeslist of plt.Axes: The axes with plots drawn.

insightsolver.visualization.make_mosaics_rule_vs_comp_for_i(solver, i: int, ncols: int = 3) → Figure#

Creates a figure showing mosaics for the whole rule and each feature.

Parameters#

solverInsightSolver: The solver object.
iint: Index of the rule.
ncolsint: Number of columns in the subplot grid.

Returns#

figplt.Figure: The created figure.

insightsolver.visualization.plot_mosaics_rule_vs_comp_for_i(solver, i: int, ncols: int = 3) → None#

Displays the mosaic plots of the rule at position i, including the whole rule and each feature.

Parameters#

solverInsightSolver: The solver object.
iint: Index of the rule.
ncolsint: Number of columns in the subplot grid.

insightsolver.visualization.draw_mosaic_rule_vs_pop_for_i(ax: Axes, solver, i: int, do_plot_comp: bool = True) → Axes#

Draws the mosaic plot that compares the purity of the rule at index i vs the population.

Parameters#

axplt.Axes: The axes to draw on.
solverInsightSolver: The solver object.
iint: Index of the rule.
do_plot_compbool: If True, show the complement.

Returns#

axplt.Axes: The axes with the plot.

insightsolver.visualization.make_mosaic_rule_vs_pop_for_i(solver, i: int, do_plot_comp: bool = True) → Figure#

Creates a figure showing the mosaic plot that compares the purity of the rule at index i vs the population.

Parameters#

solverInsightSolver: The solver object.
iint: Index of the rule.
do_plot_compbool: If True, show the complement.

Returns#

figplt.Figure: The created figure.

insightsolver.visualization.plot_mosaic_rule_vs_pop_for_i(solver, i: int, do_plot_comp: bool = True) → None#

Displays the mosaic plot that compares the purity of the rule at index i vs the population.

Parameters#

solverInsightSolver: The solver object.
iint: Index of the rule.
do_plot_compbool: If True, show the complement.

insightsolver.visualization.draw_mosaics_rule_vs_pop(axes: list[Axes], solver, do_plot_comp: bool = True) → list[Axes]#

Draws mosaics of the purity of all rules vs population (and complement if requested).

Parameters#

axeslist of plt.Axes: The axes where to draw the plots.
solverInsightSolver: The solver object.
do_plot_compbool: If True, show the complement.

Returns#

axeslist of plt.Axes: The axes with plots drawn.

insightsolver.visualization.make_mosaics_rule_vs_pop(solver, do_plot_comp: bool = True, ncols: int = 3, fig_width: float = 12) → Figure#

Creates a figure showing mosaics of purity vs population for all rules.

Parameters#

solverInsightSolver: The solver object.
do_plot_compbool: If True, show the complement.
ncolsint: Number of columns in the subplot grid.
fig_widthfloat: Width of the figure in inches.

Returns#

figplt.Figure: The created figure.

insightsolver.visualization.plot_mosaics_rule_vs_pop(solver, do_plot_comp: bool = True, ncols: int = 3, fig_width: float = 12, verbose: bool = False) → None#

Displays the mosaic plots of purity vs population for all rules.

Parameters#

solverInsightSolver: The solver object.
do_plot_compbool: If True, show the complement.
ncolsint: Number of columns in the subplot grid.
fig_widthfloat: Width of the figure in inches.
verbosebool: If True, prints debug information.

insightsolver.visualization.make_all(solver, language: str = 'en', do_mutual_information: bool = True, do_banner: bool = True, do_contributions: bool = True, do_distributions: bool = True, do_mosaics_rule_vs_comp: bool = True, do_mosaics_rule_vs_pop: bool = True, do_legend: bool = True) → List[tuple[str, Figure]]#

Creates all visualization figures for the solver.

This function generates (depending on boolean flags): - Mutual information plot - Banner for each rule - Feature contributions for each rule - Feature distributions for each rule - Mosaics of rule vs complement for each rule - Mosaics of rule vs population - Legend

Parameters#

solverInsightSolver: The fitted solver object.
languagestr: Language for the plots (‘en’ or ‘fr’).
do_mutual_informationbool: Whether to generate the mutual information figure.
do_bannerbool: Whether to generate the banner figure for each rule.
do_contributionsbool: Whether to generate the feature contributions figures.
do_distributionsbool: Whether to generate the feature distributions figures.
do_mosaics_rule_vs_compbool: Whether to generate the mosaics of rule vs complement figures.
do_mosaics_rule_vs_popbool: Whether to generate the mosaics of rule vs population figures.
do_legendbool: Whether to generate the legend figure.

Returns#

figsList[tuple[str, plt.Figure]]: A list of tuples containing (figure_name, figure_object).

insightsolver.visualization.plot_all(solver, language: str = 'en', do_mutual_information: bool = True, do_banner: bool = True, do_contributions: bool = True, do_distributions: bool = True, do_mosaics_rule_vs_comp: bool = True, do_mosaics_rule_vs_pop: bool = True, do_legend: bool = True) → None#

Displays all visualization figures for the solver.

This function displays (depending on boolean flags): - Mutual information plot - Banner for each rule - Feature contributions for each rule - Feature distributions for each rule - Mosaics of rule vs complement for each rule - Mosaics of rule vs population - Legend

Parameters#

solverInsightSolver: The fitted solver object.
languagestr: Language for the plots (‘en’ or ‘fr’).
do_mutual_informationbool: Whether to display the mutual information figure.
do_bannerbool: Whether to display the banner figure for each rule.
do_contributionsbool: Whether to display the feature contributions figures.
do_distributionsbool: Whether to display the feature distributions figures.
do_mosaics_rule_vs_compbool: Whether to display the mosaics of rule vs complement figures.
do_mosaics_rule_vs_popbool: Whether to display the mosaics of rule vs population figures.
do_legendbool: Whether to display the legend figure.

insightsolver.visualization.make_pdf(solver, output_file: str | None = None, verbose: bool = False, do_mutual_information: bool = True, do_banner: bool = True, do_contributions: bool = True, do_distributions: bool = True, do_mosaics_rule_vs_comp: bool = True, do_mosaics_rule_vs_pop: bool = True, do_legend: bool = True, language: str = 'en') → str#

Generates a PDF containing all visualization figures for the solver, using make_all() to generate figures.

Parameters#

solverInsightSolver: The fitted solver object.
output_filestr, optional: If provided, export the PDF at this path.
verbosebool: Verbosity.
do_mutual_informationbool: Include mutual information figure.
do_bannerbool: Include banner figures.
do_contributionsbool: Include contribution figures.
do_distributionsbool: Include distribution figures.
do_mosaics_rule_vs_compbool: Include mosaics of rule vs complement figures.
do_mosaics_rule_vs_popbool: Include mosaics of rule vs population figures.
do_legendbool: Include legend figure.
languagestr: Language for the plots (‘en’ or ‘fr’).

Returns#

pdf_base64str: PDF content encoded as base64 for in-memory use.

insightsolver.visualization.make_zip(solver, output_file: str | None = None, verbose: bool = False, do_png: bool = True, do_csv: bool = True, do_json: bool = True, do_excel: bool = True, do_pdf: bool = True, language: str = 'en') → str#

Export the solver content to a ZIP file.

Parameters#

solverInsightSolver: The fitted solver object.
output_filestr, optional: If provided, the ZIP will be saved to this path.
verbosebool: Whether to print progress messages.
do_pngbool: Include PNG figures.
do_csvbool: Include CSV ruleset.
do_jsonbool: Include JSON ruleset.
do_excelbool: Include Excel ruleset.
do_pdfbool: Include PDF of figures.
languagestr: Language for plots (‘en’ or ‘fr’).

Returns#

zip_base64str: ZIP content encoded as base64 for in-memory use.

visualization module

Contents

visualization module#

Description#

Naming conventions of the visualization functions#

Utilities#

Mutual information#

Banner#

Legend#

Feature contributions#

Feature distribution for feature#

Feature distributions for S#

Mosaic of rule vs complement for a feature for the rule i#

Mosaics of rule vs complement for the rule i#

Mosaic of rule vs pop vs complement for the rule i#

Mosaics of rule vs pop vs complement#

Complete plot#

Export to PDF#

Export to ZIP#

License#

Heuristic#

Parameters#

Returns#

Parameters#

Returns#

Parameters#

Returns#

Parameters#

Returns#

Parameters#

Returns#

Parameters#

Returns#

Parameters#

Parameters#

Returns#

Parameters#

Returns#

Parameters#

Returns#

Parameters#

Parameters#

Returns#

Parameters#

Parameters#

Returns#

Parameters#

Parameters#

Returns#

Parameters#

Parameters#

Returns#

Parameters#

Returns#

Parameters#

Parameters#

Parameters#

Returns#

Parameters#

Returns#

Parameters#

Parameters#

Returns#

Parameters#

Returns#

Parameters#

Parameters#

Returns#

Parameters#

Returns#

Parameters#

Parameters#

Returns#

Parameters#

Returns#

Parameters#

Parameters#

Returns#

Parameters#

Returns#