Plotting utilities

pyballmapper.plotting.generate_partitions(node_contents: dict[Any, list[Any]], original_element_values: dict[Any, Any]) → dict[Any, dict[Any, int]]

Given collections X_i of sets Y_ij and a function U Y_ij -> Z, computes for each i the count of elements in X_i that map to a given value in Z.

node_contents:

dict of lists such that node_contents[node] is a list of all elements contained in node.

original_element_values:

dict of values associated to each of the original set of elements. Not necessarily numeric.

class pyballmapper.plotting.graph_GUI(graph: networkx.Graph, my_palette: matplotlib.colors.Colormap, tooltips_variables: list[str] | None = None, figsize: tuple[int, int] = (800, 600), output_format: str = 'svg', render_seed: int = 42, render_iterations: int = 100, MIN_SIZE: int = 7, MAX_SIZE: int = 20)

Bases: object

Create a Bokeh plot rapresenting the BallMapper graph.

Parameters:

  • graph (NetworkX graph) – The Graph attribute of a BallMapper object.

  • my_palette (matplotlib palette) – The color palette used to color the nodes.

  • tooltips_variables (list of strings, default=None) – Which variables to show in the mouse hovertool.

  • figsize (tuple, default=(800, 600)) – The figure size.

  • output_format (string, default='svg') – The format in which the save button saves the plot.

  • render_seed (int, default=42) – Seed for the networkx spring_layout.

  • render_iterations (int, default=100) – Number of iterations for the networkx spring_layout.

  • MIN_SIZE (int, default=7) – Minimal size of the graph nodes.

  • int (MAX_SIZE;) – Maximal size of the graph nodes.

  • default=20 – Maximal size of the graph nodes.

Graph

The BallMapper graph. Each node correspond to a covering ball and has attributes: ‘landmark’ the id of the corresponding landmark point ‘points covered’ the ids of the points covered by the corresponding ball

Type:

NetworkX Graph object

eps

The input radius of the balls.

Type:

float

points_covered_by_landmarks

keys: landmarks ids values: list of ids of the points covered by the corresponding ball

Type:

dict

add_colorbar(num_ticks: int, low: float, high: float) → None

Add a colorbar to the right side of the Bokeh plot.

Parameters:

  • num_ticks (int) – Number of ticks, use a high number (100) for a continuos looking colorbar.

  • low (float) – Value at the bottom of the colorbar.

  • high (float) – Value at the top of the colorbar.

color_by_variable(variable: str, MIN_VALUE: float = inf, MAX_VALUE: float = -inf, logscale: bool = False) → tuple[float, float]

Color the BallMapper Bokeh plot nodes using a specific variable.

Coloring data can be added using the BallMapper.add_coloring() method. MIN_VALUE and MAX_VALUE are computed automatically, but lower (resp. higher) values can be manually specified.

Parameters:

  • variable (string) – Which coloring variable to use.

  • MIN_VALUE (float, default=np.inf) – Minimum value for the coloring palette.

  • MAX_VALUE (float, default=-np.inf) – Maximum value for the coloring palette.

  • logscale (bool, default=False) – Wheter to use a logscale coloring. If False, linear is used.

color_edges() → None

color edges by interpolating between the nodes’ colors

pyballmapper.plotting.init_bokeh_figure(g: networkx.Graph, background_fill_color: str, title: str | None, match_aspect: bool, edge_width: int, edge_color: str, plot_height: int, plot_width: int, sizing_mode: str, outline_line_width: int, outline_line_color: str) → tuple[Any, list[float], list[float]]
pyballmapper.plotting.kmapper_visualize(bm: BallMapper, coloring_df: pandas.DataFrame, path_html: str = 'output.html', title: str | None = None, **kwargs: Any) → None

leverages kepler-mapper visualization tool to produce an interactive html. https://kepler-mapper.scikit-tda.org

Parameters:

  • bm (BallMapper) – the BallMapper graph to plot

  • coloring_df (pandas.DataFrame) – pandas dataframe of shape (n_samples, n_coloring_function)

  • path_html (str, optional) – the output file, by default ‘output.html’

  • title (string, optional) – title to be displayed in the top bar, by default None

pyballmapper.plotting.pie_graph_plot(partitions: dict[Any, dict[Any, float]], g: nx.Graph | None = None, nodes: list[Any] | None = None, edges: list[tuple[Any, Any]] | None = None, graph_layout: dict[Any, tuple[float, float]] | None = None, radius: float = 0.01, node_labels: list[str] | None = None, node_scaling: str = 'old', palette: dict[Any, str] | list[str] | None = None, background_fill_color: str = 'white', title: str | None = None, match_aspect: bool = True, edge_width: int = 1, edge_color: str = 'black', plot_height: int = 600, plot_width: int = 600, sizing_mode: str = 'fixed', show_node_labels: bool = False, outline_line_width: int = 1, outline_line_color: str = 'black') → Any

Produces a bokeh plot of a graph where each node is represented by a pie chart.

Parameters:

  • g – networkx graph

  • partitions – dict of dicts such that partitions[node][class] == portion of class at node.

  • graph_layout – dict of size 2 arrays such that graph_layout[node] == position of node in plane.

  • nodes – list of nodes. If used, parameter g is ignored.

  • edges – list of edges. Used in conjunction with nodes. If used, parameter g is ignored.

  • radius – Radius of nodes. Upper bound on the radii of nodes if node_scaling is set to True.

  • node_labels – list of text labels for nodes. If provided, must be in same order as iteration over nodes in graph.

  • node_scaling –

    If set to True, nodes are scaled from

    ((smallest partition size)/(largest partition size))*radius to radius, with

    a minimum size of nodes of radius/10 for enhanced viewability.

  • palette – dict with keys nodes and values bokeh colors, or list with colors with same size as number of classes present in partitions.

  • title – Plot’s title.

  • edge_width – Graph’s edge thickness.

  • edge_color – Graph’s edge color. Used only if sizing_mode is “fixed”.

  • show_node_labels – If set to True, adds node labels to the plot.

  • match_aspect – Preserves aspect ratio when changing size. Passed as-is to bokeh’s figure initializer.

  • plot_height – Plot’s height in pixels. Used only if sizing_mode is “fixed”. Passed as-is to bokeh’s figure initializer.

  • plot_width – Plot’s width in pixels. Passed as-is to bokeh’s figure initializer.

  • sizing_mode – One of “fixed”, “stretch_both”, “stretch_width”, “stretch_height”, “scale_width”, “scale_height”. Passed as-is to bokeh’s figure initializer.

  • outline_line_width – Thickness of the frame of the plot. Passed as-is to bokeh’s figure initializer.

  • outline_line_color – Color of the frame of the plot. Passed as-is to bokeh’s figure initializer.

  • background_fill_color – bokeh color to use for plot background. Passed as-is to bokeh’s figure initializer.