Causality Base Class

The causality module provides a base class for causality analysis.

class macrostat.causality.CausalityAnalyzer(model_class: Type[Model])

Bases: object

analyze()

Analyze a model class and return dependency dictionary

build_adjacency_matrix() → DataFrame

Build adjacency matrix from dependencies

check_for_cycles()

Check for cycles in the model’s dependency graph.

While cycles are not necessarily a problem, cycling logic in a given step may indicate indeterminism. This method returns a list of cycles found in the created adjacency matrix.

Returns:

List of cycles found in the graph. Each cycle is a list of nodes. Returns empty list if no cycles are found.

Return type:

list

plot_heatmap()

Visualize the adjacency matrix organized by trophic levels.

Parameters:

• figsize (tuple, optional) – Figure size in inches (width, height), by default (12, 10)

• cmap (str, optional) – Colormap to use, by default ‘viridis’

Returns:

The figure object containing the plot

Return type:

matplotlib.figure.Figure

plot_with_cytoscape(port: int = 8050, node_styles: dict[str, dict[str, str]] = {'history': {'background-color': 'lightgray', 'border-color': 'black'}, 'hyper': {'background-color': 'lavender', 'border-color': 'purple'}, 'parameters': {'background-color': 'lightgreen', 'border-color': 'green'}, 'prior': {'background-color': 'lightpink', 'border-color': 'red'}, 'scenario': {'background-color': 'lightyellow', 'border-color': 'yellow'}, 'state': {'background-color': 'lightblue', 'border-color': 'blue'}})

Create an interactive flowchart visualization using Dash and Cytoscape.

This method creates a web-based interactive visualization of the model’s structure using Dash and Cytoscape. The visualization allows for: - Zooming and panning - Node selection and highlighting - Edge highlighting - Node dragging and repositioning - Export to various formats

Parameters:

port (int, optional) – The port number to run the Dash server on, by default 8050

Returns:

Opens a web browser with the interactive visualization

Return type:

None