causalexplain.gui package

Subpackages

• causalexplain.gui.tabs package
  • Submodules
    • DiagnosticsTab
      • DiagnosticsTab.__init__()
      • DiagnosticsTab.build()
      • DiagnosticsTab.refresh_from_state()
    • GenerateTab
      • GenerateTab.__init__()
      • GenerateTab.build()
      • GenerateTab.run_generate()
      • GenerateTab.run_save_generated()
    • LoadTab
      • LoadTab.__init__()
      • LoadTab.build()
      • LoadTab.run_load()
    • TrainTab
      • TrainTab.__init__()
      • TrainTab.build()
      • TrainTab.run_training()
      • TrainTab.start_training_task()
      • TrainTab.cancel_training_task()
      • TrainTab.save_trained_model()
      • TrainTab.save_trained_dag()
  • Module contents
    • TrainTab
      • TrainTab.__init__()
      • TrainTab.build()
      • TrainTab.run_training()
      • TrainTab.start_training_task()
      • TrainTab.cancel_training_task()
      • TrainTab.save_trained_model()
      • TrainTab.save_trained_dag()
    • LoadTab
      • LoadTab.__init__()
      • LoadTab.build()
      • LoadTab.run_load()
    • GenerateTab
      • GenerateTab.__init__()
      • GenerateTab.build()
      • GenerateTab.run_generate()
      • GenerateTab.run_save_generated()
    • DiagnosticsTab
      • DiagnosticsTab.__init__()
      • DiagnosticsTab.build()
      • DiagnosticsTab.refresh_from_state()

Submodules

NiceGUI app for local causalexplain workflows.

has_active_model(active_model_state)

Return whether the GUI currently has a trained or loaded model.

run_gui(host='127.0.0.1', port=8080)

Start the NiceGUI application.

Cytoscape helpers for NiceGUI graph rendering.

ensure_cytoscape_assets()

Expose Cytoscape asset registration for GUI apps.

Helpers for shaping ReX diagnostics for GUI consumption.

ordered_regressors(regressors)

Return regressors in preferred GUI order.

normalize_rex_diagnostics(diagnostics_by_regressor)

Stack per-regressor diagnostics into GUI-ready dataframes.

regression_errors_for_all_targets(errors_long)

Return regression errors for all target variables.

shap_values_for_target(shap_long, target)

Filter and sort SHAP mean values for a selected target variable.

bootstrap_details_for_regressor(bootstrap_matrices, bootstrap_edges_long, regressor)

Return the matrix and edge table for the selected regressor.

Graph helper utilities for the GUI.

clean_node_name(name)

Normalize a node name by stripping whitespace and quotes.

normalize_graph(graph)

Return a cleaned directed graph with normalized node names.

dag_is_valid(graph, min_edges, max_edges)

Check that the graph is a DAG within the edge constraints.

graph_from_dot(path)

Load a DOT file into a normalized directed graph.

File and path utilities for the GUI.

ensure_file(path, suffixes)

Ensure the file exists and matches the expected suffixes.

ensure_output_dir(path)

Create the output directory for the given file path when needed.

normalize_output_value(raw_value, required_ext)

Normalize an output path by enforcing the required extension.

sanitize_output_name(name)

Strip extensions and whitespace from a dataset base name.

Rendering helpers for graphs, metrics, and diagnostics in the GUI.

update_metrics_log(log_el, metrics)

Render metrics into a NiceGUI log element.

overlay_status_message(classes)

Return the overlay status message based on edge classes.

render_cytoscape_overlay(container, status_label, discoverer, ref_graph, *, persist_positions, height='420px')

Render an interactive overlay chart for predicted vs. reference DAGs.

render_cytoscape_graph(container, graph, *, height='420px')

Render a DAG in a NiceGUI container using Cytoscape.

render_regression_error_chart(errors_frame)

Return EChart options for grouped regression errors across targets.

render_shap_mean_chart(shap_for_target, target)

Return EChart options for grouped SHAP mean values.

render_bootstrap_heatmap(bootstrap_matrix, regressor)

Return EChart options for a regressor-specific bootstrap heatmap.

Default settings and persistence helpers for the GUI.

default_train_settings()

Return default training settings for the GUI.

default_load_settings()

Return default load/evaluation settings for the GUI.

default_generate_settings()

Return default dataset generation settings for the GUI.

default_diagnostics_settings()

Return default settings for the diagnostics tab.

merge_settings(stored, defaults)

Merge persisted settings into the provided defaults.

Style helpers for the causalexplain GUI.

app_styles_path()

Return the absolute path to the GUI CSS file.

read_app_styles()

Read the GUI CSS content from disk.

register_app_styles()

Inject the GUI CSS into the NiceGUI head.

UI helper utilities for NiceGUI widgets.

update_settings(storage, settings_key, target, key, value)

Update a settings dict and persist it to NiceGUI storage.

bind_setting(element, storage, settings_key, settings_ref, field)

Bind UI change events to a settings dictionary.

set_input_value(input_el, value)

Update a NiceGUI input value and refresh the widget.

async save_upload(upload_event, upload_dir, suffix=None)

Persist an uploaded file to the GUI upload folder.

make_upload_handler(input_el, storage, settings_key, settings_ref, field, upload_dir, suffix=None, status_label=None)

Create an async upload handler that updates UI state.

normalize_output_path(input_el, storage, settings_key, settings_ref, field, required_ext)

Normalize a UI output path and persist the setting.

Module contents

GUI entry points for causalexplain.

run_gui(*args, **kwargs)

Lazy import to avoid circular dependencies when loading GUI helpers.