atlas.active_learning.dashboard package

Flask monitoring dashboard for active learning loops.

Submodules

atlas.active_learning.dashboard.training_dashboard_flask module

Dashboard for monitoring and reviewing training Active Learning WorkChains.

atlas.active_learning.dashboard.training_dashboard_flask.create_cache(workchain_node_id)

Initializes a new, empty cache DataFrame.

atlas.active_learning.dashboard.training_dashboard_flask.gather_information(workchain_node_id, app)

Main function to gather all dashboard information, using a cache file.

atlas.active_learning.dashboard.training_dashboard_flask.get_complete_steps_uuid(node)

Retrieves UUIDs and details of completed steps in a workchain.

atlas.active_learning.dashboard.training_dashboard_flask.get_missing_cache_steps_uuid(node_pk, cache: DataFrame)

Identifies workchain steps that are not yet in the cache.

atlas.active_learning.dashboard.training_dashboard_flask.get_model_stats(node)

Extracts model training statistics (RMSE) if available.

atlas.active_learning.dashboard.training_dashboard_flask.get_progbar_class_name(node)

Determines the CSS class for a progress bar based on node state.

atlas.active_learning.dashboard.training_dashboard_flask.get_report(node)

Parses and styles the report from an AiiDA workchain node.

atlas.active_learning.dashboard.training_dashboard_flask.get_step_child_info(node)

Gathers information about the direct children of a given node.

atlas.active_learning.dashboard.training_dashboard_flask.run_training_dashboard(workchain_node_id, refresh_interval=60, port=8000)

Sets up and runs the Flask application for the dashboard.

atlas.active_learning.dashboard.training_dashboard_flask.update_cache(cache: DataFrame, missing_uuid_list: list, db_size_dict: dict) → DataFrame

Updates the cache DataFrame with information from new UUIDs.

Module contents

Flask monitoring dashboard for active learning loop progress.