Optimization pipeline — napari_easytrack.analysis.optim_pipeline

This module contains functions for cell tracking using Bayesian Optimization with timeout handling.

napari_easytrack.analysis.optim_pipeline.add_config_params_to_dict(params_dict, config_file_path, dataset_name)

Add default parameters from the config file to the dictionary.

Parameters:

  • params_dict (dict) – Dictionary to add parameters to.

  • config_file_path (str) – Path to the config file.

  • dataset_name (str) – Name of the dataset.

Returns:

Updated dictionary with added parameters.

Return type:

dict

napari_easytrack.analysis.optim_pipeline.add_missing_attributes(graph)

Add attributes to nodes without any attributes, essentially inserting dummy nodes where necessary.

Parameters:

graph (networkx.Graph) – The graph to update.

napari_easytrack.analysis.optim_pipeline.calculate_accuracy(lbep, segm, gt_data)

Calculate the accuracy of the tracking algorithm.

Parameters:

  • lbep – LBEP data from tracking.

  • segm – Segmentation data.

  • gt_data – Ground truth data.

Returns:

Dictionary containing accuracy metrics.

Return type:

dict

napari_easytrack.analysis.optim_pipeline.optimize_dataset(dataset, gt_data, objectives, study_name, n_trials=64, use_parallel_backend=True, sampler='tpe', search_space=None)

Optimize a dataset without timeout for each trial.

Parameters:

  • dataset – Dataset to be optimized.

  • gt_data – Ground truth data.

  • objectives (str) – Objectives type, ‘1obj’ or ‘2obj’.

  • study_name (str) – Name of the Optuna study.

  • n_trials (int, optional) – Number of trials to run. Defaults to 64.

  • use_parallel_backend (bool, optional) – Whether to use parallel backend. Defaults to True.

  • sampler (str, optional) – Sampler to use for the optimization. Defaults to ‘tpe’.

  • search_space (dict, optional) – Search space for grid sampling.

Returns:

The Optuna study object.

Return type:

optuna.study.Study

napari_easytrack.analysis.optim_pipeline.optimize_dataset_with_timeout(dataset, gt_data, objectives, study_name, n_trials=64, timeout=300, timeout_penalty=10000, use_parallel_backend=True, sampler='tpe', search_space=None)

Optimize a dataset with a timeout for each trial.

Parameters:

  • dataset – Dataset to be optimized.

  • gt_data – Ground truth data.

  • objectives (str) – Objectives type, ‘1obj’ or ‘2obj’.

  • study_name (str) – Name of the Optuna study.

  • n_trials (int, optional) – Number of trials to run. Defaults to 64.

  • timeout (int, optional) – Timeout for each trial in seconds. Defaults to 300.

  • timeout_penalty (float, optional) – Penalty value for timeouts. Defaults to 10000.

  • use_parallel_backend (bool, optional) – Whether to use parallel backend. Defaults to True.

  • sampler (str, optional) – Sampler to use for the optimization. Defaults to ‘tpe’.

  • search_space (dict, optional) – Search space for grid sampling.

Returns:

The Optuna study object.

Return type:

optuna.study.Study

napari_easytrack.analysis.optim_pipeline.read_config_params(config_file_path)

Read parameters from a config file.

Parameters:

config_file_path (str) – Path to the config file.

Returns:

Dictionary containing the config parameters.

Return type:

dict

napari_easytrack.analysis.optim_pipeline.run_cell_tracking_algorithm(objects, config, volume, max_search_radius=100, use_napari=False)

Run the cell tracking algorithm.

Parameters:

  • objects – Objects to track.

  • config – Tracking configuration.

  • volume – Volume of the tracking space.

  • max_search_radius (int, optional) – Maximum search radius. Defaults to 100.

  • use_napari (bool, optional) – Whether to use napari for visualization. Defaults to False.

Returns:

Tracking results including LBEP and tracks, and napari data if use_napari is True.

Return type:

tuple

napari_easytrack.analysis.optim_pipeline.run_with_timeout(func, args, kwargs, timeout, objectives, sem, timeout_penalty)

Run a function with a timeout and handle termination if it exceeds the timeout.

Parameters:

  • func (callable) – Function to be executed.

  • args (tuple) – Arguments to pass to the function.

  • kwargs (dict) – Keyword arguments to pass to the function.

  • timeout (int) – Timeout in seconds.

  • objectives (str) – Objectives type, ‘1obj’ or ‘2obj’.

  • sem (multiprocessing.Semaphore) – Semaphore to control concurrent execution.

  • timeout_penalty (float) – Penalty value for timeouts.

Returns:

Result of the function execution or a default value if timed out.

Return type:

result

napari_easytrack.analysis.optim_pipeline.write_best_params_to_config(params, config_file_path)

Write the best parameters to the config file in JSON format.

Parameters:

  • params (dict) – Dictionary containing the best parameters.

  • config_file_path (str) – Path to the config file where parameters will be written.