piex.explorer module¶

class piex.explorer.MongoPipelineExplorer(db, *args, **kwargs)[source]¶

Bases: piex.explorer.PipelineExplorer

get_best_pipeline(dataset, **filters)[source]¶: Get information about the best pipeline ever found for the given dataset.

get_dataset_id(dataset)[source]¶: Get the TRAIN dataset id for a given dataset.

get_datasets(**filters)[source]¶: Get a pandas DataFrame with information about the available datasets.

get_pipelines(**filters)[source]¶: Get a pandas DataFrame with information about the scored pipelines.

get_templates(**filters)[source]¶: Get a pandas DataFrame with information about the available templates.

get_test_results(**filters)[source]¶: Get a pandas DataFrame with the results of the executed tests.

get_tests(**filters)[source]¶: Get a pandas DataFrame with information about the executed tests.

load_pipeline(pipeline_id)[source]¶: Get the dict representation of the pipeline.

class piex.explorer.PipelineExplorer(data_path='data')[source]¶

Bases: object

DATASETS_COLUMNS = ['dataset', 'data_modality', 'task_type', 'task_subtype']¶

get_best_pipeline(dataset, **filters)[source]¶: Get information about the best pipeline ever found for the given dataset.

get_best_template(dataset, **filters)[source]¶: Get the name of the template that got the best score for this dataset.

get_dataset_id(dataset)[source]¶: Get the TRAIN dataset id for a given dataset.

get_datasets(**filters)[source]¶: Get a pandas DataFrame with information about the available datasets.

get_default_hyperparameters(template_name)[source]¶: Get the default hyperparmeters of the given template.

get_pipelines(**filters)[source]¶: Get a pandas DataFrame with information about the scored pipelines.

get_templates(**filters)[source]¶: Get a pandas DataFrame with information about the available templates.

get_test_results(**filters)[source]¶: Get a pandas DataFrame with the results of the executed tests.

get_tests(**filters)[source]¶: Get a pandas DataFrame with information about the executed tests.

get_tunable_hyperparameters(template_name)[source]¶: Get the tunable hyperparmeters of the given template.

load_best_pipeline(dataset, **filters)[source]¶: Get the dict representation of the best pipeline ever found for this dataset.

load_best_template(dataset, **filters)[source]¶: Get the dict representation of the best template ever found for this dataset.

load_pipeline(pipeline_id)[source]¶: Get the dict representation of the pipeline.

load_template(template_name)[source]¶: Get the dict representation of the template.

score_pipeline(pipeline_id, dataset, n_splits=5, cv=None, random_state=0)[source]¶: Cross validate the given pipeline on this dataset.

score_template(template_name, dataset, hyperparameters=None, n_splits=5, cv=None, random_state=0)[source]¶: Cross validate the given template on this dataset.

class piex.explorer.S3PipelineExplorer(bucket, *args, **kwargs)[source]¶

Bases: piex.explorer.PipelineExplorer

property client¶

get_best_pipeline(dataset, **filters)[source]¶: Get information about the best pipeline ever found for the given dataset.

get_dataset_id(dataset)[source]¶: Get the TRAIN dataset id for a given dataset.

get_datasets(**filters)[source]¶: Get a pandas DataFrame with information about the available datasets.

get_pipelines(**filters)[source]¶: Get a pandas DataFrame with information about the scored pipelines.

get_templates(**filters)[source]¶: Get a pandas DataFrame with information about the available templates.

get_test_results(**filters)[source]¶: Get a pandas DataFrame with the results of the executed tests.

get_tests(**filters)[source]¶: Get a pandas DataFrame with information about the executed tests.

load_pipeline(pipeline_id)[source]¶: Get the dict representation of the pipeline.