seismometer.data.loader.pipeline.SeismogramLoader

class seismometer.data.loader.pipeline.SeismogramLoader(config, prediction_fn=None, event_fn=None, post_predict_fn=None, post_event_fn=None, merge_fn=None, post_load_fn=None)

A data loading pipeline using three types of hooks:

  • load predictions [ConfigOnlyHook]

  • transform (type) the loaded predictions [ConfigFrameHook]

  • load events [ConfigOnlyHook]

  • transform (type) the loaded events [ConfigFrameHook]

  • merge events onto predictions [MergeFramesHook]

  • post load manipulations [ConfigFrameHook]

Each step is expected to return a dataframe, chaining the steps to get the frame driving a loaded Seismogram.

Parameters:
__init__(config, prediction_fn=None, event_fn=None, post_predict_fn=None, post_event_fn=None, merge_fn=None, post_load_fn=None)

Initialize a data loading pipeline of functions returning a dataframe for a Seismogram session.

Parameters:

  • config (ConfigProvider) – The loaded configuration object.

  • prediction_fn (ConfigOnlyHook) – A callable taking a ConfigProvider and returning a dataframe. Used to load a (predictions) dataframe based on configuration; skipped if prediction_obj is provided to load_data.

  • event_fn (ConfigOnlyHook, optional) – A callable taking a ConfigProvider and returning a dataframe. Used to load a (events) dataframe based on configuration; skipped if event_obj is provided to load_data.

  • post_predict_fn (ConfigFrameHook, optional) – A callable taking a ConfigProvider and a (predictions) dataframe and returning a dataframe. Used to do minor transforms of predictions such as type casting.

  • post_event_fn (ConfigFrameHook, optional) – A callable taking a ConfigProvider and a (events) dataframe and returning a dataframe. Used to do minor transforms of events such as type casting.

  • merge_fn (MergeFramesHook, optional) – A callable taking a ConfigProvider, a (events) dataframe, and a (predictions) dataframe and returning a dataframe. Used to merge events onto predictions based on configuration.

  • post_load_fn (ConfigFrameHook, optional) – A callable taking a ConfigProvider and the fully loaded dataframe and returning a dataframe. Used to allow any custom manipulations of the Seismogram dataframe during load. WARNING: This can completly overwrite/discard the daframe that was loaded.

Methods

__init__(config[, prediction_fn, event_fn, ...])

Initialize a data loading pipeline of functions returning a dataframe for a Seismogram session.

load_data([prediction_obj, event_obj])

Entry point for loading data for a Seismogram session.