aPriori.NN.TrainingBuilder¶

class aPriori.NN.TrainingBuilder(*args, **kwargs)¶

Bases: dict

Description:¶

A dictionary-like class that builds and manages a collection of VectorScaler instances for training purposes.

This class includes methods to add, configure, and save/load multiple VectorScaler objects, and allows batch scaling of data based on a specified Field3D input.

Attributes:¶

  • state_dict (dict): A dictionary containing the state of all scalers in the TrainingBuilder.

Methods:¶

  • __init__(*args, **kwargs):

    Initialize the TrainingBuilder with optional dictionary arguments.

  • get_subset(keys):

    Return a subset of the TrainingBuilder based on the specified keys.

  • __setitem__(key, value):

    Set an item in the TrainingBuilder with key and value type validation.

  • add(variable, *args, **kwargs):

    Add a new VectorScaler to the TrainingBuilder.

  • build_x(field):

    Construct a feature matrix by transforming data from a Field3D instance.

  • fit(field):

    Fit each VectorScaler in the TrainingBuilder to the corresponding data in the Field3D.

  • load(file_path):

    Load the state of each scaler from a JSON file and reinitialize the TrainingBuilder.

  • save(file_path):

    Save the current state of the TrainingBuilder to a JSON file.

  • state_dict:

    Return the current state of the TrainingBuilder, including all scaler states.

  • __str__():

    Return a string representation of the TrainingBuilder.

get_subset(keys)¶

Return a subset of the TrainingBuilder based on the specified keys.

Parameters:¶

  • keys (list):

    The keys to include in the subset.

Returns:¶

:
  • TrainingBuilder:

    A new TrainingBuilder containing only the specified keys.

add(variable, *args, **kwargs)¶

Add a new VectorScaler to the TrainingBuilder.

Parameters:¶

  • variable (str):

    The key to associate with the VectorScaler.

  • *args:

    Arguments to pass to the VectorScaler initializer.

  • **kwargs:

    Keyword arguments to pass to the VectorScaler initializer.

Raises:¶

  • KeyError:

    If the specified variable already exists in the TrainingBuilder.

build_x(field)¶

Construct a feature matrix by transforming data from a Field3D instance.

Parameters:¶

  • field (Field3D): The input field containing data to scale.

Returns:¶

:

  • np.ndarray: A 2D array containing the scaled data for all variables in the TrainingBuilder.

Raises:¶

  • KeyError: If the field is not an instance of Field3D.

fit(field)¶

Fit each VectorScaler in the TrainingBuilder to the corresponding data in the Field3D or list of Field3D objects.

Parameters:¶

  • field (Field3D or list of Field3D): The field(s) with data to fit each scaler.

load(file_path)¶

Load the state of a previously saved object from a JSON file and reinitialize the TrainingBuilder.

Parameters:¶

  • file_path (str): The path to the JSON file to load.

Raises:¶

  • ValueError: If the file is not a valid JSON file.

save(file_path)¶

Save the current state of the TrainingBuilder to a JSON file.

Parameters:¶

  • file_path (str): The path to save the JSON file.

property state_dict¶

Return the current state of the TrainingBuilder, including all scaler states.

Returns:¶

  • dict: A dictionary representing the state of all scalers in the TrainingBuilder.