TransformableDataBlock

class aixd.data.TransformableDataBlock(name: str, dobj_list: List[DataObject], **kwargs)[source]

Bases: DataBlock

A special data block that is transformable. Does not contain an uid column, and holds no data, it only captures the structure of the data. It allows to apply transformation to the data. In particular, it saves the transformed data objects for the inverse transformation.

Parameters:
  • name (str) – The name of the data block.

  • dobj_list (List[DataObject]) – A list of data objects.

  • kwargs – Arguments passed to DataBlock.__init__(…).

Methods

get_dobj_dimensions

Returns the dimensions of the data objects.

inverse_transform

Inverse transforms the data matrix.

to_data_block

Converts a transformable data block to back to an ordinary data block.

transform

Transforms the data matrix.

transformation_is_fitted

Returns True if the normalization is fitted.

Inherited Methods

check_data_consistency

Checks the data consistency.

get_cols_dobjs

In this case, the attributes can be either names of dimensions, or names of full data objects, and we need to return the precise column names that correspond to the attributes.

get_data_mat

Returns the data belonging to the requested DataObject's, as a single data matrix.

get_data_mat_dtypes

Returns the data types of the columns of the data matrix for the requested DataObject's.

get_data_mats

Returns the data belonging to the requested DataObject's, as a list of data matrices.

get_dobjs

Returns the data objects.

info_dobj

Returns information of all the DataObject's of the DataBlock.

print_objects

Prints the names of the DataObject's.

print_transf

Prints the applied transformations for the DataBlock.

read

Internal function for saving.

update_obj

Updates the DataObject instance according to the passed data (e.g., changing its domain or type)

update_position_indices

Updates the position and position_index attributes of the DataObject's.

write

Internal function for writing.