Slicers

class sekupy.preprocessing.slicers.DatasetMasker(mask=None, **kwargs)[source]
transform(ds, axis=0)[source]

Transform the provided dataset.

This method applies the transformation to the dataset and records the transformation in the dataset’s preprocessing history.

Parameters:

ds (Dataset) – The dataset to transform

Returns:

The transformed dataset

Return type:

Dataset

class sekupy.preprocessing.slicers.FeatureExpressionSlicer(fx=<ufunc 'greater'>)[source]
transform(ds)[source]

[summary]

Parameters:
  • ds (pymvpa dataset) – The dataset to be used

  • value (int or fx, optional) – The function used to generate a value to be compared with the attribute using the compare_fx funtion, by default lambdax:np.mean(x)+1.5*np.std(x)

Returns:

ds – The sliced dataset

Return type:

pymvpa dataset

class sekupy.preprocessing.slicers.FeatureSlicer(**kwargs)[source]

This transformer filters the dataset using features as specified on a dictionary The dictionary indicates the feature attributes to be used as key and a list with conditions to be selected:

selection_dict = {

‘accuracy’: [‘I’], ‘frame’:[1,2,3] }

This dictionary means that we will select all features with frame attribute equal to 1 OR 2 OR 3 AND all samples with accuracy equal to ‘I’.

transform(ds)[source]

Transform the provided dataset.

This method applies the transformation to the dataset and records the transformation in the dataset’s preprocessing history.

Parameters:

ds (Dataset) – The dataset to transform

Returns:

The transformed dataset

Return type:

Dataset

class sekupy.preprocessing.slicers.SampleExpressionSlicer(attr, compare_fx=<ufunc 'greater'>, attr_transformer=None)[source]
transform(ds, value=<function SampleExpressionSlicer.<lambda>>)[source]

[summary]

Parameters:
  • ds (pymvpa dataset) – The dataset to be used

  • value (int or fx, optional) – The function used to generate a value to be compared with the attribute using the compare_fx funtion, by default lambdax:np.mean(x)+1.5*np.std(x)

Returns:

ds – The sliced dataset

Return type:

pymvpa dataset

class sekupy.preprocessing.slicers.SampleSlicer(**kwargs)[source]

Selects only portions of the dataset based on a dictionary The dictionary indicates the sample attributes to be used as key and a list with conditions to be selected:

selection_dict = {

‘frame’: [1,2,3] }

This dictionary means that we will select all samples with frame attribute equal to 1 OR 2 OR 3 AND all samples with accuracy equal to ‘I’.

transform(ds)[source]

Transform the provided dataset.

This method applies the transformation to the dataset and records the transformation in the dataset’s preprocessing history.

Parameters:

ds (Dataset) – The dataset to transform

Returns:

The transformed dataset

Return type:

Dataset