rise_text
=========

.. py:module:: rise_text


Classes
-------

.. autoapisummary::

   rise_text.RISEText


Module Contents
---------------

.. py:class:: RISEText(n_masks=1000, feature_res=8, p_keep=None, preprocess_function=None)

   RISE implementation for text based on https://github.com/eclique/RISE/blob/master/Easy_start.ipynb.


   .. py:attribute:: n_masks


   .. py:attribute:: feature_res


   .. py:attribute:: p_keep


   .. py:attribute:: preprocess_function


   .. py:attribute:: masks
      :value: None



   .. py:attribute:: predictions
      :value: None



   .. py:method:: explain(model_or_function, input_text, labels, tokenizer=None, batch_size=100)

      Runs the RISE explainer on text.

         The model will be called with masked versions of the input text.

      :param model_or_function: The function that runs the model to be explained _or_
                                the path to a ONNX model on disk.
      :type model_or_function: callable or str
      :param input_text: Text to be explained
      :type input_text: np.ndarray
      :param tokenizer: Tokenizer class with tokenize and convert_tokens_to_string methods, and mask_token attribute
      :param labels: Labels to be explained
      :type labels: Iterable(int)
      :param batch_size: Batch size to use for running the model.
      :type batch_size: int

      :returns: List of tuples (word, index of word in raw text, importance for target class) for each class.



   .. py:method:: _determine_p_keep(input_text, runner, tokenizer, n_masks, batch_size)

      See n_mask default value https://github.com/dianna-ai/dianna/issues/24#issuecomment-1000152233.



   .. py:method:: _calculate_mean_class_std(p_keep, runner, input_text, tokenizer, n_masks, batch_size)


   .. py:method:: _generate_masks(input_shape, p_keep, n_masks)


   .. py:method:: _get_saliencies(runner, sentences, num_tokens, batch_size, p_keep)


   .. py:method:: _reshape_result(input_tokens, labels, saliencies)
      :staticmethod:



   .. py:method:: _create_masked_sentences(tokens, masks, tokenizer)