synthval.metrics.PRScores
=========================

.. py:class:: synthval.metrics.PRScores(row_batch_size = 25000, col_batch_size = 50000, num_nearest_n = 3)

   Bases: :py:obj:`SimilarityMetric`


   A Similarity Metric class that computes the Precision and Recall scores between two distributions
   (real_dist and synth_dist).

   .. attribute:: row_batch_size

      Size of the row batches used when computing pairwise distances. This provides a trade-off between memory
      usage and performance (default: 25000).

      :type: int, optional

   .. attribute:: col_batch_size

      Size of the column batches used for computing pairwise distances (default: 50000).

      :type: int, optional

   .. attribute:: num_nearest_n

      Number of nearest neighbors used to estimate the manifold. The manifold is used for computing precision
      and recall (default: 2).

      :type: int, optional

   .. rubric:: References

   Tuomas Kynkäänniemi, Tero Karras, Samuli Laine, Jaakko Lehtinen, Timo Aila - Improved Precision and Recall
   Metric for Assessing Generative Models - Annual Conference on Neural Information Processing Systems, 2019.


   .. py:method:: calculate(real_dist_df, synth_dist_df)

      Compute Precision and Recall metrics between two distributions.

      :param real_dist_df: DataFrame containing samples from distribution real_dist.
      :type real_dist_df: pandas.DataFrame
      :param synth_dist_df: DataFrame containing samples from distribution synth_dist.
      :type synth_dist_df: pandas.DataFrame

      :returns: A numpy array containing the Precision and the Recall metrics between distribution real_dist and synth_dist.
      :rtype: numpy.ndarray



   .. py:method:: __compute_distances(row_features, col_features, col_batch_size)
      :staticmethod:


      Compute pairwise distances between row and column features using batches to optimize memory usage.

      :param row_features: Tensor of feature vectors representing the rows.
      :type row_features: torch.Tensor
      :param col_features: Tensor of feature vectors representing the columns.
      :type col_features: torch.Tensor
      :param col_batch_size: Size of the column batches for computing distances.
      :type col_batch_size: int

      :returns: **dist_batches** -- Tensor of pairwise distances.
      :rtype: torch.Tensor