Adding dummy class and test, correcting some docstrings

Author: baraline

aeon/similarity_search/_dummy.py

@@ -0,0 +1,93 @@
+"""Dummy similarity seach estimator."""
+
+__author__ = ["baraline"]
+__all__ = ["DummySimilaritySearch"]
+
+
+from aeon.similarity_search.base import BaseSimiliaritySearch
+
+
+class DummySimilaritySearch(BaseSimiliaritySearch):
+    """
+    DummySimilaritySearch for testing of the BaseSimiliaritySearch class.
+
+    Parameters
+    ----------
+    distance : str, default ="euclidean"
+        Name of the distance function to use.
+    normalize : bool, default = False
+        Whether the distance function should be z-normalized.
+    store_distance_profile : bool, default = =False.
+        Whether to store the computed distance profile in the attribute
+        "_distance_profile" after calling the predict method.
+    """
+
+    def __init__(
+        self, distance="euclidean", normalize=False, store_distance_profile=False
+    ):
+        super(DummySimilaritySearch, self).__init__(
+            distance=distance,
+            normalize=normalize,
+            store_distance_profile=store_distance_profile,
+        )
+
+    def _fit(self, X, y):
+        """
+        Private fit method, does nothing more than the base class.
+
+        Parameters
+        ----------
+        X : array, shape (n_instances, n_channels, n_timestamps)
+            Input array to used as database for the similarity search
+        y : optional
+            Not used.
+
+        Returns
+        -------
+        self
+
+        """
+        return self
+
+    def _predict(self, q, mask):
+        """
+        Private predict method for DummySimilaritySearch.
+
+        It compute the distance profiles and then returns the best match
+
+        Parameters
+        ----------
+        q :  array, shape (n_channels, q_length)
+            Input query used for similarity search.
+        mask : array, shape (n_instances, n_channels, n_timestamps - (q_length - 1))
+            Boolean mask of the shape of the distance profile indicating for which part
+            of it the distance should be computed.
+
+        Returns
+        -------
+        array
+            An array containing the index of the best match between q and _X.
+
+        """
+        if self.normalize:
+            distance_profile = self.distance_profile_function(
+                self._X,
+                q,
+                mask,
+                self._X_means,
+                self._X_stds,
+                self._q_means,
+                self._q_stds,
+            )
+        else:
+            distance_profile = self.distance_profile_function(self._X, q, mask)
+
+        if self.store_distance_profile:
+            self._distance_profile = distance_profile
+
+        # For now, deal with the multidimensional case as "dependent", so we sum.
+        search_size = distance_profile.shape[-1]
+        distance_profile = distance_profile.sum(axis=1)
+        _id_best = distance_profile.argmin(axis=None)
+
+        return [(_id_best // search_size, _id_best % search_size)]

aeon/similarity_search/base.py

@@ -19,7 +19,7 @@
 class BaseSimiliaritySearch(BaseEstimator, ABC):
     """BaseSimilaritySearch.
 
-    Attributes
+    Parameters
     ----------
     distance : str, default ="euclidean"
         Name of the distance function to use.
@@ -28,6 +28,15 @@ class BaseSimiliaritySearch(BaseEstimator, ABC):
     store_distance_profile : bool, default = False.
         Whether to store the computed distance profile in the attribute
         "_distance_profile" after calling the predict method.
+
+    Attributes
+    ----------
+    _X : array, shape (n_instances, n_channels, n_timestamps)
+        The input time series stored during the fit method.
+    distance_profile_function : function
+        The function used to compute the distance profile affected
+        during the fit method based on the distance and normalize
+        parameters.
     """
 
     _tags = {

aeon/similarity_search/tests/test_dummy.py

@@ -0,0 +1,36 @@
+"""Tests for DummySimilaritySearch."""
+
+__author__ = ["baraline"]
+
+
+import numpy as np
+import pytest
+from numpy.testing import assert_array_equal
+
+from aeon.similarity_search._dummy import DummySimilaritySearch
+
+DATATYPES = ["int64", "float64"]
+
+
+@pytest.mark.parametrize("dtype", DATATYPES)
+def test_DummySimilaritySearch(dtype):
+    X = np.asarray(
+        [[[1, 2, 3, 4, 5, 6, 7, 8]], [[1, 2, 4, 4, 5, 6, 5, 4]]], dtype=dtype
+    )
+    q = np.asarray([[3, 4, 5]], dtype=dtype)
+
+    search = DummySimilaritySearch()
+    search.fit(X)
+    idx = search.predict(q)
+    assert_array_equal(idx, [(0, 2)])
+
+    search = DummySimilaritySearch(normalize=True)
+    search.fit(X)
+    q = np.asarray([[8, 8, 10]], dtype=dtype)
+    idx = search.predict(q)
+    assert_array_equal(idx, [(1, 2)])
+
+    search = DummySimilaritySearch(normalize=True)
+    search.fit(X)
+    idx = search.predict(q, q_index=(1, 2))
+    assert_array_equal(idx, [(1, 0)])

aeon/similarity_search/tests/test_top_k_similarity.py

@@ -1,9 +1,6 @@
-"""
-Created on Sat Sep  9 14:12:58 2023
-
-@author: antoi
-"""
+"""Tests for TopKSimilaritySearch."""
 
+__author__ = ["baraline"]
 
 import numpy as np
 import pytest

aeon/similarity_search/top_k_similarity.py

@@ -11,7 +11,7 @@ class TopKSimilaritySearch(BaseSimiliaritySearch):
 
     Finds the closest k series to the query series based on a distance function.
 
-    Attributes
+    Parameters
     ----------
     k : int, default=1
         The number of nearest matches from Q to return.
@@ -22,6 +22,27 @@ class TopKSimilaritySearch(BaseSimiliaritySearch):
     store_distance_profile : bool, default = =False.
         Whether to store the computed distance profile in the attribute
         "_distance_profile" after calling the predict method.
+
+    Attributes
+    ----------
+    _X : array, shape (n_instances, n_channels, n_timestamps)
+        The input time series stored during the fit method.
+    distance_profile_function : function
+        The function used to compute the distance profile affected
+        during the fit method based on the distance and normalize
+        parameters.
+
+    Examples
+    --------
+    >>> from aeon.similarity_search import TopKSimilaritySearch
+    >>> from aeon.datasets import load_unit_test
+    >>> X_train, y_train = load_unit_test(split="train")
+    >>> X_test, y_test = load_unit_test(split="test")
+    >>> clf = TopKSimilaritySearch(k=1)
+    >>> clf.fit(X_train, y_train)
+    TopKSimilaritySearch(...)
+    >>> q = X_test[0, :, 5:15]
+    >>> y_pred = clf.predict(q)
     """
 
     def __init__(