Refactored methods and added its proper documentation #195

Author: dpuenteramirez

instance_selection/_CNN.py

@@ -3,34 +3,29 @@
 # @Filename:    CNN.py
 # @Author:      Daniel Puente Ramírez
 # @Time:        19/11/21 07:13
-# @Version:     4.0
+# @Version:     5.0
 
 import numpy as np
 import pandas as pd
 
 from .utils import transform, delete_multiple_element
 
 
-def check_store(store, sample, store_classes):
-    euc = []
-    for s in store:
-        euc.append(np.linalg.norm(s - sample))
-    euc = np.array(euc)
-    euc_nn = np.amin(euc)
-    index_nn = np.ravel(np.where(euc == euc_nn))
-    return store_classes[index_nn[0]]
+class CNN:
+    """
+    Hart, P. (1968). The condensed nearest neighbor rule (corresp.). IEEE
+    transactions on information theory, 14(3), 515-516.
 
+    Parameters
+    ----------
 
-class CNN:
+    """
 
     def __init__(self):
         self.x_attr = None
 
     def filter(self, samples, y):
         """
-        Hart, P. (1968). The condensed nearest neighbor rule (corresp.). IEEE
-            transactions on information theory, 14(3), 515-516.
-
         Implementation of The Condensed Nearest Neighbor Rule
 
         The first sample of each class is placed in *store*. Thus we only have
@@ -48,6 +43,7 @@ def filter(self, samples, y):
         Extracted from:
         The condensed nearest neighbor rule. IEEE Transactions on Information
         Theory ( Volume: 14, Issue: 3, May 1968)
+
         :param samples: DataFrame.
         :param y: DataFrame.
         :return: the input dataset with the remaining samples.
@@ -61,7 +57,7 @@ def filter(self, samples, y):
         handbag = []
 
         for sample_class, sample in zip(samples.target, samples.data):
-            nn_class = check_store(store, sample, store_classes)
+            nn_class = self._check_store(store, sample, store_classes)
 
             if nn_class == sample_class:
                 handbag.append((sample_class, sample))
@@ -75,7 +71,7 @@ def filter(self, samples, y):
             indexes = []
             for index, s2 in enumerate(handbag):
                 sample_class, sample = s2
-                nn_class = check_store(store, sample, store_classes)
+                nn_class = self._check_store(store, sample, store_classes)
                 if nn_class != sample_class:
                     store.append(sample)
                     store_classes.append(sample_class)
@@ -88,3 +84,24 @@ def filter(self, samples, y):
             int))
 
         return samples, y
+
+    @staticmethod
+    def _check_store(store, sample, store_classes):
+        """
+        > The function takes in a sample, a store of samples, and the classes of
+         the store of samples. It then calculates the Euclidean distance
+         between the sample and each sample in the store. It then returns the
+         class of the sample in the store that is closest to the sample
+
+        :param store: the list of samples that have been stored
+        :param sample: the sample we want to classify
+        :param store_classes: the classes of the samples in the store
+        :return: The class of the nearest neighbor.
+        """
+        euc = []
+        for s in store:
+            euc.append(np.linalg.norm(s - sample))
+        euc = np.array(euc)
+        euc_nn = np.amin(euc)
+        index_nn = np.ravel(np.where(euc == euc_nn))
+        return store_classes[index_nn[0]]