Added the mistake_finder.py (forgot in last commit). Used for filtering out...

Added the mistake_finder.py (forgot in last commit). Used for filtering out the mistakes. Skeleton now. Integrated its call in statistics.py for the Anchor Algorithm (commented out now). Small PEP8 change in anchor_algorithm.py.

umbra/anchor_algorithm.py

@@ -39,8 +39,8 @@ class AnchorAlgorithm(AlignmentStrategy):
                 if src_last_anchor < src_index:
                     shd_index = self._shadow.index(word.anchor)
                     self._search_between_anchors(src_last_anchor,
-                                                shd_last_anchor, src_index,
-                                                shd_index)
+                                                 shd_last_anchor, src_index,
+                                                 shd_index)
                 shd_last_anchor = self._shadow.index(word.anchor) + 1
                 src_last_anchor = src_index + 1
 

umbra/mistake_finder.py

+from mistake_enum import Mistake
+
+
+# TODO: Finish this skeleton.
+class MistakeFinder:
+    """Finds the mistakes in already aligned lists of Words"""
+
+    @staticmethod
+    def start(self, source, shadow):
+        """Find all the mistakes and classify them."""
+        # Loop over the shadow:
+        for index, word in enumerate(shadow):
+            if not self._check_repetition(word, index):  # If not a repetition,
+                if not self._check_semantic_mistake(word):  # semantic,
+                    if not self._check_phonentic_mistake(word):  # or phonetic,
+                        # ... then shadow word is random:
+                        word.mistake = Mistake.RANDOM
+        # Loop over the source:
+        for word in source:
+            if not word.shadowed and word.mistake is None:
+                # If not yet marked as mistake, then it is skipped:
+                word.mistake = Mistake.SKIPPED
+
+    # def _check_repetition(self, word, index):
+    #
+
+    # TODO: WordNet integration. Function below should be usable already!
+    # def _check_semantic_mistake(self, shd_word):
+    #     shd_string = shd_word.word
+    #     src_string = shd_word.source.word
+    #     if model.seman_check.semantically_related(src_string, shd_string):
+    #         shd_word.mistake = Mistake.SEMANTIC
+    #         src_word.mistake = Mistake.SEMANTIC
+    #         return True
+    #     return False
+
+    # def _check_phonetic_mistake(self, shd_word):
+    #

umbra/statistics.py

 from saa_algorithm import SaaAlgorithm
 from saa_Romeo import SaaRomeo
 from anchor_algorithm import AnchorAlgorithm
+from mistake_finder import MistakeFinder
 
 
 class Statistics:
@@ -27,7 +28,7 @@ class Statistics:
             shadow: the words in the shadow file
         """
 
-        #Alignment 0
+        # Alignment 0
         print('Romeo')
         self._strategy = SaaRomeo()
         source_align, shadow_align = self._strategy.align(source, shadow)
@@ -51,6 +52,8 @@ class Statistics:
             if s_word.has_source():
                 print(f'source: {s_word.source} shadow: {s_word}')
         correctness = self._strategy.correctly_shadowed(source)
-
+        # TODO: Make the mistake finding work with the statement below
+        # MistakeFinder.start(source_align_em, shadow_align_em)
+        
         return source_align, shadow_align, correctness