#376 refactoring, removed id-depended parsing methods

arekit/common/data/const.py

@@ -3,6 +3,10 @@
 TEXT = 'text_a'
 LABEL = 'label'
 
+# Global identifier of the opinion in the sampled data.
+OPINION_ID = "opinion_id"
+OPINION_LINKAGE_ID = "linkage_id"
+
 # Corresponds to fields with attitude ends. (values, STRING)
 SOURCE = 'source'
 TARGET = 'target'

arekit/common/data/input/providers/rows/samples.py

@@ -70,6 +70,10 @@ def __assign_value(column, value):
             index_in_linked=index_in_linked,
             label_scaler=self._label_provider.LabelScaler)
 
+        row[const.OPINION_ID] = text_opinion_linkage.First.TextOpinionID
+
+        row[const.OPINION_LINKAGE_ID] = index_in_linked
+
         row[const.DOC_ID] = text_opinion_linkage.First.DocID
 
         row[const.SENT_IND] = sentence_ind

arekit/common/data/row_ids/base.py

@@ -54,26 +54,3 @@ def create_pattern(id_value, p_type):
     def convert_sample_id_to_opinion_id(sample_id):
         assert(isinstance(sample_id, str))
         return sample_id[:sample_id.index(BaseIDProvider.INDEX[0])] + BaseIDProvider.INDEX.format(0)
-
-    # region 'parse' methods
-
-    @staticmethod
-    def _parse(row_id, pattern):
-        assert(isinstance(pattern, str))
-
-        _from = row_id.index(pattern[0]) + 1
-        _to = row_id.index(BaseIDProvider.SEPARATOR, _from, len(row_id))
-
-        return int(row_id[_from:_to])
-
-    @staticmethod
-    def parse_opinion_in_opinion_id(opinion_id):
-        assert(isinstance(opinion_id, str))
-        return BaseIDProvider._parse(opinion_id, BaseIDProvider.OPINION)
-
-    @staticmethod
-    def parse_opinion_in_sample_id(sample_id):
-        assert(isinstance(sample_id, str))
-        return BaseIDProvider._parse(sample_id, BaseIDProvider.OPINION)
-
-    # endregion

arekit/common/data/views/samples.py

@@ -15,9 +15,7 @@ def iter_from_storage(self, storage):
 
         linked = []
         current_opinion_id = undefined
-        for row_index, sample_id in enumerate(storage.iter_column_values(const.ID)):
-            sample_id = str(sample_id)
-            opinion_id = self.__row_ids_provider.parse_opinion_in_sample_id(sample_id)
+        for row_index, opinion_id in enumerate(storage.iter_column_values(const.OPINION_ID)):
             if current_opinion_id != undefined:
                 if opinion_id != current_opinion_id:
                     yield linked

arekit/contrib/networks/input/rows_parser.py

@@ -36,6 +36,8 @@ def __handle_text(value):
     const.S_IND: lambda value: int(value),
     const.T_IND: lambda value: int(value),
     const.SENT_IND: lambda value: int(value),
+    const.OPINION_ID: lambda value: int(value),
+    const.OPINION_LINKAGE_ID: lambda value: int(value),
     const.ENTITY_VALUES: lambda value: __process_values_list(value),
     const.ENTITY_TYPES: lambda value: __process_values_list(value),
     const.ENTITIES: lambda value: __process_indices_list(value),
@@ -82,7 +84,15 @@ def __value_or_none(self, key):
     @property
     def SampleID(self):
         return self.__params[const.ID]
-
+
+    @property
+    def OpinionID(self):
+        return self.__params[const.OPINION_ID]
+
+    @property
+    def OpinionLinkageID(self):
+        return self.__params[const.OPINION_LINKAGE_ID]
+
     @property
     def Terms(self):
         return self.__params[const.TEXT]

arekit/contrib/utils/data/views/linkages/base.py

@@ -20,10 +20,6 @@ def __init__(self, ids_provider, storage):
 
     # region private methods
 
-    def __iter_doc_opinion_ids(self, row_ids):
-        for row_id in row_ids:
-            yield self._ids_provider.parse_opinion_in_opinion_id(row_id)
-
     def __iter_opinions_by_linkages(self, linkages_df, opinions_view):
         for df_linkage in linkages_df:
             assert (isinstance(df_linkage, pd.DataFrame))
@@ -42,9 +38,8 @@ def _iter_by_opinions(self, linked_df, opinions_view):
 
     def iter_opinion_linkages(self, doc_id, opinions_view):
         assert(isinstance(opinions_view, BaseOpinionStorageView))
-        doc_df = self._storage.find_by_value(column_name=const.DOC_ID, value=doc_id)
-        row_ids = [row_id for row_id in doc_df[const.ID]]  # TODO. Adopt storage.
-        doc_opin_ids = self.__iter_doc_opinion_ids(row_ids=row_ids)
+        doc_df = self._storage.find_by_value(column_name=const.OPINION_ID, value=doc_id)
+        doc_opin_ids = [opinion_id for opinion_id in doc_df[const.OPINION_ID]]
 
         doc_opin_id_patterns = map(
             lambda opinion_id: self._ids_provider.create_pattern(id_value=opinion_id, p_type=BaseIDProvider.OPINION),