#398 related

arekit/contrib/source/brat/entities/compound.py

@@ -0,0 +1,24 @@
+from arekit.contrib.source.brat.entities.entity import BratEntity
+
+
+class BratCompoundEntity(BratEntity):
+    """ Entity which contains the hierarchy of the other entities.
+    """
+
+    def __init__(self, id_in_doc, value, e_type, root, entities, index_begin, index_end, group_index=None):
+        assert(isinstance(entities, list))
+        assert(isinstance(root, BratCompoundEntity) or root is None)
+        super(BratCompoundEntity, self).__init__(value=value, e_type=e_type,
+                                                 id_in_doc=id_in_doc,
+                                                 index_begin=index_begin,
+                                                 index_end=index_end,
+                                                 group_index=group_index)
+        self.__entities = entities
+        self.__root = root
+
+    @property
+    def Root(self):
+        return self.__root
+
+    def iter_childs(self):
+        return iter(self.__entities)

arekit/contrib/source/brat/sentences_reader.py

@@ -10,10 +10,10 @@ def from_file(input_file, entities, line_handler=None, skip_entity_func=None):
         assert(isinstance(entities, EntityCollection))
         assert(callable(skip_entity_func) or skip_entity_func is None)
 
-        sentences_data = BratDocumentSentencesReader.__parse_sentences(
+        sentences_data = BratDocumentSentencesReader._parse_sentences(
             input_file=input_file, line_handler=line_handler)
 
-        sentence_entities = BratDocumentSentencesReader.__parse_entities(
+        sentence_entities = BratDocumentSentencesReader._parse_entities(
             sentences_data=sentences_data,
             entities=entities,
             skip_entity_func=skip_entity_func)
@@ -28,9 +28,24 @@ def from_file(input_file, entities, line_handler=None, skip_entity_func=None):
 
         return brat_sentences
 
-    # endregion
+    @staticmethod
+    def from_sentences_data(entities, sentences_data, skip_entity_func=None):
+        assert(isinstance(entities, EntityCollection))
+
+        sentence_entities = BratDocumentSentencesReader._parse_entities(
+            sentences_data=sentences_data,
+            entities=entities,
+            skip_entity_func=skip_entity_func)
 
-    # region private methods
+        # Convert all the content to brat sentences.
+        brat_sentences = []
+        for s_ind, s_dict in enumerate(sentences_data):
+            brat_sentence = BratSentence(text=s_dict["text"],
+                                         index_begin=s_dict["ind_begin"],
+                                         entities=sentence_entities[s_ind])
+            brat_sentences.append(brat_sentence)
+
+        return brat_sentences
 
     @staticmethod
     def __is_sentence_contains(sentence_data, entity):
@@ -40,7 +55,7 @@ def __is_sentence_contains(sentence_data, entity):
                entity.IndexEnd <= sentence_data["ind_end"]
 
     @staticmethod
-    def __parse_entities(sentences_data, entities, skip_entity_func):
+    def _parse_entities(sentences_data, entities, skip_entity_func):
         """ Sentences is a list of json-like data (dictionaries).
         """
         assert(isinstance(sentences_data, list))
@@ -91,7 +106,7 @@ def __parse_entities(sentences_data, entities, skip_entity_func):
         return entities_in_sentences
 
     @staticmethod
-    def __parse_sentences(input_file, line_handler):
+    def _parse_sentences(input_file, line_handler):
         assert(callable(line_handler) or line_handler is None)
         sentences = []
         line_start = 0

tests/contrib/source/test_brat_compound_entities.py

@@ -0,0 +1,45 @@
+import unittest
+
+from arekit.common.bound import Bound
+from arekit.common.entities.collection import EntityCollection
+from arekit.common.synonyms.grouping import SynonymsCollectionValuesGroupingProviders
+from arekit.contrib.source.brat.entities.entity import BratEntity
+from arekit.contrib.source.brat.sentences_reader import BratDocumentSentencesReader
+from arekit.contrib.utils.processing.lemmatization.mystem import MystemWrapper
+from arekit.contrib.utils.synonyms.stemmer_based import StemmerBasedSynonymCollection
+
+
+class TestCompoundEntites(unittest.TestCase):
+
+    text = "мама мыла раму"
+    entities = [
+        BratEntity(id_in_doc="T1", e_type="PERSON", index_begin=0, index_end=4, value="мама"),
+        BratEntity(id_in_doc="T2", e_type="VERB", index_begin=5, index_end=9, value="мыла"),
+        BratEntity(id_in_doc="T3", e_type="OBJECT", index_begin=10, index_end=14, value="раму"),
+        BratEntity(id_in_doc="T3", e_type="ACTION", index_begin=0, index_end=9, value="мама мыла")
+    ]
+
+    def test(self):
+        s_data = [
+            {"text": self.text, "ind_begin": 0, "ind_end": len(self.text)}
+        ]
+
+        synonyms = StemmerBasedSynonymCollection(
+            iter_group_values_lists=[], stemmer=MystemWrapper(), is_read_only=False, debug=False)
+
+        collection = EntityCollection(
+            self.entities,
+            value_to_group_id_func=lambda value:
+                SynonymsCollectionValuesGroupingProviders.provide_existed_or_register_missed_value(synonyms, value))
+
+        sentences = BratDocumentSentencesReader.from_sentences_data(entities=collection,
+                                                                    sentences_data=s_data)
+
+        for sentence in sentences:
+            for e, b in sentence.iter_entity_with_local_bounds():
+                assert(isinstance(b, Bound))
+                print(e.Value, b.Position, b.Position + b.Length)
+
+
+if __name__ == '__main__':
+    unittest.main()