Train using NER tags in BIO format

ahisto_named_entity_search/recognition/model.py

@@ -65,7 +65,7 @@ class NerModel:
             ner_texts, all_ner_tags = [], []
             for tagged_sentence in TaggedSentence.load(tagged_sentence_basename):
                 ner_texts.append(tagged_sentence.sentence)
-                all_ner_tags.append(tagged_sentence.ner_tags)
+                all_ner_tags.append(tagged_sentence.bio_ner_tags)
             return ner_texts, all_ner_tags
 
         ner_training_texts, ner_training_labels = load_ner_dataset(training_tagged_sentence_basename)

ahisto_named_entity_search/search/result.py

@@ -64,6 +64,23 @@ class TaggedSentence:
         ner_tags = ' '.join(self.ner_tags_tuple)
         return ner_tags
 
+    @property
+    def bio_ner_tags(self) -> NerTags:
+        previous_ner_tag = None
+        bio_ner_tags_list = []
+        for ner_tag in self.ner_tags_tuple:
+            if ner_tag == 'O':
+                bio_ner_tag = ner_tag
+            else:
+                if previous_ner_tag is None or ner_tag != previous_ner_tag:
+                    bio_ner_tag = f'B-{ner_tag}'
+                else:
+                    bio_ner_tag = f'I-{ner_tag}'
+            bio_ner_tags_list.append(bio_ner_tag)
+            previous_ner_tag = ner_tag
+        bio_ner_tags = ' '.join(bio_ner_tags_list)
+        return bio_ner_tags
+
     @classmethod
     def save(cls, basename: str, tagged_sentences: Iterable['TaggedSentence']) -> None:
         sentences_filename = cls._get_sentences_filename(basename)