def translateNCRFPPintoEntities(doc_token, predict_results, doc_name):
entity_id = 1
results = []
sent_num = len(predict_results)
for idx in range(sent_num):
sent_length = len(predict_results[idx][0])
sent_token = doc_token[(doc_token['sent_idx'] == idx)]
assert sent_token.shape[0] == sent_length, "file {}, sent {}".format(
doc_name, idx)
labelSequence = []
for idy in range(sent_length):
token = sent_token.iloc[idy]
label = predict_results[idx][0][idy]
labelSequence.append(label)
if label[0] == 'S' or label[0] == 'B':
entity = Entity()
entity.create(str(entity_id), label[2:], token['start'],
token['end'], token['text'], idx, idy, idy)
results.append(entity)
entity_id += 1
elif label[0] == 'M' or label[0] == 'E':
if checkWrongState(labelSequence):
entity = results[-1]
entity.append(token['start'], token['end'], token['text'],
idy)
return results
def read_one_file(fileName, annotation_dir, entities_overlapped_types):
annotation_file = get_bioc_file(join(annotation_dir, fileName))
bioc_passage = annotation_file[0].passages[0]
entities = []
for entity in bioc_passage.annotations:
entity_ = Entity()
entity_.create(entity.id, entity.infons['type'],
entity.locations[0].offset, entity.locations[0].end,
entity.text, None, None, None)
for old_entity in entities:
if is_overlapped(entity_, old_entity):
logging.debug(
"entity overlapped: doc:{}, entity1_id:{}, entity1_type:{}, entity1_span:{} {}, entity2_id:{}, entity2_type:{}, entity2_span:{} {}"
.format(fileName, old_entity.id, old_entity.type,
old_entity.start, old_entity.end, entity_.id,
entity_.type, entity_.start, entity_.end))
overlapped_types = entity_.type + "_" + old_entity.type if cmp(
entity_.type, old_entity.type
) > 0 else old_entity.type + "_" + entity_.type
if overlapped_types in entities_overlapped_types:
count = entities_overlapped_types[overlapped_types]
count += 1
entities_overlapped_types[overlapped_types] = count
else:
entities_overlapped_types[overlapped_types] = 1
entities.append(entity_)
