|
|
@ -329,6 +329,18 @@ def find_entities(tsv_file, tsv_out_file, ner_rest_endpoint, ned_rest_endpoint,
|
|
|
|
|
|
|
|
|
|
|
|
tsv, ner_result = ner(tsv, ner_rest_endpoint)
|
|
|
|
tsv, ner_result = ner(tsv, ner_rest_endpoint)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
elif os.path.exists(tsv_file):
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
print('Using NER information that is already contained in file: {}'.format(tsv_file))
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
tmp = tsv.copy()
|
|
|
|
|
|
|
|
tmp['sen'] = (tmp['No.'] == 0).cumsum()
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
ner_result = [[{'word': row.TOKEN, 'prediction': row['NE-TAG']} for _, row in sen.iterrows]
|
|
|
|
|
|
|
|
for _, sen in tmp.groupby('sen')]
|
|
|
|
|
|
|
|
else:
|
|
|
|
|
|
|
|
raise RuntimeError("Either NER rest endpoint or NER-TAG information within tsv_file required.")
|
|
|
|
|
|
|
|
|
|
|
|
if ned_rest_endpoint is not None:
|
|
|
|
if ned_rest_endpoint is not None:
|
|
|
|
|
|
|
|
|
|
|
|
tsv, ned_result = ned(tsv, ner_result, ned_rest_endpoint, json_file=ned_json_file, threshold=ned_threshold)
|
|
|
|
tsv, ned_result = ned(tsv, ner_result, ned_rest_endpoint, json_file=ned_json_file, threshold=ned_threshold)
|
|
|
|