def parse(self, doc, doc_id):
"""Parses the document and returns all candidate mention-entity pairs.
:return: candidate entities {men:{en:cmn, ...}, ...}
"""
query = Document(doc_id, doc)
candidate_ens = defaultdict(list)
for ngram in query.get_ngrams():
for en in self.commonness.get(ngram, {}):
if en != "_total":
candidate_ens[ngram].append(en)
return candidate_ens
def main():
snippets = Document.load_test_snippets(SNIPPETS)
out_file_name = OUTPUT_DIR + "/output_cmn.txt"
out_file = open(out_file_name, "w")
el = ELCmn()
for doc_id, doc in sorted(snippets.items(), key=lambda item: int(item[0])):
print("[" + doc_id + "]\t" + doc)
linked_ens = el.annotate(doc, doc_id)
ELUtils.write_to_file(doc_id, out_file, linked_ens)
print("------")
print("output written to", out_file_name)