def greedy_process(text, _, allcase=False, topn=1, **kwargs):
entities = []
mentions = mention.greedy_detector(erd_db, text, allcase=allcase)
for m in mentions:
if stopped(m[0]):
continue
mid = None
score = 0.9
for title in erd_db.gen_texts(m[0], allcase):
mid, score = erd_db.freebase_text(title, topn=topn)
if mid is not None:
break
if mid is None:
continue
entities.append((m[1], m[2], mid, second_id(mid), m[0], score, 0.99))
# filter less popular texts
mentioned = set()
for e in entities:
if e[-2] >= 0.9:
mentioned.add(e[2])
filtered = []
for e in entities:
if e[2] in mentioned:
filtered.append(e)
return filter_over(filtered), []
def greedy_process(text, _, allcase=False, topn=1, **kwargs):
entities = []
mentions = mention.greedy_detector(erd_db, text, allcase=allcase)
for m in mentions:
if stopped(m[0]):
continue
mid = None
score = 0.9
for title in erd_db.gen_texts(m[0], allcase):
mid, score = erd_db.freebase_text(title, topn=topn)
if mid is not None:
break
if mid is None:
continue
entities.append((m[1], m[2], mid, second_id(mid), m[0], score, 0.99))
# filter less popular texts
mentioned = set()
for e in entities:
if e[-2] >= 0.9:
mentioned.add(e[2])
filtered = []
for e in entities:
if e[2] in mentioned:
filtered.append(e)
return filter_over(filtered), []
"""test
test.py DB-PATH QUERY_TEXT
"""
import sys
import mention
import db
erd_db = db.Database()
erd_db.open(sys.argv[1])
print('start testing')
with open(sys.argv[2], 'r') as f:
print(mention.greedy_detector(erd_db, f.read()))
while True:
get = input()
if len(get) == 0:
break
else:
print(mention.greedy_detector(erd_db, get))
