def _convert_row_to_claim(self, row):
claim = Claim()
# claim_id = unicode(row['claim_id'])
title = unicode(row['title'], errors='replace')
claim.title = title
description = unicode(row['description'], errors='replace')
claim.description = description
url = unicode(row['url'])
claim.url = url
verdict_date = row['verdict_date']
claim.verdict_date = date(verdict_date)
post_guid = compute_post_guid(self._social_network_url, url,
verdict_date)
claim.claim_id = post_guid
claim.domain = self._domain
keywords = unicode(row['keywords'])
claim.keywords = keywords
verdict = unicode(row['verdict'])
claim.verdict = verdict
claim.category = unicode(row['main_category'])
claim.sub_category = unicode(row['secondary_category'])
return claim
def create_claim_from_record(self, series):
claim = Claim()
claim.title = str(series['title'])
claim.description = str(series['description'])
claim.url = str(series['url'])
claim.verdict_date = str_to_date(series['verdict_date'])
claim.verdict = str(series['verdict'])
claim.category = str(series['main_category'])
claim.keywords = str(series['keywords'])
claim.domain = str(series['domain'])
post_guid = compute_post_guid(self._social_network_url, claim.url, date_to_str(claim.verdict_date))
claim.claim_id = post_guid
return claim
def _create_claim(self, original_liar_dataset_id, speaker, targeted_label,
statement, subject):
claim = Claim()
osn_id = str(original_liar_dataset_id)
claim.claim_id = compute_post_guid(self._social_network_url, osn_id,
'2007-01-01 00:00:00')
claim.domain = 'LiarLiar'
claim.description = statement
claim.title = statement
claim.url = ''
claim.category = 'Politic'
claim.sub_category = subject
claim.verdict = targeted_label
claim.verdict_date = str_to_date('2007-01-01 00:00:00')
claim.keywords = ''
return claim
def get_claim_from_article(self, article, source):
claim = Claim()
claim.title = article['title']
claim.description = article['description']
claim.url = article['url']
try:
publish_date = str_to_date(article['publishedAt'].split('+')[0],
'%Y-%m-%dT%H:%M:%SZ')
except:
publish_date = str_to_date(article['publishedAt'].split('+')[0],
'%Y-%m-%dT%H:%M:%S')
claim.verdict_date = publish_date
claim.domain = article['source'].get('id', str(source))
claim.verdict = True
claim.claim_id = compute_post_guid(self._social_network_url, claim.url,
date_to_str(claim.verdict_date))
claim.category = str(source)
return claim
def import_to_db(self, file_path):
with open(file_path, 'rb') as f:
data = csv.reader(f)
res = []
for row in data:
try:
c = Claim()
c.title = str(row[3])
c.claim_id = str(uuid.uuid4().__str__())
c.description = str(row[0])
c.category = str(row[4])
c.verdict = str(row[1])
c.domain= "Claim"
c.verdict_date = datetime.datetime.strptime(row[6],'%d %B %Y')
c.url = str(row[2])
except Exception as e:
continue
res.append(c)
print(','.join(row))
self._db.add_claims(res)
