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['publication_date']
claim.verdict_date = date(verdict_date)
post_guid = compute_post_guid(self._social_network_url, claim_id, verdict_date)
claim.claim_id = post_guid
claim.domain = self._domain
keywords = unicode(row['keywords'])
claim.keywords = keywords
verdict = unicode(row['post_type'])
claim.verdict = verdict
return claim
def _add_claim(self, claim_id, content, date_str, post_type=None):
claim = Claim()
claim.claim_id = claim_id
claim.verdict = post_type
claim.title = claim_id
claim.description = content
claim.verdict_date = convert_str_to_unicode_datetime(date_str)
claim.url = "claim url"
self._db.addPost(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 _add_claim(self,
claim_id,
content,
date_str=u"2017-06-14 05:00:00",
keywords=u"",
post_type=None):
claim = Claim()
claim.claim_id = claim_id
claim.verdict = post_type
claim.title = claim_id
claim.description = content
claim.verdict_date = convert_str_to_unicode_datetime(date_str)
claim.keywords = keywords
claim.url = u"claim url"
self._db.addPost(claim)
self._claim_dictionary[claim.claim_id] = 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)
def _convert_row_to_claim(self, row):
claim = Claim()
claim_id = unicode(row['campaign_id'])
title = unicode(row['title'], errors='replace')
claim.title = title
try:
verdict_date = datetime.datetime.strftime(
datetime.datetime.strptime(row['date'], '%d-%m-%y'),
'%Y-%m-%d %H:%M:%S')
except:
verdict_date = getConfig().eval(self.__class__.__name__,
'default_date')
claim.verdict_date = date(verdict_date)
post_guid = compute_post_guid(self._social_network_url, claim_id,
verdict_date)
claim.claim_id = post_guid
claim.domain = self._domain
keywords = unicode(row['category'])
claim.keywords = keywords
verdict = unicode(row['campaign_class'])
claim.verdict = verdict
#claim.claim_topic = unicode(row['category'])
return claim
def _convert_row_to_claim(self, claim_row):
claim = Claim()
claim.claim_id = compute_author_guid_by_author_name(
str(claim_row['social_id']))
claim.domain = str(claim_row['site'])
claim.verdict = str(claim_row['ruling_val'])
claim.verdict_date = datetime.datetime.fromtimestamp(
claim_row['ruling_time'])
self._claims.append(claim)
attribute_name = self.__class__.__name__ + "_claim_type"
author_feature = BaseFeatureGenerator.create_author_feature(
attribute_name, claim.claim_id, claim.verdict, self._window_start,
self._window_end)
self._author_features.append(author_feature)
attribute_name = self.__class__.__name__ + "_claim_id"
author_feature = BaseFeatureGenerator.create_author_feature(
attribute_name, claim.claim_id, claim.claim_id, self._window_start,
self._window_end)
self._author_features.append(author_feature)
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 _add_claim(self, claim_id):
claim = Claim()
claim.claim_id = claim_id
self._db.addPosts([claim])
def _parse_articles_lst_to_articles(self, all_articles_lst_of_dics):
print("###### 'Entering _parse_articles_lst_to_articles'")
parsed_articles_lst = []
claims_lst = []
posts_lst = []
articles_lst = []
article_items_lst = []
# Parsing articles list of dictionaries data, received using the API.
for news_articles_dic in all_articles_lst_of_dics:
print("###### 'PARSING: {}'".format(str(news_articles_dic)))
parsed_articles_lst += self._parse_news_article(news_articles_dic)
# For each news article dictionary commit:
for parsed_news_article in parsed_articles_lst:
print("###### 'Iterating parsed_articles_lst single item: {}'".
format(str(parsed_news_article)))
# Building: claim & News_Article & News_Article_Item objects.
claim = Claim()
post = Post()
article = News_Article()
article_item = News_Article_Item()
# Initializing Claim object with data:
identifier = compute_post_guid(parsed_news_article['url'],
parsed_news_article['author'],
parsed_news_article['publishedAt'])
claim.claim_id = post.post_id = post.guid = post.post_osn_guid = article.post_id = article_item.post_id = unicode(
identifier)
author_guid = compute_author_guid_by_author_name(
parsed_news_article['author'])
post.author_guid = article.author_guid = article_item.author_guid = unicode(
author_guid)
post.author = article.author = unicode(
parsed_news_article['author'])
post.title = claim.title = article.title = unicode(
parsed_news_article['title'])
post.content = article_item.content = unicode(
parsed_news_article['content'])
post.description = claim.description = article.description = unicode(
parsed_news_article['description'])
post.date = post.created_at = claim.verdict_date = article.published_date = datetime.datetime.strptime(
parsed_news_article['publishedAt'], '%Y-%m-%d %H:%M:%S')
article_item.source_newsapi_internal_id = unicode(
parsed_news_article['source_id'])
article_item.source_newsapi_internal_name = unicode(
parsed_news_article['source_name'])
post.url = claim.url = article.url = unicode(
parsed_news_article['url'])
article_item.img_url = unicode(parsed_news_article['urlToImage'])
post.post_type = claim.verdict = unicode(
"TRUE"
) # todo: Add constant. We assume all news articles are true.
post.domain = claim.domain = unicode(
"NewsSite") # todo: Add constant.
# Update objects lists:
posts_lst.append(post)
claims_lst.append(claim)
articles_lst.append(article)
article_items_lst.append(article_item)
print("###### 'EXITING _parse_articles_lst_to_articles'")
return posts_lst, claims_lst, articles_lst, article_items_lst
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
