def test_merge_messages(self):
new_df = merge_messages(self.df2,
max_time_diff=td(weeks=4),
string_similar_threshold=50
)
self._check_mid_year(new_df)
assert_true(len(new_df[new_df['message_id'] == 71808]) > 0)
def main():
import argparse
parser = argparse.ArgumentParser()
parser.add_argument('-d', '--dataset', required=True)
parser.add_argument('--hashtag_ban')
args = parser.parse_args()
try:
df = pd.read_json('data/{}/interactions.json'.format(args.dataset))
except (ValueError, IOError):
df = pd.read_pickle('data/{}/interactions.pkl'.format(args.dataset))
df = df.drop_duplicates(subset=['message_id'])
df['hashtags'] = df['hashtags'].apply(
lambda hs: list(set(map(lambda s: s.lower(), hs)))
)
if args.hashtag_ban:
df['hashtags'] = df['hashtags'].apply(
lambda hs: filter(lambda h: h != args.hashtag_ban, hs)
)
df = remove_mentions_and_urls(df)
df = df[df['body'].map(len) > 10] # filter short body
# df = df[df['body'].map(detect_lan) == 'en'] # non english
df = merge_messages(df,
timedelta(minutes=30),
50,
'datetime')
# df.to_json('data/{}/interactions_new.json'.format(args.dataset),
# orient='records')
df.to_pickle('data/{}/interactions.pkl'.format(args.dataset))
def test_twitter_case(self):
df = merge_messages(self.df3,
max_time_diff=td(days=1),
string_similar_threshold=50,
time_field='datetime')
assert_equal(592820666782580736, df.iloc[0]['message_id'])
# raise
assert_equal(1, len(df))
def test_twitter_case(self):
df = merge_messages(
self.df3,
max_time_diff=td(days=1),
string_similar_threshold=50,
time_field='datetime'
)
assert_equal(
592820666782580736,
df.iloc[0]['message_id']
)
# raise
assert_equal(1, len(df))
def test_merge_messages(self):
new_df = merge_messages(self.df2,
max_time_diff=td(weeks=4),
string_similar_threshold=50)
self._check_mid_year(new_df)
assert_true(len(new_df[new_df['message_id'] == 71808]) > 0)
def filter_and_save(df):
# df = process_message_body(df)
# df = df[df['body'].map(len) > 10] # filter short body
# potential: 14557
black_list = [256, 1175]
for s in black_list:
df = df[df['sender_id'] != s]
df.to_json('data/enron/interactions.json', orient="records")
if __name__ == '__main__':
import argparse
parser = argparse.ArgumentParser()
parser.add_argument('-d', '--dataset')
args = parser.parse_args()
df = pd.read_json('data/{}/interactions.json'.format(args.dataset))
df = merge_messages(df, timedelta(weeks=4), 70, 'datetime')
df.to_json('data/{}/interactions.json'.format(args.dataset))
# frequent_senders = df['sender_id'].value_counts().index[20:40]
# for s in frequent_senders:
# print(s)
# print(df[df['sender_id'] == s]['subject'][:10])
# print('*' * 20)
# filter_and_save(df)