def extractHardFeatures(DIR):
#TODO uncomment
posts = pd.read_csv(DIR + 'posts/users_posts_times.csv')
#activities = useractivity.usersActivityFast(posts)
year, month, day = useractivity.theLatestPostTime(posts)
#extract last week and last day posts
posts_week, posts_day = useractivity.extractDayWeekActivity(posts, year, month, day)
#print posts_week
df_result = useractivity.extractTimeIntervalFeatures(posts_week, posts_day)
print 'extracted Q_LAST_WEEK, A_LAST_WEEK, P_NUM_LAST_WEEK'
df_result.to_csv(DIR + 'users/temp_features.csv', index=False)
activities = useractivity.usersActivityFast(posts)
df_tr = useractivity.userActivityTransform(activities)
df_tr.to_csv(DIR + 'users/temporal_user_activities.csv', index=False)
df = pd.read_csv(DIR + 'posts/quest_stats.csv')
# extract features when a question was asked
q_wknd = timefeatures.dateWeekend(df)
q_wknd.to_csv(DIR + 'posts/quest_weekend.csv', index=False)
extractLocs(DIR)
df_tags = pd.read_csv(DIR + 'posts/quest_stats.csv')
extractTagFeatures(DIR, df_tags)
def dataCut(data, time_cutoff):
data_sorted = data.sort(['TimeAsked'], ascending=False)
data_entries = 0
for (i, timestamp) in enumerate(data_sorted['TimeAsked']):
time_answered = tmpf.parseTime(timestamp)
if time_answered < time_cutoff:
data_entries = i
print data_entries, 'questions to cut'
break
data_cut = data_sorted[1:data_entries]
return data_cut
def dataCut(data, time_cutoff):
data_sorted = data.sort(['TimeAsked'], ascending=False)
data_entries = 0
for (i, timestamp) in enumerate(data_sorted['TimeAsked']):
time_answered = tmpf.parseTime(timestamp)
if time_answered < time_cutoff:
data_entries = i
print data_entries, 'questions to cut'
break
data_cut = data_sorted[1:data_entries]
return data_cut
