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 extractUserActivityFeatures(c, DIR):
if not os.path.exists(DIR + 'users'):
os.makedirs(DIR + 'users')
users_activity = pd.DataFrame(query.usersActivity(c), columns=header['users_activity'])
users_activity.to_csv(DIR + 'posts/users_activity.csv', index=False)
posts = useractivity.makePosts(users_activity)
posts = posts[posts['UserId'] != -1]
return posts