summarize_activity.py

#!/usr/bin/python                                                                                                                                                                             

import os
import sys
from datetime import date, timedelta, datetime
import simplejson as json


def parse(filex):
    data = json.loads(filex.read(), 'utf8')
    activity_summary = {}

    for d in data:
        dt = datetime.strptime(d["datetime"], '%Y-%m-%dT%H:%M:%S').strftime('%Y-%m-%d')

        act_tag = '%s_%s' % (dt, d["extra"]["type"])

        # check whether we have users for that activity                                                                                                                                       
        if (act_tag in activity_summary):
            users = activity_summary[act_tag]
        else:
            activity_summary[act_tag] = users = {}

        users[d['email']] = users.setdefault(d['email'], 0) + 1

    return activity_summary


if __name__ == '__main__':

    activity = parse(sys.stdin)

    for act_tag in activity:
        for user in activity[act_tag]:
            dt = act_tag.split("_")[0]
            cnt = activity[act_tag][user]

            print '%s,%s,%s' % (dt, user, cnt)