#!/usr/bin/python

import sys
import json
import zipfile
import tarfile
import re
import datetime

ROOMS = 'hipchat_export/rooms/'
USERLIST = 'hipchat_export/users/list.json'
ONEDAY = datetime.timedelta(1)
HOURSECS = 3600

MINDATE = None
MINDATE = datetime.datetime(2017, 1, 1)

T_ROOMS = 'rooms.json'


class Rooms(object):
  def __init__(self, tar):
    fp = tar.extractfile(T_ROOMS)
    self.rooms = [d['Room'] for d in json.load(fp)]

  def print_list(self):
    for room in self.rooms:
      print room['id'], room['name']


def print_room(tar, room):
  for line in _get_messages(tar, room):
    print('[%s] %s:%s: %s' % line)


def histogram_room(zf, room):
  users = { }
  data = { }
  for line in _get_chats(zf, room):
    uid = line[1]
    if uid not in users:
      users[uid] = set([line[2]])
    else:
      users[uid].add(line[2])
    hour = line[0][:line[0].index(':')]
    if hour not in data:
      data[hour] = { uid: 1 }
    elif uid not in data[hour]:
      data[hour][uid] = 1
    else:
      data[hour][uid] += 1

  for uid in users:
    print uid, users[uid]

  uids = users.keys()
  print '\t'.join([''] + uids)
  for hour in sorted(data.keys()):
    columns = [hour] + [str(data[hour].get(uid, '')) for uid in uids]
    print '\t'.join(columns)
    #for uid in sorted(data[hour]):
    #  print hour, uid, data[hour][uid]


def print_users(zf):
  users = _get_users(zf)
  for uid, (name, email) in sorted(users.items(), key=lambda x: x[1][0]):
    print uid, name, email


def histogram_user(zf, user, rooms, for_chart=False):
  data = { }
  for room in rooms:
    for line in _get_chats(zf, room):
      if line[1] != user:
        continue
      y,m,d = [int(x) for x in line[0][:10].split('-')]
      h = int(line[0][11:13])

      slot = datetime.datetime(y, m, d, h)
      if MINDATE and slot < MINDATE:
        continue

      if slot not in data:
        data[slot] = 1
      else:
        data[slot] += 1

  totalhours = 0
  totalcount = 0
  weekhours = 0
  startblock = None
  lasthour = None
  for hour in sorted(data.keys()):
    if not startblock:
      startblock = hour
    if lasthour:
      hours = (hour - lasthour).total_seconds() / HOURSECS
      if hours > 3:
        # a block of interaction has completed. add 1 hour for the LASTHOUR
        # that was observed.
        weekhours += 1

        startweek = startblock.isocalendar()[1]
        thisweek = hour.isocalendar()[1]

        # if the work block was started on/before Friday, and this new
        # block-start after that, then let's note the workweek hours
        #print 'START:', startblock.weekday(), 'HOUR:', hour.weekday()
        if startblock.weekday() <= 4 and \
            (hour.weekday() >= 5 or thisweek > startweek):
          if not for_chart:
            print 'Week Hours:', weekhours

        # if the work block was started in the week prior to this block,
        # then let's wrap up the prior week.
        if thisweek > startweek:
          if for_chart:
            print startweek, weekhours
          else:
            print 'Week+end Hours:', weekhours
          totalhours += weekhours
          weekhours = 0

        # start/label the next block
        startblock = hour
        if not for_chart:
          print '...'
          #print 'HOURS:', weekhours
          print hour.strftime('%A:')
      else:
        weekhours += hours

    lasthour = hour
    if not for_chart:
      print '%s %d' % (hour, data[hour])
    totalcount += data[hour]

  print
  print '---'
  print 'TOTAL (hours):', totalhours
  print 'TOTAL (count):', totalcount


def _get_users(zf):
  users = { }
  fp = zf.open(USERLIST)
  for user in json.load(fp)['users']:
    users[user['user_id']] = (user['name'], user['email'])
  return users


def _get_chats(zf, room):
  chats = [ ]
  # Little hack: only get files like $room/2nnn-* and avoid $room/
  files = [f for f in zf.namelist() if f.startswith(ROOMS + room + '/2')]
  for fname in sorted(files):
    fp = zf.open(fname)
    for line in json.load(fp):
      if not line:
        continue
      date = line['date'].encode('utf8')
      user = line['from']['name'].encode('utf8')
      msg = _clean(line['message'].encode('utf8'))
      if line['from']['user_id'] == 'api':
        uid = 'api'
      else:
        uid = str(line['from']['user_id'])
      chats.append((date, uid, user, msg))
  return chats


def _get_messages(tar, room, only_user=True):
  messages = [ ]
  fp = tar.extractfile('rooms/%s/history.json' % (room,))
  for line in json.load(fp):
    if only_user and 'UserMessage' not in line:
      continue
    try:
      msg = line['UserMessage']
    except KeyError:
      print 'ERROR:', line
      continue
    messages.append((msg['timestamp'].encode('utf8'),
                     msg['sender']['id'],
                     msg['sender']['name'].encode('utf8'),
                     _clean(msg['message'].encode('utf8'))))
  return messages


def _clean(msg):
  return _RE_STRIP_HTML.sub('', msg)
_RE_STRIP_HTML = re.compile('<.*?>')


if __name__ == '__main__':
  cmd = sys.argv[1]
  fname = sys.argv[2]
  if fname.endswith('.zip'):
    zf = zipfile.ZipFile(fname)
    tar = None
  else:
    zf = None
    tar = tarfile.open(fname)

  ### this used to work with zipfiles. HipChat now downloads gzips.
  ### some commands have been converted.

  if cmd == 'list':
    Rooms(tar).print_list()
  elif cmd == 'room':
    print_room(tar, sys.argv[3])
  elif cmd == 'hist':
    histogram_room(zf, sys.argv[3])
  elif cmd == 'users':
    print_users(zf)
  elif cmd == 'user':
    histogram_user(zf, sys.argv[3], sys.argv[4:])
  elif cmd == 'chart':
    histogram_user(zf, sys.argv[3], sys.argv[4:], True)
  else:
    print 'UNKNOWN COMMAND:', cmd
    sys.exit(1)