sessionize_logic.py

# requires data to be in format <user_sk>\t<timestamp>, clustered by user_sk and sorted by <timestamp> ascending
line = ''
current_uid = ''
last_click_time = -1
perUser_sessionID_counter = 1
timeout  = 60*60
  
for line in click_rows:	
  user_sk, tstamp_str = line.strip().split("\t")
  tstamp = long(tstamp_str)

  # reset if next partition beginns
  if current_uid != user_sk:
    current_uid = user_sk
    perUser_sessionID_counter = 1
    last_click_time = tstamp

  # time between clicks exceeds session timeout?
  if tstamp - last_click_time > timeout:
    perUser_sessionID_counter += 1

  last_click_time = tstamp
  print(f"{user_sk}, {str(perUser_sessionID_counter)}")