|
| 1 | +""" |
| 2 | +Import sample data for recommendation engine |
| 3 | +""" |
| 4 | + |
| 5 | +import predictionio |
| 6 | +import argparse |
| 7 | +import random |
| 8 | +import datetime |
| 9 | +import pytz |
| 10 | +import dateutil.parser |
| 11 | + |
| 12 | +RATE_ACTIONS_DELIMITER = "," |
| 13 | +SEED = 1 |
| 14 | + |
| 15 | + |
| 16 | +def import_events(client, file): |
| 17 | + random.seed(SEED) |
| 18 | + count = 0 |
| 19 | + |
| 20 | + # year, month, day[, hour[, minute[, second[ |
| 21 | + #event_date = datetime.datetime(2015, 8, 13, 12, 24, 41) |
| 22 | + # These date values are chosen to match the multi-query-handmade.sh queries for testing purposes |
| 23 | + num_samples = 50 |
| 24 | + num_usage_event_lines = 17 |
| 25 | + start_date = dateutil.parser.parse("2015-08-12T11:28:45.114-07:00") |
| 26 | + item_date = start_date |
| 27 | + end_date = dateutil.parser.parse("2015-10-01T11:28:45.114-07:00") |
| 28 | + date_increment = (end_date - start_date) / (num_usage_event_lines * num_samples) |
| 29 | + event_date = item_date |
| 30 | + print "Date increment: " + str(date_increment) |
| 31 | + available_date = item_date + datetime.timedelta(days=-2) |
| 32 | + expire_date = item_date + datetime.timedelta(days=2) |
| 33 | + end = event_date.isoformat() |
| 34 | + start = event_date.isoformat() |
| 35 | + |
| 36 | + #date_increment = datetime.timedelta(minutes=1) |
| 37 | + #item_date = datetime.datetime.now(pytz.timezone("America/New_York")) |
| 38 | + # event_date = datetime.now(pytz.utc) |
| 39 | + print "Importing data." |
| 40 | + |
| 41 | + for it in range(1, num_samples, 1): |
| 42 | + f = open(file, 'r') |
| 43 | + for line in f: |
| 44 | + data = line.rstrip('\r\n').split(RATE_ACTIONS_DELIMITER) |
| 45 | + # For demonstration purpose action names are taken from input along with secondary actions on |
| 46 | + # For the UR add some item metadata |
| 47 | + user = "u" + str(random.randint(0, 1000)) |
| 48 | + if (data[1] == "purchase"): |
| 49 | + client.create_event( |
| 50 | + event=data[1], |
| 51 | + entity_type="user", |
| 52 | + entity_id=data[0], |
| 53 | + target_entity_type="item", |
| 54 | + target_entity_id=data[2], |
| 55 | + event_time=event_date |
| 56 | + ) |
| 57 | + print "Event: " + data[1] + " user: " + data[0] + " target_entity_id: " + data[2] + " eventDate: " + event_date.isoformat() |
| 58 | + elif (data[1] == "view"): # assumes other event type is 'view' |
| 59 | + client.create_event( |
| 60 | + event=data[1], |
| 61 | + entity_type="user", |
| 62 | + entity_id=data[0], |
| 63 | + target_entity_type="item", # type of item in this action |
| 64 | + target_entity_id=data[2], |
| 65 | + event_time=event_date |
| 66 | + ) |
| 67 | + print "Event: " + data[1] + " user: " + data[0] + " target_entity_id: " + data[2] + " eventDate: " + event_date.isoformat() |
| 68 | + event_date += date_increment |
| 69 | + |
| 70 | + f.close() |
| 71 | + |
| 72 | + #now set some properties once per items with various dates |
| 73 | + num_items = 5 |
| 74 | + start_date = dateutil.parser.parse("2015-08-12T11:28:45.114-07:00") |
| 75 | + item_date = start_date |
| 76 | + end_date = dateutil.parser.parse("2015-10-01T11:28:45.114-07:00") |
| 77 | + date_increment = (end_date - start_date)/num_items |
| 78 | + print "Date increment: " + str(date_increment) |
| 79 | + event_date = item_date |
| 80 | + available_date = item_date + datetime.timedelta(days=-2) |
| 81 | + expire_date = item_date + datetime.timedelta(days=2) |
| 82 | + end = event_date.isoformat() |
| 83 | + start = event_date.isoformat() |
| 84 | + |
| 85 | + |
| 86 | + f = open(file, 'r') |
| 87 | + count = 0 |
| 88 | + for line in f: |
| 89 | + data = line.rstrip('\r\n').split(RATE_ACTIONS_DELIMITER) |
| 90 | + if (data[1] == "$set"): # must be a set event |
| 91 | + date_choice = count % 3 |
| 92 | + print "data_choice = " + str(date_choice) |
| 93 | + if (date_choice == 2): # both bounds and date for daterange |
| 94 | + client.create_event( |
| 95 | + event=data[1], |
| 96 | + entity_type="item", |
| 97 | + entity_id=data[0], |
| 98 | + properties={"category": [data[2]], "expiredate": expire_date.isoformat(), |
| 99 | + "availabledate": available_date.isoformat(), "date": item_date.isoformat()}, |
| 100 | + event_time=event_date |
| 101 | + ) |
| 102 | + print "Event: " + data[1] + " item: " + data[0] + " properties/catagory: " + data[2] + " eventDate: " + \ |
| 103 | + event_date.isoformat() + \ |
| 104 | + " properties/availabledate: " + available_date.isoformat() + \ |
| 105 | + " properties/date: " + item_date.isoformat() + \ |
| 106 | + " properties/expiredate: " + expire_date.isoformat() |
| 107 | + elif (date_choice == 1): # available bound and date for daterange |
| 108 | + client.create_event( |
| 109 | + event=data[1], |
| 110 | + entity_type="item", |
| 111 | + entity_id=data[0], |
| 112 | + properties={"category": [data[2]], |
| 113 | + "availabledate": available_date.isoformat(), "date": item_date.isoformat()}, |
| 114 | + event_time=event_date |
| 115 | + ) |
| 116 | + print "Event: " + data[1] + " item: " + data[0] + " properties/catagory: " + data[2] + " eventDate: " + \ |
| 117 | + event_date.isoformat() + \ |
| 118 | + " properties/availabledate: " + available_date.isoformat() + \ |
| 119 | + " properties/date: " + item_date.isoformat() |
| 120 | + else: # expire bound and date for daterange |
| 121 | + client.create_event( |
| 122 | + event=data[1], |
| 123 | + entity_type="item", |
| 124 | + entity_id=data[0], |
| 125 | + properties={"category": [data[2]], "expiredate": expire_date.isoformat(), |
| 126 | + "date": item_date.isoformat()}, |
| 127 | + event_time=event_date |
| 128 | + ) |
| 129 | + print "Event: " + data[1] + " item: " + data[0] + " properties/catagory: " + data[2] + " eventDate: " + \ |
| 130 | + event_date.isoformat() + \ |
| 131 | + " properties/date: " + item_date.isoformat() + \ |
| 132 | + " properties/expiredate: " + expire_date.isoformat() |
| 133 | + count += 1 |
| 134 | + expire_date += date_increment |
| 135 | + item_date += date_increment |
| 136 | + event_date += date_increment |
| 137 | + available_date += date_increment |
| 138 | + f.close() |
| 139 | + print "%s events are imported." % count |
| 140 | + print "Start at " + start + " end at " + end + " with increment in seconds of " + str(date_increment.seconds) |
| 141 | + |
| 142 | + |
| 143 | +if __name__ == '__main__': |
| 144 | + parser = argparse.ArgumentParser( |
| 145 | + description="Import sample data for recommendation engine") |
| 146 | + parser.add_argument('--access_key', default='invald_access_key') |
| 147 | + parser.add_argument('--url', default="http://localhost:7070") |
| 148 | + parser.add_argument('--file', default="./data/sample-handmade-data.txt") |
| 149 | + |
| 150 | + args = parser.parse_args() |
| 151 | + print args |
| 152 | + |
| 153 | + client = predictionio.EventClient( |
| 154 | + access_key=args.access_key, |
| 155 | + url=args.url, |
| 156 | + threads=5, |
| 157 | + qsize=500) |
| 158 | + import_events(client, args.file) |
0 commit comments