AlgoliaHN - Reddit ranking script

gistfile1.txt

from algoliasearch import algoliasearch
from datetime import datetime, timedelta
from math import log

# reddit hot score computing tools, courtesy of Mika S!
epoch = datetime(1970, 1, 1)
oldest_article = 1160418111


def bucket_date(date):
    res = date - timedelta(days=date.weekday())
    return res.replace(hour=0, minute=0, second=0, microsecond=0)


def epoch_seconds(date):
    td = date - epoch
    return td.days * 86400 + td.seconds + (float(td.microseconds) / 1000000)


def hot(score, date):
    try:
        order = log(score, 10)
    except:
        order = 0
    seconds = epoch_seconds(bucket_date(date)) - oldest_article
    return round(order + seconds / 45000, 7)


# Algolia tools
client = algoliasearch.Client('UJ5WYC0L7X', 'd693b2435c0d9429ce25e4fbd4ec6d41')
index = client.init_index('Item_production')
client2 = algoliasearch.Client('RSBCBF0EG8', '521ea9434ec24fec6767b6f40e8e8356')
index2 = client2.init_index('HN_Item_production')

to_transfer = []
total = 0
batch_size = 10000

# and there we go
res = index.browse_all({'query': '', 'attributesToRetrieve': ['*']})
for hit in res:
    hit['hot'] = hot(hit.get('points', 1), datetime.fromtimestamp(hit['created_at_i']))
    to_transfer.append(hit)

    if len(to_transfer) % batch_size == 0:
        index2.partial_update_objects(to_transfer)
        total += batch_size
        print('Updated %s products (total: %s)' % (batch_size, total))
        to_transfer = []

if len(to_transfer) > 0:
    index2.partial_update_objects(to_transfer)
    total += len(to_transfer)
    print('Updated %s products (total: %s)' % (len(to_transfer), total))