def calcualte_useless_articles(cls):
from people.mixins import default_key_user
full_warm_seq_ids = default_key_user.get_full_dataset()
articles = cls.objects(seq_id__in=full_warm_seq_ids).only('id', 'seq_id', 'category', 'published_at', 'quality', 'hot')
line_article = cls.get_earliest_valid_obj()
outdate_articles = [article for article in articles if article.seq_id < line_article.seq_id]
usable_articles = set(articles) - set(outdate_articles)
useless_articles = outdate_articles
today = date.today()
now = datetime.utcnow()
valid_deadline = now - VALID_DURATION
update_time = get_global_cal_time()
for pubdate, _articles in groupby(sorted(useless_articles, key=lambda a: a.published_at.date()), lambda a: a.published_at.date()):
if pubdate >= now.date():
continue
__articles = list(_articles)
__aritlces = sorted(__articles, key=lambda _ar: calculate_scores(0, 0, _ar.published_at, _ar.quality, _ar.hot, 0)[0], reverse=True)
count = len(__articles)
if pubdate < valid_deadline.date():
useless_articles.extend(__articles)
else:
useless_articles.extend(__articles[count * 3 / 4:])
return useless_articles
def get_all_candidate_article_seq_ids(cls):
from people.mixins import default_key_user
return default_key_user.get_full_dataset()
