Mercurial > public > sg101
view gpp/news/utils.py @ 552:9e42e6618168
For bitbucket issue #2, tweak the admin settings for the Post model to
reduce slow queries. Define our own queryset() method so we can control the
select_related(), and not have it cascade from post to topics to forums to
categories. Removed 'topic' from list_display because MySQL still sucked with
2 inner joins. Now it seems to be tolerable with only one join to User.
author | Brian Neal <bgneal@gmail.com> |
---|---|
date | Wed, 25 Jan 2012 20:07:03 -0600 |
parents | 24f1230f3ee3 |
children |
line wrap: on
line source
""" Common utility/helper code for the news app. """ from django.contrib.contenttypes.models import ContentType from comments.models import Comment from tagging.models import TaggedItem from news.models import Story def attach_extra_attrs(stories): """ For each story in the input stories list, attach 2 new attributes: tag_list and comment_count. The tags and comment count info is pulled from the database in bulk. This saves database queries when lots of news stories are displayed at once. For best results, use ".defer('tags')" when retrieve the stories from the database. """ stories_dict = dict((story.id, story) for story in stories) story_ids = stories_dict.keys() # Get all the tags out in one query ct = ContentType.objects.get_for_model(Story) tagged_items = TaggedItem.objects.filter(content_type=ct, object_id__in=story_ids).select_related('tag') for story in stories_dict.values(): story.tag_list = [] story.comment_count = 0 # attach tags for item in tagged_items: stories_dict[item.object_id].tag_list.append(item.tag.name) # Now get all the comment counts out in one fell swoop story_ids = Comment.objects.filter(content_type=ct, object_id__in=story_ids).values_list('object_id', flat=True) # compute comment_count for story_id in story_ids: stories_dict[story_id].comment_count += 1