X-Git-Url: https://git.mdrn.pl/wolnelektury.git/blobdiff_plain/6ba1659e33c6aa1de178b5ab9a7a4cf27385e7a4..bd6eceb9cd8e613c289f612ade962535741e2699:/apps/newtagging/models.py diff --git a/apps/newtagging/models.py b/apps/newtagging/models.py index 6d05e287e..ead442f31 100644 --- a/apps/newtagging/models.py +++ b/apps/newtagging/models.py @@ -3,13 +3,7 @@ Models and managers for generic tagging. """ -# Python 2.3 compatibility -try: - set -except NameError: - from sets import Set as set - -from django.contrib.contenttypes import generic +from django.contrib.contenttypes.fields import GenericForeignKey from django.contrib.contenttypes.models import ContentType from django.db import connection, models from django.utils.translation import ugettext_lazy as _ @@ -19,11 +13,6 @@ from django.dispatch import Signal qn = connection.ops.quote_name -try: - from django.db.models.query import parse_lookup -except ImportError: - parse_lookup = None - tags_updated = Signal(providing_args=["affected_tags"]) @@ -101,64 +90,7 @@ class TagManager(models.Manager): return self.filter(items__content_type__pk=ctype.pk, items__object_id=obj.pk) - def _get_usage(self, model, counts=False, min_count=None, extra_joins=None, extra_criteria=None, params=None, extra=None): - """ - Perform the custom SQL query for ``usage_for_model`` and - ``usage_for_queryset``. - """ - if min_count is not None: counts = True - - model_table = qn(model._meta.db_table) - model_pk = '%s.%s' % (model_table, qn(model._meta.pk.column)) - tag_columns = self._get_tag_columns() - - if extra is None: extra = {} - extra_where = '' - if 'where' in extra: - extra_where = 'AND ' + ' AND '.join(extra['where']) - - query = """ - SELECT DISTINCT %(tag_columns)s%(count_sql)s - FROM - %(tag)s - INNER JOIN %(tagged_item)s - ON %(tag)s.id = %(tagged_item)s.tag_id - INNER JOIN %(model)s - ON %(tagged_item)s.object_id = %(model_pk)s - %%s - WHERE %(tagged_item)s.content_type_id = %(content_type_id)s - %%s - %(extra_where)s - GROUP BY %(tag_columns)s, %(tag)s.id, %(tag)s.name - %%s - ORDER BY %(tag)s.%(ordering)s ASC""" % { - 'tag': qn(self.model._meta.db_table), - 'ordering': ', '.join(qn(field) for field in self.model._meta.ordering), - 'tag_columns': tag_columns, - 'count_sql': counts and (', COUNT(%s)' % model_pk) or '', - 'tagged_item': qn(self.intermediary_table_model._meta.db_table), - 'model': model_table, - 'model_pk': model_pk, - 'extra_where': extra_where, - 'content_type_id': ContentType.objects.get_for_model(model).pk, - } - - min_count_sql = '' - if min_count is not None: - min_count_sql = 'HAVING COUNT(%s) >= %%s' % model_pk - params.append(min_count) - - cursor = connection.cursor() - cursor.execute(query % (extra_joins, extra_criteria, min_count_sql), params) - tags = [] - for row in cursor.fetchall(): - t = self.model(*row[:len(self.model._meta.fields)]) - if counts: - t.count = row[len(self.model._meta.fields)] - tags.append(t) - return tags - - def usage_for_model(self, model, counts=False, min_count=None, filters=None, extra=None): + def usage_for_model(self, model, counts=False, filters=None): """ Obtain a list of tags associated with instances of the given Model class. @@ -167,39 +99,21 @@ class TagManager(models.Manager): each tag, indicating how many times it has been used against the Model class in question. - If ``min_count`` is given, only tags which have a ``count`` - greater than or equal to ``min_count`` will be returned. - Passing a value for ``min_count`` implies ``counts=True``. - To limit the tags (and counts, if specified) returned to those used by a subset of the Model's instances, pass a dictionary of field lookups to be applied to the given Model as the ``filters`` argument. """ - if extra is None: extra = {} + # TODO: Do we really need this filters stuff? if filters is None: filters = {} - if not parse_lookup: - # post-queryset-refactor (hand off to usage_for_queryset) - queryset = model._default_manager.filter() - for f in filters.items(): - queryset.query.add_filter(f) - usage = self.usage_for_queryset(queryset, counts, min_count, extra) - else: - # pre-queryset-refactor - extra_joins = '' - extra_criteria = '' - params = [] - if len(filters) > 0: - joins, where, params = parse_lookup(filters.items(), model._meta) - extra_joins = ' '.join(['%s %s AS %s ON %s' % (join_type, table, alias, condition) - for (alias, (table, join_type, condition)) in joins.items()]) - extra_criteria = 'AND %s' % (' AND '.join(where)) - usage = self._get_usage(model, counts, min_count, extra_joins, extra_criteria, params, extra) - + queryset = model._default_manager.filter() + for f in filters.items(): + queryset.query.add_filter(f) + usage = self.usage_for_queryset(queryset, counts) return usage - def usage_for_queryset(self, queryset, counts=False, min_count=None, extra=None): + def usage_for_queryset(self, queryset, counts=False): """ Obtain a list of tags associated with instances of a model contained in the given queryset. @@ -207,33 +121,18 @@ class TagManager(models.Manager): If ``counts`` is True, a ``count`` attribute will be added to each tag, indicating how many times it has been used against the Model class in question. - - If ``min_count`` is given, only tags which have a ``count`` - greater than or equal to ``min_count`` will be returned. - Passing a value for ``min_count`` implies ``counts=True``. """ - if parse_lookup: - raise AttributeError("'TagManager.usage_for_queryset' is not compatible with pre-queryset-refactor versions of Django.") - - if getattr(queryset.query, 'get_compiler', None): - # Django 1.2+ - compiler = queryset.query.get_compiler(using='default') - extra_joins = ' '.join(compiler.get_from_clause()[0][1:]) - where, params = queryset.query.where.as_sql( - compiler.quote_name_unless_alias, compiler.connection + usage = self.model._default_manager.filter( + items__content_type=ContentType.objects.get_for_model(queryset.model), + items__object_id__in=queryset ) + if counts: + usage = usage.annotate(count=models.Count('id')) else: - # Django pre-1.2 - extra_joins = ' '.join(queryset.query.get_from_clause()[0][1:]) - where, params = queryset.query.where.as_sql() - - if where: - extra_criteria = 'AND %s' % where - else: - extra_criteria = '' - return self._get_usage(queryset.model, counts, min_count, extra_joins, extra_criteria, params, extra) + usage = usage.distinct() + return usage - def related_for_model(self, tags, model, counts=False, min_count=None, extra=None): + def related_for_model(self, tags, model, counts=False): """ Obtain a list of tags related to a given list of tags - that is, other tags used by items which have all the given tags. @@ -241,90 +140,14 @@ class TagManager(models.Manager): If ``counts`` is True, a ``count`` attribute will be added to each tag, indicating the number of items which have it in addition to the given list of tags. - - If ``min_count`` is given, only tags which have a ``count`` - greater than or equal to ``min_count`` will be returned. - Passing a value for ``min_count`` implies ``counts=True``. """ - if min_count is not None: counts = True - tags = self.model.get_tag_list(tags) - tag_count = len(tags) - tagged_item_table = qn(self.intermediary_table_model._meta.db_table) - tag_columns = self._get_tag_columns() - - if extra is None: extra = {} - extra_where = '' - if 'where' in extra: - extra_where = 'AND ' + ' AND '.join(extra['where']) - - # Temporary table in this query is a hack to prevent MySQL from executing - # inner query as dependant query (which could result in severe performance loss) - query = """ - SELECT %(tag_columns)s%(count_sql)s - FROM %(tagged_item)s INNER JOIN %(tag)s ON %(tagged_item)s.tag_id = %(tag)s.id - WHERE %(tagged_item)s.content_type_id = %(content_type_id)s - AND %(tagged_item)s.object_id IN - ( - SELECT * - FROM ( - SELECT %(tagged_item)s.object_id - FROM %(tagged_item)s, %(tag)s - WHERE %(tagged_item)s.content_type_id = %(content_type_id)s - AND %(tag)s.id = %(tagged_item)s.tag_id - AND %(tag)s.id IN (%(tag_id_placeholders)s) - GROUP BY %(tagged_item)s.object_id - HAVING COUNT(%(tagged_item)s.object_id) = %(tag_count)s - ) AS temporary - ) - AND %(tag)s.id NOT IN (%(tag_id_placeholders)s) - %(extra_where)s - GROUP BY %(tag_columns)s - %(min_count_sql)s - ORDER BY %(tag)s.%(ordering)s ASC""" % { - 'tag': qn(self.model._meta.db_table), - 'ordering': ', '.join(qn(field) for field in self.model._meta.ordering), - 'tag_columns': tag_columns, - 'count_sql': counts and ', COUNT(%s.object_id)' % tagged_item_table or '', - 'tagged_item': tagged_item_table, - 'content_type_id': ContentType.objects.get_for_model(model).pk, - 'tag_id_placeholders': ','.join(['%s'] * tag_count), - 'extra_where': extra_where, - 'tag_count': tag_count, - 'min_count_sql': min_count is not None and ('HAVING COUNT(%s.object_id) >= %%s' % tagged_item_table) or '', - } - - params = [tag.pk for tag in tags] * 2 - if min_count is not None: - params.append(min_count) - - cursor = connection.cursor() - cursor.execute(query, params) - related = [] - for row in cursor.fetchall(): - tag = self.model(*row[:len(self.model._meta.fields)]) - if counts is True: - tag.count = row[len(self.model._meta.fields)] - related.append(tag) - return related - - def _get_tag_columns(self): - tag_table = qn(self.model._meta.db_table) - return ', '.join('%s.%s' % (tag_table, qn(field.column)) for field in self.model._meta.fields) + objs = self.model.intermediary_table_model.objects.get_by_model(model, tags) + qs = self.usage_for_queryset(objs, counts) + qs = qs.exclude(pk__in=[tag.pk for tag in tags]) + return qs class TaggedItemManager(models.Manager): - """ - FIXME There's currently no way to get the ``GROUP BY`` and ``HAVING`` - SQL clauses required by many of this manager's methods into - Django's ORM. - - For now, we manually execute a query to retrieve the PKs of - objects we're interested in, then use the ORM's ``__in`` - lookup to return a ``QuerySet``. - - Once the queryset-refactor branch lands in trunk, this can be - tidied up significantly. - """ def __init__(self, tag_model): super(TaggedItemManager, self).__init__() self.tag_model = tag_model @@ -334,176 +157,41 @@ class TaggedItemManager(models.Manager): Create a ``QuerySet`` containing instances of the specified model associated with a given tag or list of tags. """ - tags = self.tag_model.get_tag_list(tags) - tag_count = len(tags) - if tag_count == 0: - # No existing tags were given - queryset, model = get_queryset_and_model(queryset_or_model) - return model._default_manager.none() - elif tag_count == 1: - # Optimisation for single tag - fall through to the simpler - # query below. - tag = tags[0] - else: - return self.get_intersection_by_model(queryset_or_model, tags) - queryset, model = get_queryset_and_model(queryset_or_model) - content_type = ContentType.objects.get_for_model(model) - opts = self.model._meta - tagged_item_table = qn(opts.db_table) - return queryset.extra( - tables=[opts.db_table], - where=[ - '%s.content_type_id = %%s' % tagged_item_table, - '%s.tag_id = %%s' % tagged_item_table, - '%s.%s = %s.object_id' % (qn(model._meta.db_table), - qn(model._meta.pk.column), - tagged_item_table) - ], - params=[content_type.pk, tag.pk], - ) - - def get_intersection_by_model(self, queryset_or_model, tags): - """ - Create a ``QuerySet`` containing instances of the specified - model associated with *all* of the given list of tags. - """ tags = self.tag_model.get_tag_list(tags) - tag_count = len(tags) - queryset, model = get_queryset_and_model(queryset_or_model) + if not tags: + # No existing tags were given + return queryset.none() - if not tag_count: - return model._default_manager.none() - - model_table = qn(model._meta.db_table) - # This query selects the ids of all objects which have all the - # given tags. - query = """ - SELECT %(model_pk)s - FROM %(model)s, %(tagged_item)s - WHERE %(tagged_item)s.content_type_id = %(content_type_id)s - AND %(tagged_item)s.tag_id IN (%(tag_id_placeholders)s) - AND %(model_pk)s = %(tagged_item)s.object_id - GROUP BY %(model_pk)s - HAVING COUNT(%(model_pk)s) = %(tag_count)s""" % { - 'model_pk': '%s.%s' % (model_table, qn(model._meta.pk.column)), - 'model': model_table, - 'tagged_item': qn(self.model._meta.db_table), - 'content_type_id': ContentType.objects.get_for_model(model).pk, - 'tag_id_placeholders': ','.join(['%s'] * tag_count), - 'tag_count': tag_count, - } - - cursor = connection.cursor() - cursor.execute(query, [tag.pk for tag in tags]) - object_ids = [row[0] for row in cursor.fetchall()] - if len(object_ids) > 0: - return queryset.filter(pk__in=object_ids) - else: - return model._default_manager.none() + # TODO: presumes reverse generic relation + # Multiple joins are WAY faster than having-count, at least on Postgres 9.1. + for tag in tags: + queryset = queryset.filter(tag_relations__tag=tag) + return queryset def get_union_by_model(self, queryset_or_model, tags): """ Create a ``QuerySet`` containing instances of the specified model associated with *any* of the given list of tags. """ - tags = self.tag_model.get_tag_list(tags) - tag_count = len(tags) queryset, model = get_queryset_and_model(queryset_or_model) + tags = self.tag_model.get_tag_list(tags) + if not tags: + return queryset.none() + # TODO: presumes reverse generic relation + return queryset.filter(tag_relations__tag__in=tags) - if not tag_count: - return model._default_manager.none() - - model_table = qn(model._meta.db_table) - # This query selects the ids of all objects which have any of - # the given tags. - query = """ - SELECT %(model_pk)s - FROM %(model)s, %(tagged_item)s - WHERE %(tagged_item)s.content_type_id = %(content_type_id)s - AND %(tagged_item)s.tag_id IN (%(tag_id_placeholders)s) - AND %(model_pk)s = %(tagged_item)s.object_id - GROUP BY %(model_pk)s""" % { - 'model_pk': '%s.%s' % (model_table, qn(model._meta.pk.column)), - 'model': model_table, - 'tagged_item': qn(self.model._meta.db_table), - 'content_type_id': ContentType.objects.get_for_model(model).pk, - 'tag_id_placeholders': ','.join(['%s'] * tag_count), - } - - cursor = connection.cursor() - cursor.execute(query, [tag.pk for tag in tags]) - object_ids = [row[0] for row in cursor.fetchall()] - if len(object_ids) > 0: - return queryset.filter(pk__in=object_ids) - else: - return model._default_manager.none() - - def get_related(self, obj, queryset_or_model, num=None, ignore_by_tag=None): + def get_related(self, obj, queryset_or_model): """ Retrieve a list of instances of the specified model which share tags with the model instance ``obj``, ordered by the number of shared tags in descending order. - - If ``num`` is given, a maximum of ``num`` instances will be - returned. - - If ``ignore_by_tag`` is given, object tagged with it will be ignored. """ queryset, model = get_queryset_and_model(queryset_or_model) - model_table = qn(model._meta.db_table) - content_type = ContentType.objects.get_for_model(obj) - related_content_type = ContentType.objects.get_for_model(model) - query = """ - SELECT %(model_pk)s, COUNT(related_tagged_item.object_id) AS %(count)s - FROM %(model)s, %(tagged_item)s, %(tag)s, %(tagged_item)s related_tagged_item - WHERE %(tagged_item)s.object_id = %%s - AND %(tagged_item)s.content_type_id = %(content_type_id)s - AND %(tag)s.id = %(tagged_item)s.tag_id - AND related_tagged_item.content_type_id = %(related_content_type_id)s - AND related_tagged_item.tag_id = %(tagged_item)s.tag_id - AND %(model_pk)s = related_tagged_item.object_id""" - if content_type.pk == related_content_type.pk: - # Exclude the given instance itself if determining related - # instances for the same model. - query += """ - AND related_tagged_item.object_id != %(tagged_item)s.object_id""" - if ignore_by_tag is not None: - query += """ - AND NOT EXISTS ( - SELECT * FROM %(tagged_item)s - WHERE %(tagged_item)s.object_id = %(model_pk)s - AND %(tagged_item)s.content_type_id = %(content_type_id)s - AND %(ignore_id)s = %(tagged_item)s.tag_id - ) - """ - query += """ - GROUP BY %(model_pk)s - ORDER BY %(count)s DESC - %(limit_offset)s""" - query = query % { - 'model_pk': '%s.%s' % (model_table, qn(model._meta.pk.column)), - 'count': qn('count'), - 'model': model_table, - 'tagged_item': qn(self.model._meta.db_table), - 'tag': qn(self.model._meta.get_field('tag').rel.to._meta.db_table), - 'content_type_id': content_type.pk, - 'related_content_type_id': related_content_type.pk, - 'limit_offset': num is not None and connection.ops.limit_offset_sql(num) or '', - 'ignore_id': ignore_by_tag.id if ignore_by_tag else None, - } - - cursor = connection.cursor() - cursor.execute(query, [obj.pk]) - object_ids = [row[0] for row in cursor.fetchall()] - if len(object_ids) > 0: - # Use in_bulk here instead of an id__in lookup, because id__in would - # clobber the ordering. - object_dict = queryset.in_bulk(object_ids) - return [object_dict[object_id] for object_id in object_ids \ - if object_id in object_dict] - else: - return [] + # TODO: presumes reverse generic relation. + # Do we know it's 'tags'? + return queryset.filter(tag_relations__tag__in=obj.tags).annotate( + count=models.Count('pk')).order_by('-count').exclude(pk=obj.pk) ########## @@ -516,6 +204,7 @@ def create_intermediary_table_model(model): class Meta: db_table = '%s_relation' % model._meta.db_table unique_together = (('tag', 'content_type', 'object_id'),) + app_label = model._meta.app_label def obj_unicode(self): try: @@ -530,7 +219,7 @@ def create_intermediary_table_model(model): 'tag': models.ForeignKey(model, verbose_name=_('tag'), related_name='items'), 'content_type': models.ForeignKey(ContentType, verbose_name=_('content type')), 'object_id': models.PositiveIntegerField(_('object id'), db_index=True), - 'content_object': generic.GenericForeignKey('content_type', 'object_id'), + 'content_object': GenericForeignKey('content_type', 'object_id'), '__unicode__': obj_unicode, } @@ -539,8 +228,8 @@ def create_intermediary_table_model(model): class TagMeta(ModelBase): "Metaclass for tag models (models inheriting from TagBase)." - def __new__(cls, name, bases, attrs): - model = super(TagMeta, cls).__new__(cls, name, bases, attrs) + def __new__(mcs, name, bases, attrs): + model = super(TagMeta, mcs).__new__(mcs, name, bases, attrs) if not model._meta.abstract: # Create an intermediary table and register custom managers for concrete models model.intermediary_table_model = create_intermediary_table_model(model)