|
- """
- Various data structures used in query construction.
-
- Factored out from django.db.models.query to avoid making the main module very
- large and/or so that they can be used by other modules without getting into
- circular import difficulties.
- """
- from __future__ import unicode_literals
-
- import inspect
- from collections import namedtuple
-
- from django.core.exceptions import FieldDoesNotExist
- from django.db.backends import utils
- from django.db.models.constants import LOOKUP_SEP
- from django.utils import tree
-
- # PathInfo is used when converting lookups (fk__somecol). The contents
- # describe the relation in Model terms (model Options and Fields for both
- # sides of the relation. The join_field is the field backing the relation.
- PathInfo = namedtuple('PathInfo', 'from_opts to_opts target_fields join_field m2m direct')
-
-
- class InvalidQuery(Exception):
- """
- The query passed to raw isn't a safe query to use with raw.
- """
- pass
-
-
- class QueryWrapper(object):
- """
- A type that indicates the contents are an SQL fragment and the associate
- parameters. Can be used to pass opaque data to a where-clause, for example.
- """
- contains_aggregate = False
-
- def __init__(self, sql, params):
- self.data = sql, list(params)
-
- def as_sql(self, compiler=None, connection=None):
- return self.data
-
-
- class Q(tree.Node):
- """
- Encapsulates filters as objects that can then be combined logically (using
- `&` and `|`).
- """
- # Connection types
- AND = 'AND'
- OR = 'OR'
- default = AND
-
- def __init__(self, *args, **kwargs):
- super(Q, self).__init__(children=list(args) + list(kwargs.items()))
-
- def _combine(self, other, conn):
- if not isinstance(other, Q):
- raise TypeError(other)
- obj = type(self)()
- obj.connector = conn
- obj.add(self, conn)
- obj.add(other, conn)
- return obj
-
- def __or__(self, other):
- return self._combine(other, self.OR)
-
- def __and__(self, other):
- return self._combine(other, self.AND)
-
- def __invert__(self):
- obj = type(self)()
- obj.add(self, self.AND)
- obj.negate()
- return obj
-
- def clone(self):
- clone = self.__class__._new_instance(
- children=[], connector=self.connector, negated=self.negated)
- for child in self.children:
- if hasattr(child, 'clone'):
- clone.children.append(child.clone())
- else:
- clone.children.append(child)
- return clone
-
- def resolve_expression(self, query=None, allow_joins=True, reuse=None, summarize=False, for_save=False):
- # We must promote any new joins to left outer joins so that when Q is
- # used as an expression, rows aren't filtered due to joins.
- clause, joins = query._add_q(self, reuse, allow_joins=allow_joins, split_subq=False)
- query.promote_joins(joins)
- return clause
-
- @classmethod
- def _refs_aggregate(cls, obj, existing_aggregates):
- if not isinstance(obj, tree.Node):
- aggregate, aggregate_lookups = refs_aggregate(obj[0].split(LOOKUP_SEP), existing_aggregates)
- if not aggregate and hasattr(obj[1], 'refs_aggregate'):
- return obj[1].refs_aggregate(existing_aggregates)
- return aggregate, aggregate_lookups
- for c in obj.children:
- aggregate, aggregate_lookups = cls._refs_aggregate(c, existing_aggregates)
- if aggregate:
- return aggregate, aggregate_lookups
- return False, ()
-
- def refs_aggregate(self, existing_aggregates):
- if not existing_aggregates:
- return False
-
- return self._refs_aggregate(self, existing_aggregates)
-
-
- class DeferredAttribute(object):
- """
- A wrapper for a deferred-loading field. When the value is read from this
- object the first time, the query is executed.
- """
- def __init__(self, field_name, model):
- self.field_name = field_name
-
- def __get__(self, instance, owner):
- """
- Retrieves and caches the value from the datastore on the first lookup.
- Returns the cached value.
- """
- non_deferred_model = instance._meta.proxy_for_model
- opts = non_deferred_model._meta
-
- assert instance is not None
- data = instance.__dict__
- if data.get(self.field_name, self) is self:
- # self.field_name is the attname of the field, but only() takes the
- # actual name, so we need to translate it here.
- try:
- f = opts.get_field(self.field_name)
- except FieldDoesNotExist:
- f = [f for f in opts.fields if f.attname == self.field_name][0]
- name = f.name
- # Let's see if the field is part of the parent chain. If so we
- # might be able to reuse the already loaded value. Refs #18343.
- val = self._check_parent_chain(instance, name)
- if val is None:
- instance.refresh_from_db(fields=[self.field_name])
- val = getattr(instance, self.field_name)
- data[self.field_name] = val
- return data[self.field_name]
-
- def __set__(self, instance, value):
- """
- Deferred loading attributes can be set normally (which means there will
- never be a database lookup involved.
- """
- instance.__dict__[self.field_name] = value
-
- def _check_parent_chain(self, instance, name):
- """
- Check if the field value can be fetched from a parent field already
- loaded in the instance. This can be done if the to-be fetched
- field is a primary key field.
- """
- opts = instance._meta
- f = opts.get_field(name)
- link_field = opts.get_ancestor_link(f.model)
- if f.primary_key and f != link_field:
- return getattr(instance, link_field.attname)
- return None
-
-
- class RegisterLookupMixin(object):
- def _get_lookup(self, lookup_name):
- try:
- return self.class_lookups[lookup_name]
- except KeyError:
- # To allow for inheritance, check parent class' class_lookups.
- for parent in inspect.getmro(self.__class__):
- if 'class_lookups' not in parent.__dict__:
- continue
- if lookup_name in parent.class_lookups:
- return parent.class_lookups[lookup_name]
- except AttributeError:
- # This class didn't have any class_lookups
- pass
- return None
-
- def get_lookup(self, lookup_name):
- from django.db.models.lookups import Lookup
- found = self._get_lookup(lookup_name)
- if found is None and hasattr(self, 'output_field'):
- return self.output_field.get_lookup(lookup_name)
- if found is not None and not issubclass(found, Lookup):
- return None
- return found
-
- def get_transform(self, lookup_name):
- from django.db.models.lookups import Transform
- found = self._get_lookup(lookup_name)
- if found is None and hasattr(self, 'output_field'):
- return self.output_field.get_transform(lookup_name)
- if found is not None and not issubclass(found, Transform):
- return None
- return found
-
- @classmethod
- def register_lookup(cls, lookup, lookup_name=None):
- if lookup_name is None:
- lookup_name = lookup.lookup_name
- if 'class_lookups' not in cls.__dict__:
- cls.class_lookups = {}
- cls.class_lookups[lookup_name] = lookup
- return lookup
-
- @classmethod
- def _unregister_lookup(cls, lookup, lookup_name=None):
- """
- Remove given lookup from cls lookups. For use in tests only as it's
- not thread-safe.
- """
- if lookup_name is None:
- lookup_name = lookup.lookup_name
- del cls.class_lookups[lookup_name]
-
-
- def select_related_descend(field, restricted, requested, load_fields, reverse=False):
- """
- Returns True if this field should be used to descend deeper for
- select_related() purposes. Used by both the query construction code
- (sql.query.fill_related_selections()) and the model instance creation code
- (query.get_klass_info()).
-
- Arguments:
- * field - the field to be checked
- * restricted - a boolean field, indicating if the field list has been
- manually restricted using a requested clause)
- * requested - The select_related() dictionary.
- * load_fields - the set of fields to be loaded on this model
- * reverse - boolean, True if we are checking a reverse select related
- """
- if not field.remote_field:
- return False
- if field.remote_field.parent_link and not reverse:
- return False
- if restricted:
- if reverse and field.related_query_name() not in requested:
- return False
- if not reverse and field.name not in requested:
- return False
- if not restricted and field.null:
- return False
- if load_fields:
- if field.attname not in load_fields:
- if restricted and field.name in requested:
- raise InvalidQuery("Field %s.%s cannot be both deferred"
- " and traversed using select_related"
- " at the same time." %
- (field.model._meta.object_name, field.name))
- return False
- return True
-
-
- # This function is needed because data descriptors must be defined on a class
- # object, not an instance, to have any effect.
-
- def deferred_class_factory(model, attrs):
- """
- Returns a class object that is a copy of "model" with the specified "attrs"
- being replaced with DeferredAttribute objects. The "pk_value" ties the
- deferred attributes to a particular instance of the model.
- """
- if not attrs:
- return model
- opts = model._meta
- # Never create deferred models based on deferred model
- if model._deferred:
- # Deferred models are proxies for the non-deferred model. We never
- # create chains of defers => proxy_for_model is the non-deferred
- # model.
- model = opts.proxy_for_model
- # The app registry wants a unique name for each model, otherwise the new
- # class won't be created (we get an exception). Therefore, we generate
- # the name using the passed in attrs. It's OK to reuse an existing class
- # object if the attrs are identical.
- name = "%s_Deferred_%s" % (model.__name__, '_'.join(sorted(attrs)))
- name = utils.truncate_name(name, 80, 32)
-
- try:
- return opts.apps.get_model(model._meta.app_label, name)
-
- except LookupError:
-
- class Meta:
- proxy = True
- apps = opts.apps
- app_label = opts.app_label
-
- overrides = {attr: DeferredAttribute(attr, model) for attr in attrs}
- overrides["Meta"] = Meta
- overrides["__module__"] = model.__module__
- overrides["_deferred"] = True
- return type(str(name), (model,), overrides)
-
-
- # The above function is also used to unpickle model instances with deferred
- # fields.
- deferred_class_factory.__safe_for_unpickling__ = True
-
-
- def refs_aggregate(lookup_parts, aggregates):
- """
- A helper method to check if the lookup_parts contains references
- to the given aggregates set. Because the LOOKUP_SEP is contained in the
- default annotation names we must check each prefix of the lookup_parts
- for a match.
- """
- for n in range(len(lookup_parts) + 1):
- level_n_lookup = LOOKUP_SEP.join(lookup_parts[0:n])
- if level_n_lookup in aggregates and aggregates[level_n_lookup].contains_aggregate:
- return aggregates[level_n_lookup], lookup_parts[n:]
- return False, ()
-
-
- def refs_expression(lookup_parts, annotations):
- """
- A helper method to check if the lookup_parts contains references
- to the given annotations set. Because the LOOKUP_SEP is contained in the
- default annotation names we must check each prefix of the lookup_parts
- for a match.
- """
- for n in range(len(lookup_parts) + 1):
- level_n_lookup = LOOKUP_SEP.join(lookup_parts[0:n])
- if level_n_lookup in annotations and annotations[level_n_lookup]:
- return annotations[level_n_lookup], lookup_parts[n:]
- return False, ()
-
-
- def check_rel_lookup_compatibility(model, target_opts, field):
- """
- Check that self.model is compatible with target_opts. Compatibility
- is OK if:
- 1) model and opts match (where proxy inheritance is removed)
- 2) model is parent of opts' model or the other way around
- """
- def check(opts):
- return (
- model._meta.concrete_model == opts.concrete_model or
- opts.concrete_model in model._meta.get_parent_list() or
- model in opts.get_parent_list()
- )
- # If the field is a primary key, then doing a query against the field's
- # model is ok, too. Consider the case:
- # class Restaurant(models.Model):
- # place = OnetoOneField(Place, primary_key=True):
- # Restaurant.objects.filter(pk__in=Restaurant.objects.all()).
- # If we didn't have the primary key check, then pk__in (== place__in) would
- # give Place's opts as the target opts, but Restaurant isn't compatible
- # with that. This logic applies only to primary keys, as when doing __in=qs,
- # we are going to turn this into __in=qs.values('pk') later on.
- return (
- check(target_opts) or
- (getattr(field, 'primary_key', False) and check(field.model._meta))
- )
|