Add LazyReferenceField

docs/changelog.rst

@@ -2,9 +2,9 @@
 Changelog
 =========
 
-Development
+Changes in 0.15.0
-===========
+=================
-- (Fill this out as you fix issues and develop your features).
+- Add LazyReferenceField to address #1230
 
 Changes in 0.14.1
 =================

mongoengine/base/__init__.py

@@ -15,7 +15,7 @@ __all__ = (
     'UPDATE_OPERATORS', '_document_registry', 'get_document',
 
     # datastructures
-    'BaseDict', 'BaseList', 'EmbeddedDocumentList',
+    'BaseDict', 'BaseList', 'EmbeddedDocumentList', 'LazyReference',
 
     # document
     'BaseDocument',

mongoengine/base/datastructures.py

@@ -2,11 +2,12 @@ import itertools
 import weakref
 
 import six
+from bson import DBRef
 
 from mongoengine.common import _import_class
 from mongoengine.errors import DoesNotExist, MultipleObjectsReturned
 
-__all__ = ('BaseDict', 'BaseList', 'EmbeddedDocumentList')
+__all__ = ('BaseDict', 'BaseList', 'EmbeddedDocumentList', 'LazyReference')
 
 
 class BaseDict(dict):
@@ -445,3 +446,42 @@ class StrictDict(object):
 
             cls._classes[allowed_keys] = SpecificStrictDict
         return cls._classes[allowed_keys]
+
+
+class LazyReference(DBRef):
+    __slots__ = ('_cached_doc', 'passthrough', 'document_type')
+
+    def fetch(self, force=False):
+        if not self._cached_doc or force:
+            self._cached_doc = self.document_type.objects.get(pk=self.pk)
+            if not self._cached_doc:
+                raise DoesNotExist('Trying to dereference unknown document %s' % (self))
+        return self._cached_doc
+
+    @property
+    def pk(self):
+        return self.id
+
+    def __init__(self, document_type, pk, cached_doc=None, passthrough=False):
+        self.document_type = document_type
+        self._cached_doc = cached_doc
+        self.passthrough = passthrough
+        super(LazyReference, self).__init__(self.document_type._get_collection_name(), pk)
+
+    def __getitem__(self, name):
+        if not self.passthrough:
+            raise KeyError()
+        document = self.fetch()
+        return document[name]
+
+    def __getattr__(self, name):
+        if not object.__getattribute__(self, 'passthrough'):
+            raise AttributeError()
+        document = self.fetch()
+        try:
+            return document[name]
+        except KeyError:
+            raise AttributeError()
+
+    def __repr__(self):
+        return "<LazyReference(%s, %r)>" % (self.document_type, self.pk)

mongoengine/fields.py

@@ -26,7 +26,8 @@ except ImportError:
     Int64 = long
 
 from mongoengine.base import (BaseDocument, BaseField, ComplexBaseField,
-                              GeoJsonBaseField, ObjectIdField, get_document)
+                              GeoJsonBaseField, ObjectIdField, get_document,
+                              LazyReference)
 from mongoengine.connection import DEFAULT_CONNECTION_NAME, get_db
 from mongoengine.document import Document, EmbeddedDocument
 from mongoengine.errors import DoesNotExist, InvalidQueryError, ValidationError
@@ -46,6 +47,8 @@ __all__ = (
     'GenericEmbeddedDocumentField', 'DynamicField', 'ListField',
     'SortedListField', 'EmbeddedDocumentListField', 'DictField',
     'MapField', 'ReferenceField', 'CachedReferenceField',
+    'LazyReferenceField',
+    # 'GenericLazyReferenceField',
     'GenericReferenceField', 'BinaryField', 'GridFSError', 'GridFSProxy',
     'FileField', 'ImageGridFsProxy', 'ImproperlyConfigured', 'ImageField',
     'GeoPointField', 'PointField', 'LineStringField', 'PolygonField',
@@ -953,6 +956,15 @@ class ReferenceField(BaseField):
     """A reference to a document that will be automatically dereferenced on
     access (lazily).
 
+    Note this means you will get a database I/O access everytime you access
+    this field. This is necessary because the field returns a :class:`~mongoengine.Document`
+    which precise type can depend of the value of the `_cls` field present in the
+    document in database.
+    In short, using this type of field can lead to poor performances (especially
+    if you access this field only to retrieve it `pk` field which is already
+    known before dereference). To solve this you should consider using the
+    :class:`~mongoengine.fields.LazyReferenceField`.
+
     Use the `reverse_delete_rule` to handle what should happen if the document
     the field is referencing is deleted.  EmbeddedDocuments, DictFields and
     MapFields does not support reverse_delete_rule and an `InvalidDocumentError`
@@ -1087,8 +1099,8 @@ class ReferenceField(BaseField):
 
     def validate(self, value):
 
-        if not isinstance(value, (self.document_type, DBRef, ObjectId)):
+        if not isinstance(value, (self.document_type, LazyReference, DBRef, ObjectId)):
-            self.error('A ReferenceField only accepts DBRef, ObjectId or documents')
+            self.error('A ReferenceField only accepts DBRef, LazyReference, ObjectId or documents')
 
         if isinstance(value, Document) and value.id is None:
             self.error('You can only reference documents once they have been '
@@ -2141,3 +2153,124 @@ class MultiPolygonField(GeoJsonBaseField):
     .. versionadded:: 0.9
     """
     _type = 'MultiPolygon'
+
+
+class LazyReferenceField(BaseField):
+    """A really lazy reference to a document.
+    Unlike the :class:`~mongoengine.fields.ReferenceField` it must be manually
+    dereferenced using it ``fetch()`` method.
+    """
+
+    def __init__(self, document_type, passthrough=False, dbref=False,
+                 reverse_delete_rule=DO_NOTHING, **kwargs):
+        """Initialises the Reference Field.
+
+        :param dbref:  Store the reference as :class:`~pymongo.dbref.DBRef`
+          or as the :class:`~pymongo.objectid.ObjectId`.id .
+        :param reverse_delete_rule: Determines what to do when the referring
+          object is deleted
+        :param passthrough: When trying to access unknown fields, the
+        :class:`~mongoengine.base.datastructure.LazyReference` instance will
+        automatically call `fetch()` and try to retrive the field on the fetched
+        document. Note this only work getting field (not setting or deleting).
+        """
+        if (
+            not isinstance(document_type, six.string_types) and
+            not issubclass(document_type, Document)
+        ):
+            self.error('Argument to LazyReferenceField constructor must be a '
+                       'document class or a string')
+
+        self.dbref = dbref
+        self.passthrough = passthrough
+        self.document_type_obj = document_type
+        self.reverse_delete_rule = reverse_delete_rule
+        super(LazyReferenceField, self).__init__(**kwargs)
+
+    @property
+    def document_type(self):
+        if isinstance(self.document_type_obj, six.string_types):
+            if self.document_type_obj == RECURSIVE_REFERENCE_CONSTANT:
+                self.document_type_obj = self.owner_document
+            else:
+                self.document_type_obj = get_document(self.document_type_obj)
+        return self.document_type_obj
+
+    def __get__(self, instance, owner):
+        """Descriptor to allow lazy dereferencing."""
+        if instance is None:
+            # Document class being used rather than a document object
+            return self
+
+        value = instance._data.get(self.name)
+        if isinstance(value, LazyReference):
+            if value.passthrough != self.passthrough:
+                instance._data[self.name] = LazyReference(
+                    value.document_type, value.pk, passthrough=self.passthrough)
+        elif value is not None:
+            if isinstance(value, self.document_type):
+                value = LazyReference(self.document_type, value.pk, passthrough=self.passthrough)
+            elif isinstance(value, DBRef):
+                value = LazyReference(self.document_type, value.id, passthrough=self.passthrough)
+            else:
+                # value is the primary key of the referenced document
+                value = LazyReference(self.document_type, value, passthrough=self.passthrough)
+            instance._data[self.name] = value
+
+        return super(LazyReferenceField, self).__get__(instance, owner)
+
+    def to_mongo(self, value):
+        if isinstance(value, LazyReference):
+            pk = value.pk
+        elif isinstance(value, self.document_type):
+            pk = value.pk
+        elif isinstance(value, DBRef):
+            pk = value.id
+        else:
+            # value is the primary key of the referenced document
+            pk = value
+        id_field_name = self.document_type._meta['id_field']
+        id_field = self.document_type._fields[id_field_name]
+        pk = id_field.to_mongo(pk)
+        if self.dbref:
+            return DBRef(self.document_type._get_collection_name(), pk)
+        else:
+            return pk
+
+    def validate(self, value):
+        if isinstance(value, LazyReference):
+            if not issubclass(value.document_type, self.document_type):
+                self.error('Reference must be on a `%s` document.' % self.document_type)
+            pk = value.pk
+        elif isinstance(value, self.document_type):
+            pk = value.pk
+        elif isinstance(value, DBRef):
+            # TODO: check collection ?
+            collection = self.document_type._get_collection_name()
+            if value.collection != collection:
+                self.error("DBRef on bad collection (must be on `%s`)" % collection)
+            pk = value.id
+        else:
+            # value is the primary key of the referenced document
+            id_field_name = self.document_type._meta['id_field']
+            id_field = getattr(self.document_type, id_field_name)
+            pk = value
+            try:
+                id_field.validate(pk)
+            except ValidationError:
+                self.error("value should be `{0}` document, LazyReference or DBRef on `{0}` "
+                           "or `{0}`'s primary key (i.e. `{1}`)".format(
+                            self.document_type.__name__, type(id_field).__name__))
+
+        if pk is None:
+            self.error('You can only reference documents once they have been '
+                       'saved to the database')
+
+    def prepare_query_value(self, op, value):
+        if value is None:
+            return None
+        super(LazyReferenceField, self).prepare_query_value(op, value)
+        return self.to_mongo(value)
+
+    def lookup_member(self, member_name):
+        return self.document_type._fields.get(member_name)

tests/fields/fields.py

@@ -26,7 +26,7 @@ except ImportError:
 from mongoengine import *
 from mongoengine.connection import get_db
 from mongoengine.base import (BaseDict, BaseField, EmbeddedDocumentList,
-                              _document_registry)
+                              _document_registry, LazyReference)
 
 from tests.utils import MongoDBTestCase
 
@@ -931,7 +931,9 @@ class FieldTest(MongoDBTestCase):
             comments = ListField(EmbeddedDocumentField(Comment))
             tags = ListField(StringField())
             authors = ListField(ReferenceField(User))
+            authors_as_lazy = ListField(LazyReferenceField(User))
             generic = ListField(GenericReferenceField())
+            # generic_as_lazy = ListField(LazyGenericReferenceField())
 
         User.drop_collection()
         BlogPost.drop_collection()
@@ -969,6 +971,15 @@ class FieldTest(MongoDBTestCase):
         post.authors = [user]
         post.validate()
 
+        post.authors_as_lazy = [Comment()]
+        self.assertRaises(ValidationError, post.validate)
+
+        post.authors_as_lazy = [User()]
+        self.assertRaises(ValidationError, post.validate)
+
+        post.authors_as_lazy = [user]
+        post.validate()
+
         post.generic = [1, 2]
         self.assertRaises(ValidationError, post.validate)
 
@@ -981,6 +992,18 @@ class FieldTest(MongoDBTestCase):
         post.generic = [user]
         post.validate()
 
+        # post.generic_as_lazy = [1, 2]
+        # self.assertRaises(ValidationError, post.validate)
+
+        # post.generic_as_lazy = [User(), Comment()]
+        # self.assertRaises(ValidationError, post.validate)
+
+        # post.generic_as_lazy = [Comment()]
+        # self.assertRaises(ValidationError, post.validate)
+
+        # post.generic_as_lazy = [user]
+        # post.validate()
+
     def test_sorted_list_sorting(self):
         """Ensure that a sorted list field properly sorts values.
         """
@@ -4598,5 +4621,256 @@ class CachedReferenceFieldTest(MongoDBTestCase):
         self.assertTrue(isinstance(ocorrence.animal, Animal))
 
 
+class LazyReferenceFieldTest(MongoDBTestCase):
+    def test_lazy_reference_config(self):
+        # Make sure ReferenceField only accepts a document class or a string
+        # with a document class name.
+        self.assertRaises(ValidationError, LazyReferenceField, EmbeddedDocument)
+
+    def test_lazy_reference_simple(self):
+        class Animal(Document):
+            name = StringField()
+            tag = StringField()
+
+        class Ocurrence(Document):
+            person = StringField()
+            animal = LazyReferenceField(Animal)
+
+        Animal.drop_collection()
+        Ocurrence.drop_collection()
+
+        animal = Animal(name="Leopard", tag="heavy").save()
+        Ocurrence(person="test", animal=animal).save()
+        p = Ocurrence.objects.get()
+        self.assertIsInstance(p.animal, LazyReference)
+        fetched_animal = p.animal.fetch()
+        self.assertEqual(fetched_animal, animal)
+        # `fetch` keep cache on referenced document by default...
+        animal.tag = "not so heavy"
+        animal.save()
+        double_fetch = p.animal.fetch()
+        self.assertIs(fetched_animal, double_fetch)
+        self.assertEqual(double_fetch.tag, "heavy")
+        # ...unless specified otherwise
+        fetch_force = p.animal.fetch(force=True)
+        self.assertIsNot(fetch_force, fetched_animal)
+        self.assertEqual(fetch_force.tag, "not so heavy")
+
+    def test_lazy_reference_fetch_invalid_ref(self):
+        class Animal(Document):
+            name = StringField()
+            tag = StringField()
+
+        class Ocurrence(Document):
+            person = StringField()
+            animal = LazyReferenceField(Animal)
+
+        Animal.drop_collection()
+        Ocurrence.drop_collection()
+
+        animal = Animal(name="Leopard", tag="heavy").save()
+        Ocurrence(person="test", animal=animal).save()
+        animal.delete()
+        p = Ocurrence.objects.get()
+        self.assertIsInstance(p.animal, LazyReference)
+        with self.assertRaises(DoesNotExist):
+            p.animal.fetch()
+
+    def test_lazy_reference_set(self):
+        class Animal(Document):
+            meta = {'allow_inheritance': True}
+
+            name = StringField()
+            tag = StringField()
+
+        class Ocurrence(Document):
+            person = StringField()
+            animal = LazyReferenceField(Animal)
+
+        Animal.drop_collection()
+        Ocurrence.drop_collection()
+
+        class SubAnimal(Animal):
+            nick = StringField()
+
+        animal = Animal(name="Leopard", tag="heavy").save()
+        sub_animal = SubAnimal(nick='doggo', name='dog').save()
+        for ref in (
+                animal,
+                animal.pk,
+                DBRef(animal._get_collection_name(), animal.pk),
+                LazyReference(Animal, animal.pk),
+
+                sub_animal,
+                sub_animal.pk,
+                DBRef(sub_animal._get_collection_name(), sub_animal.pk),
+                LazyReference(SubAnimal, sub_animal.pk),
+                ):
+            p = Ocurrence(person="test", animal=ref).save()
+            p.reload()
+            self.assertIsInstance(p.animal, LazyReference)
+            p.animal.fetch()
+
+    def test_lazy_reference_bad_set(self):
+        class Animal(Document):
+            name = StringField()
+            tag = StringField()
+
+        class Ocurrence(Document):
+            person = StringField()
+            animal = LazyReferenceField(Animal)
+
+        Animal.drop_collection()
+        Ocurrence.drop_collection()
+
+        class BadDoc(Document):
+            pass
+
+        animal = Animal(name="Leopard", tag="heavy").save()
+        baddoc = BadDoc().save()
+        for bad in (
+                42,
+                'foo',
+                baddoc,
+                DBRef(baddoc._get_collection_name(), animal.pk),
+                LazyReference(BadDoc, animal.pk)
+                ):
+            with self.assertRaises(ValidationError):
+                p = Ocurrence(person="test", animal=bad).save()
+
+    def test_lazy_reference_query_conversion(self):
+        """Ensure that LazyReferenceFields can be queried using objects and values
+        of the type of the primary key of the referenced object.
+        """
+        class Member(Document):
+            user_num = IntField(primary_key=True)
+
+        class BlogPost(Document):
+            title = StringField()
+            author = LazyReferenceField(Member, dbref=False)
+
+        Member.drop_collection()
+        BlogPost.drop_collection()
+
+        m1 = Member(user_num=1)
+        m1.save()
+        m2 = Member(user_num=2)
+        m2.save()
+
+        post1 = BlogPost(title='post 1', author=m1)
+        post1.save()
+
+        post2 = BlogPost(title='post 2', author=m2)
+        post2.save()
+
+        post = BlogPost.objects(author=m1).first()
+        self.assertEqual(post.id, post1.id)
+
+        post = BlogPost.objects(author=m2).first()
+        self.assertEqual(post.id, post2.id)
+
+        # Same thing by passing a LazyReference instance
+        post = BlogPost.objects(author=LazyReference(Member, m2.pk)).first()
+        self.assertEqual(post.id, post2.id)
+
+    def test_lazy_reference_query_conversion_dbref(self):
+        """Ensure that LazyReferenceFields can be queried using objects and values
+        of the type of the primary key of the referenced object.
+        """
+        class Member(Document):
+            user_num = IntField(primary_key=True)
+
+        class BlogPost(Document):
+            title = StringField()
+            author = LazyReferenceField(Member, dbref=True)
+
+        Member.drop_collection()
+        BlogPost.drop_collection()
+
+        m1 = Member(user_num=1)
+        m1.save()
+        m2 = Member(user_num=2)
+        m2.save()
+
+        post1 = BlogPost(title='post 1', author=m1)
+        post1.save()
+
+        post2 = BlogPost(title='post 2', author=m2)
+        post2.save()
+
+        post = BlogPost.objects(author=m1).first()
+        self.assertEqual(post.id, post1.id)
+
+        post = BlogPost.objects(author=m2).first()
+        self.assertEqual(post.id, post2.id)
+
+        # Same thing by passing a LazyReference instance
+        post = BlogPost.objects(author=LazyReference(Member, m2.pk)).first()
+        self.assertEqual(post.id, post2.id)
+
+    def test_lazy_reference_passthrough(self):
+        class Animal(Document):
+            name = StringField()
+            tag = StringField()
+
+        class Ocurrence(Document):
+            animal = LazyReferenceField(Animal, passthrough=False)
+            animal_passthrough = LazyReferenceField(Animal, passthrough=True)
+
+        Animal.drop_collection()
+        Ocurrence.drop_collection()
+
+        animal = Animal(name="Leopard", tag="heavy").save()
+        Ocurrence(animal=animal, animal_passthrough=animal).save()
+        p = Ocurrence.objects.get()
+        self.assertIsInstance(p.animal, LazyReference)
+        with self.assertRaises(KeyError):
+            p.animal['name']
+        with self.assertRaises(AttributeError):
+            p.animal.name
+        self.assertEqual(p.animal.pk, animal.pk)
+
+        self.assertEqual(p.animal_passthrough.name, "Leopard")
+        self.assertEqual(p.animal_passthrough['name'], "Leopard")
+
+        # Should not be able to access referenced document's methods
+        with self.assertRaises(AttributeError):
+            p.animal.save
+        with self.assertRaises(KeyError):
+            p.animal['save']
+
+    def test_lazy_reference_not_set(self):
+        class Animal(Document):
+            name = StringField()
+            tag = StringField()
+
+        class Ocurrence(Document):
+            person = StringField()
+            animal = LazyReferenceField(Animal)
+
+        Animal.drop_collection()
+        Ocurrence.drop_collection()
+
+        Ocurrence(person='foo').save()
+        p = Ocurrence.objects.get()
+        self.assertIs(p.animal, None)
+
+    def test_lazy_reference_equality(self):
+        class Animal(Document):
+            name = StringField()
+            tag = StringField()
+
+        Animal.drop_collection()
+
+        animal = Animal(name="Leopard", tag="heavy").save()
+        animalref = LazyReference(Animal, animal.pk)
+        self.assertEqual(animal, animalref)
+        self.assertEqual(animalref, animal)
+
+        other_animalref = LazyReference(Animal, ObjectId("54495ad94c934721ede76f90"))
+        self.assertNotEqual(animal, other_animalref)
+        self.assertNotEqual(other_animalref, animal)
+
+
 if __name__ == '__main__':
     unittest.main()