Fixed #7052 -- Added support for natural keys in serialization.

git-svn-id: http://code.djangoproject.com/svn/django/trunk@11863 bcc190cf-cafb-0310-a4f2-bffc1f526a37

django/contrib/auth/models.py

@@ -47,6 +47,13 @@ def check_password(raw_password, enc_password):
 class SiteProfileNotAvailable(Exception):
     pass
 
+class PermissionManager(models.Manager):
+    def get_by_natural_key(self, codename, app_label, model):
+        return self.get(
+            codename=codename,
+            content_type=ContentType.objects.get_by_natural_key(app_label, model)
+        )
+
 class Permission(models.Model):
     """The permissions system provides a way to assign permissions to specific users and groups of users.
 
@@ -63,6 +70,7 @@ class Permission(models.Model):
     name = models.CharField(_('name'), max_length=50)
     content_type = models.ForeignKey(ContentType)
     codename = models.CharField(_('codename'), max_length=100)
+    objects = PermissionManager()
 
     class Meta:
         verbose_name = _('permission')
@@ -76,6 +84,10 @@ class Permission(models.Model):
             unicode(self.content_type),
             unicode(self.name))
 
+    def natural_key(self):
+        return (self.codename,) + self.content_type.natural_key()
+    natural_key.dependencies = ['contenttypes.contenttype']
+
 class Group(models.Model):
     """Groups are a generic way of categorizing users to apply permissions, or some other label, to those users. A user can belong to any number of groups.
 

django/contrib/contenttypes/models.py

@@ -8,6 +8,13 @@ class ContentTypeManager(models.Manager):
     # This cache is shared by all the get_for_* methods.
     _cache = {}
 
+    def get_by_natural_key(self, app_label, model):
+        try:
+            ct = self.__class__._cache[(app_label, model)]
+        except KeyError:
+            ct = self.get(app_label=app_label, model=model)
+        return ct
+
     def get_for_model(self, model):
         """
         Returns the ContentType object for a given model, creating the
@@ -93,3 +100,6 @@ class ContentType(models.Model):
         so code that calls this method should catch it.
         """
         return self.model_class()._default_manager.get(**kwargs)
+
+    def natural_key(self):
+        return (self.app_label, self.model)

django/core/management/commands/dumpdata.py

@@ -13,6 +13,8 @@ class Command(BaseCommand):
             help='Specifies the indent level to use when pretty-printing output'),
         make_option('-e', '--exclude', dest='exclude',action='append', default=[],
             help='App to exclude (use multiple --exclude to exclude multiple apps).'),
+        make_option('-n', '--natural', action='store_true', dest='use_natural_keys', default=False,
+            help='Use natural keys if they are available.'),
     )
     help = 'Output the contents of the database as a fixture of the given format.'
     args = '[appname ...]'
@@ -24,6 +26,7 @@ class Command(BaseCommand):
         indent = options.get('indent',None)
         exclude = options.get('exclude',[])
         show_traceback = options.get('traceback', False)
+        use_natural_keys = options.get('use_natural_keys', False)
 
         excluded_apps = [get_app(app_label) for app_label in exclude]
 
@@ -67,18 +70,86 @@ class Command(BaseCommand):
         except KeyError:
             raise CommandError("Unknown serialization format: %s" % format)
 
+        # Now collate the objects to be serialized.
         objects = []
-        for app, model_list in app_list.items():
-            if model_list is None:
-                model_list = get_models(app)
-
-            for model in model_list:
-                if not model._meta.proxy:
-                    objects.extend(model._default_manager.all())
+        for model in sort_dependencies(app_list.items()):
+            if not model._meta.proxy:
+                objects.extend(model._default_manager.all())
 
         try:
-            return serializers.serialize(format, objects, indent=indent)
+            return serializers.serialize(format, objects, indent=indent,
+                        use_natural_keys=use_natural_keys)
         except Exception, e:
             if show_traceback:
                 raise
             raise CommandError("Unable to serialize database: %s" % e)
+
+def sort_dependencies(app_list):
+    """Sort a list of app,modellist pairs into a single list of models.
+
+    The single list of models is sorted so that any model with a natural key
+    is serialized before a normal model, and any model with a natural key
+    dependency has it's dependencies serialized first.
+    """
+    from django.db.models import get_model, get_models
+    # Process the list of models, and get the list of dependencies
+    model_dependencies = []
+    models = set()
+    for app, model_list in app_list:
+        if model_list is None:
+            model_list = get_models(app)
+
+        for model in model_list:
+            models.add(model)
+            # Add any explicitly defined dependencies
+            if hasattr(model, 'natural_key'):
+                deps = getattr(model.natural_key, 'dependencies', [])
+                if deps:
+                    deps = [get_model(*d.split('.')) for d in deps]
+            else:
+                deps = []
+
+            # Now add a dependency for any FK or M2M relation with
+            # a model that defines a natural key
+            for field in model._meta.fields:
+                if hasattr(field.rel, 'to'):
+                    rel_model = field.rel.to
+                    if hasattr(rel_model, 'natural_key'):
+                        deps.append(rel_model)
+            for field in model._meta.many_to_many:
+                rel_model = field.rel.to
+                if hasattr(rel_model, 'natural_key'):
+                    deps.append(rel_model)
+            model_dependencies.append((model, deps))
+
+    model_dependencies.reverse()
+    # Now sort the models to ensure that dependencies are met. This
+    # is done by repeatedly iterating over the input list of models.
+    # If all the dependencies of a given model are in the final list,
+    # that model is promoted to the end of the final list. This process
+    # continues until the input list is empty, or we do a full iteration
+    # over the input models without promoting a model to the final list.
+    # If we do a full iteration without a promotion, that means there are
+    # circular dependencies in the list.
+    model_list = []
+    while model_dependencies:
+        skipped = []
+        changed = False
+        while model_dependencies:
+            model, deps = model_dependencies.pop()
+            if all((d not in models or d in model_list) for d in deps):
+                # If all of the models in the dependency list are either already
+                # on the final model list, or not on the original serialization list,
+                # then we've found another model with all it's dependencies satisfied.
+                model_list.append(model)
+                changed = True
+            else:
+                skipped.append((model, deps))
+        if not changed:
+            raise CommandError("Can't resolve dependencies for %s in serialized app list." %
+                ', '.join('%s.%s' % (model._meta.app_label, model._meta.object_name)
+                for model, deps in sorted(skipped, key=lambda obj: obj[0].__name__))
+            )
+        model_dependencies = skipped
+
+    return model_list

django/core/serializers/base.py

@@ -33,6 +33,7 @@ class Serializer(object):
 
         self.stream = options.get("stream", StringIO())
         self.selected_fields = options.get("fields")
+        self.use_natural_keys = options.get("use_natural_keys", False)
 
         self.start_serialization()
         for obj in queryset:

django/core/serializers/json.py

@@ -24,6 +24,7 @@ class Serializer(PythonSerializer):
     def end_serialization(self):
         self.options.pop('stream', None)
         self.options.pop('fields', None)
+        self.options.pop('use_natural_keys', None)
         simplejson.dump(self.objects, self.stream, cls=DjangoJSONEncoder, **self.options)
 
     def getvalue(self):

django/core/serializers/python.py

@@ -47,17 +47,24 @@ class Serializer(base.Serializer):
     def handle_fk_field(self, obj, field):
         related = getattr(obj, field.name)
         if related is not None:
-            if field.rel.field_name == related._meta.pk.name:
-                # Related to remote object via primary key
-                related = related._get_pk_val()
+            if self.use_natural_keys and hasattr(related, 'natural_key'):
+                related = related.natural_key()
             else:
-                # Related to remote object via other field
-                related = getattr(related, field.rel.field_name)
-        self._current[field.name] = smart_unicode(related, strings_only=True)
+                if field.rel.field_name == related._meta.pk.name:
+                    # Related to remote object via primary key
+                    related = related._get_pk_val()
+                else:
+                    # Related to remote object via other field
+                    related = smart_unicode(getattr(related, field.rel.field_name), strings_only=True)
+        self._current[field.name] = related
 
     def handle_m2m_field(self, obj, field):
         if field.rel.through._meta.auto_created:
-            self._current[field.name] = [smart_unicode(related._get_pk_val(), strings_only=True)
+            if self.use_natural_keys and hasattr(field.rel.to, 'natural_key'):
+                m2m_value = lambda value: value.natural_key()
+            else:
+                m2m_value = lambda value: smart_unicode(value._get_pk_val(), strings_only=True)
+            self._current[field.name] = [m2m_value(related)
                                for related in getattr(obj, field.name).iterator()]
 
     def getvalue(self):
@@ -86,13 +93,28 @@ def Deserializer(object_list, **options):
 
             # Handle M2M relations
             if field.rel and isinstance(field.rel, models.ManyToManyRel):
-                m2m_convert = field.rel.to._meta.pk.to_python
-                m2m_data[field.name] = [m2m_convert(smart_unicode(pk)) for pk in field_value]
+                if hasattr(field.rel.to._default_manager, 'get_by_natural_key'):
+                    def m2m_convert(value):
+                        if hasattr(value, '__iter__'):
+                            return field.rel.to._default_manager.get_by_natural_key(*value).pk
+                        else:
+                            return smart_unicode(field.rel.to._meta.pk.to_python(value))
+                else:
+                    m2m_convert = lambda v: smart_unicode(field.rel.to._meta.pk.to_python(v))
+                m2m_data[field.name] = [m2m_convert(pk) for pk in field_value]
 
             # Handle FK fields
             elif field.rel and isinstance(field.rel, models.ManyToOneRel):
                 if field_value is not None:
-                    data[field.attname] = field.rel.to._meta.get_field(field.rel.field_name).to_python(field_value)
+                    if hasattr(field.rel.to._default_manager, 'get_by_natural_key'):
+                        if hasattr(field_value, '__iter__'):
+                            obj = field.rel.to._default_manager.get_by_natural_key(*field_value)
+                            value = getattr(obj, field.rel.field_name)
+                        else:
+                            value = field.rel.to._meta.get_field(field.rel.field_name).to_python(field_value)
+                        data[field.attname] = value
+                    else:
+                        data[field.attname] = field.rel.to._meta.get_field(field.rel.field_name).to_python(field_value)
                 else:
                     data[field.attname] = None
 

django/core/serializers/pyyaml.py

@@ -26,9 +26,9 @@ class Serializer(PythonSerializer):
     """
     Convert a queryset to YAML.
     """
-    
+
     internal_use_only = False
-    
+
     def handle_field(self, obj, field):
         # A nasty special case: base YAML doesn't support serialization of time
         # types (as opposed to dates or datetimes, which it does support). Since
@@ -40,10 +40,11 @@ class Serializer(PythonSerializer):
             self._current[field.name] = str(getattr(obj, field.name))
         else:
             super(Serializer, self).handle_field(obj, field)
-    
+
     def end_serialization(self):
         self.options.pop('stream', None)
         self.options.pop('fields', None)
+        self.options.pop('use_natural_keys', None)
         yaml.dump(self.objects, self.stream, Dumper=DjangoSafeDumper, **self.options)
 
     def getvalue(self):

django/core/serializers/xml_serializer.py

@@ -81,13 +81,22 @@ class Serializer(base.Serializer):
         self._start_relational_field(field)
         related = getattr(obj, field.name)
         if related is not None:
-            if field.rel.field_name == related._meta.pk.name:
-                # Related to remote object via primary key
-                related = related._get_pk_val()
+            if self.use_natural_keys and hasattr(related, 'natural_key'):
+                # If related object has a natural key, use it
+                related = related.natural_key()
+                # Iterable natural keys are rolled out as subelements
+                for key_value in related:
+                    self.xml.startElement("natural", {})
+                    self.xml.characters(smart_unicode(key_value))
+                    self.xml.endElement("natural")
             else:
-                # Related to remote object via other field
-                related = getattr(related, field.rel.field_name)
-            self.xml.characters(smart_unicode(related))
+                if field.rel.field_name == related._meta.pk.name:
+                    # Related to remote object via primary key
+                    related = related._get_pk_val()
+                else:
+                    # Related to remote object via other field
+                    related = getattr(related, field.rel.field_name)
+                self.xml.characters(smart_unicode(related))
         else:
             self.xml.addQuickElement("None")
         self.xml.endElement("field")
@@ -100,8 +109,25 @@ class Serializer(base.Serializer):
         """
         if field.rel.through._meta.auto_created:
             self._start_relational_field(field)
+            if self.use_natural_keys and hasattr(field.rel.to, 'natural_key'):
+                # If the objects in the m2m have a natural key, use it
+                def handle_m2m(value):
+                    natural = value.natural_key()
+                    # Iterable natural keys are rolled out as subelements
+                    self.xml.startElement("object", {})
+                    for key_value in natural:
+                        self.xml.startElement("natural", {})
+                        self.xml.characters(smart_unicode(key_value))
+                        self.xml.endElement("natural")
+                    self.xml.endElement("object")
+            else:
+                def handle_m2m(value):
+                    self.xml.addQuickElement("object", attrs={
+                        'pk' : smart_unicode(value._get_pk_val())
+                    })
             for relobj in getattr(obj, field.name).iterator():
-                self.xml.addQuickElement("object", attrs={"pk" : smart_unicode(relobj._get_pk_val())})
+                handle_m2m(relobj)
+
             self.xml.endElement("field")
 
     def _start_relational_field(self, field):
@@ -187,16 +213,40 @@ class Deserializer(base.Deserializer):
         if node.getElementsByTagName('None'):
             return None
         else:
-            return field.rel.to._meta.get_field(field.rel.field_name).to_python(
-                       getInnerText(node).strip())
+            if hasattr(field.rel.to._default_manager, 'get_by_natural_key'):
+                keys = node.getElementsByTagName('natural')
+                if keys:
+                    # If there are 'natural' subelements, it must be a natural key
+                    field_value = [getInnerText(k).strip() for k in keys]
+                    obj = field.rel.to._default_manager.get_by_natural_key(*field_value)
+                    obj_pk = getattr(obj, field.rel.field_name)
+                else:
+                    # Otherwise, treat like a normal PK
+                    field_value = getInnerText(node).strip()
+                    obj_pk = field.rel.to._meta.get_field(field.rel.field_name).to_python(field_value)
+                return obj_pk
+            else:
+                field_value = getInnerText(node).strip()
+                return field.rel.to._meta.get_field(field.rel.field_name).to_python(field_value)
 
     def _handle_m2m_field_node(self, node, field):
         """
         Handle a <field> node for a ManyToManyField.
         """
-        return [field.rel.to._meta.pk.to_python(
-                    c.getAttribute("pk"))
-                    for c in node.getElementsByTagName("object")]
+        if hasattr(field.rel.to._default_manager, 'get_by_natural_key'):
+            def m2m_convert(n):
+                keys = n.getElementsByTagName('natural')
+                if keys:
+                    # If there are 'natural' subelements, it must be a natural key
+                    field_value = [getInnerText(k).strip() for k in keys]
+                    obj_pk = field.rel.to._default_manager.get_by_natural_key(*field_value).pk
+                else:
+                    # Otherwise, treat like a normal PK value.
+                    obj_pk = field.rel.to._meta.pk.to_python(n.getAttribute('pk'))
+                return obj_pk
+        else:
+            m2m_convert = lambda n: field.rel.to._meta.pk.to_python(n.getAttribute('pk'))
+        return [m2m_convert(c) for c in node.getElementsByTagName("object")]
 
     def _get_model_from_node(self, node, attr):
         """

docs/ref/django-admin.txt

@@ -234,6 +234,17 @@ name to ``dumpdata``, the dumped output will be restricted to that model,
 rather than the entire application. You can also mix application names and
 model names.
 
+.. django-admin-option:: --natural
+
+.. versionadded:: 1.2
+
+Use :ref:`natural keys <topics-serialization-natural-keys>` to represent
+any foreign key and many-to-many relationship with a model that provides
+a natural key definition. If you are dumping ``contrib.auth`` ``Permission``
+objects or ``contrib.contenttypes`` ``ContentType`` objects, you should
+probably be using this flag.
+
+
 flush
 -----
 
@@ -701,7 +712,7 @@ information.
 
 .. versionadded:: 1.2
 
-Use the ``--failfast`` option to stop running tests and report the failure 
+Use the ``--failfast`` option to stop running tests and report the failure
 immediately after a test fails.
 
 testserver <fixture fixture ...>

docs/releases/1.2.txt

@@ -267,3 +267,13 @@ include %}`` tags).
 As a side effect, it is now much easier to support non-Django template
 languages. For more details, see the :ref:`notes on supporting
 non-Django template languages<topic-template-alternate-language>`.
+
+Natural keys in fixtures
+------------------------
+
+Fixtures can refer to remote objects using
+:ref:`topics-serialization-natural-keys`. This lookup scheme is an
+alternative to the normal primary-key based object references in a
+fixture, improving readability, and resolving problems referring to
+objects whose primary key value may not be predictable or known.
+

docs/topics/serialization.txt

@@ -154,10 +154,10 @@ to install third-party Python modules:
 .. _PyYAML: http://www.pyyaml.org/
 
 Notes for specific serialization formats
-----------------------------------------
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 json
-~~~~
+^^^^
 
 If you're using UTF-8 (or any other non-ASCII encoding) data with the JSON
 serializer, you must pass ``ensure_ascii=False`` as a parameter to the
@@ -191,3 +191,191 @@ them. Something like this will work::
 
 .. _special encoder: http://svn.red-bean.com/bob/simplejson/tags/simplejson-1.7/docs/index.html
 
+.. _topics-serialization-natural-keys:
+
+Natural keys
+------------
+
+The default serialization strategy for foreign keys and many-to-many
+relations is to serialize the value of the primary key(s) of the
+objects in the relation. This strategy works well for most types of
+object, but it can cause difficulty in some circumstances.
+
+Consider the case of a list of objects that have foreign key on
+:class:`ContentType`. If you're going to serialize an object that
+refers to a content type, you need to have a way to refer to that
+content type. Content Types are automatically created by Django as
+part of the database synchronization process, so you don't need to
+include content types in a fixture or other serialized data. As a
+result, the primary key of any given content type isn't easy to
+predict - it will depend on how and when :djadmin:`syncdb` was
+executed to create the content types.
+
+There is also the matter of convenience. An integer id isn't always
+the most convenient way to refer to an object; sometimes, a
+more natural reference would be helpful.
+
+Deserialization of natural keys
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+It is for these reasons that Django provides `natural keys`. A natural
+key is a tuple of values that can be used to uniquely identify an
+object instance without using the primary key value.
+
+Consider the following two models::
+
+    from django.db import models
+
+    class Person(models.Model):
+        first_name = models.CharField(max_length=100)
+        last_name = models.CharField(max_length=100)
+
+        birthdate = models.DateField()
+
+    class Book(models.Model):
+        name = models.CharField(max_length=100)
+        author = models.ForeignKey(Person)
+
+Ordinarily, serialized data for ``Book`` would use an integer to refer to
+the author. For example, in JSON, a Book might be serialized as::
+
+    ...
+    {
+        "pk": 1,
+        "model": "store.book",
+        "fields": {
+            "name": "Mostly Harmless",
+            "author": 42
+        }
+    }
+    ...
+
+This isn't a particularly natural way to refer to an author. It
+requires that you know the primary key value for the author; it also
+requires that this primary key value is stable and predictable.
+
+However, if we add natural key handling to Person, the fixture becomes
+much more humane. To add natural key handling, you define a default
+Manager for Person with a ``get_by_natural_key()`` method. In the case
+of a Person, a good natural key might be the pair of first and last
+name::
+
+    from django.db import models
+
+    class PersonManager(models.Manager):
+        def get_by_natural_key(self, first_name, last_name):
+            return self.filter(first_name=first_name, last_name=last_name)
+
+    class Person(models.Model):
+        objects = PersonManager()
+
+        first_name = models.CharField(max_length=100)
+        last_name = models.CharField(max_length=100)
+
+        birthdate = models.DateField()
+
+Now books can use that natural key to refer to ``Person`` objects::
+
+    ...
+    {
+        "pk": 1,
+        "model": "store.book",
+        "fields": {
+            "name": "Mostly Harmless",
+            "author": ["Douglas", "Adams"]
+        }
+    }
+    ...
+
+When you try to load this serialized data, Django will use the
+``get_by_natural_key()`` method to resolve ``["Douglas", "Adams"]``
+into the primary key of an actual ``Person`` object.
+
+Serialization of natural keys
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+So how do you get Django to emit a natural key when serializing an object?
+Firstly, you need to add another method -- this time to the model itself::
+
+    class Person(models.Model):
+        objects = PersonManager()
+
+        first_name = models.CharField(max_length=100)
+        last_name = models.CharField(max_length=100)
+
+        birthdate = models.DateField()
+
+        def natural_key(self):
+            return (self.first_name, self.last_name)
+
+Then, when you call ``serializers.serialize()``, you provide a
+``use_natural_keys=True`` argument::
+
+    >>> serializers.serialize([book1, book2], format='json', indent=2, use_natural_keys=True)
+
+When ``use_natural_keys=True`` is specified, Django will use the
+``natural_key()`` method to serialize any reference to objects of the
+type that defines the method.
+
+If you are using :djadmin:`dumpdata` to generate serialized data, you
+use the `--natural` command line flag to generate natural keys.
+
+.. note::
+
+    You don't need to define both ``natural_key()`` and
+    ``get_by_natural_key()``. If you don't want Django to output
+    natural keys during serialization, but you want to retain the
+    ability to load natural keys, then you can opt to not implement
+    the ``natural_key()`` method.
+
+    Conversely, if (for some strange reason) you want Django to output
+    natural keys during serialization, but *not* be able to load those
+    key values, just don't define the ``get_by_natural_key()`` method.
+
+Dependencies during serialization
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+Since natural keys rely on database lookups to resolve references, it
+is important that data exists before it is referenced. You can't make
+a `forward reference` with natural keys - the data you are referencing
+must exist before you include a natural key reference to that data.
+
+To accommodate this limitation, calls to :djadmin:`dumpdata` that use
+the :djadminopt:`--natural` optionwill serialize any model with a
+``natural_key()`` method before it serializes normal key objects.
+
+However, this may not always be enough. If your natural key refers to
+another object (by using a foreign key or natural key to another object
+as part of a natural key), then you need to be able to ensure that
+the objects on which a natural key depends occur in the serialized data
+before the natural key requires them.
+
+To control this ordering, you can define dependencies on your
+``natural_key()`` methods. You do this by setting a ``dependencies``
+attribute on the ``natural_key()`` method itself.
+
+For example, consider the ``Permission`` model in ``contrib.auth``.
+The following is a simplified version of the ``Permission`` model::
+
+    class Permission(models.Model):
+        name = models.CharField(max_length=50)
+        content_type = models.ForeignKey(ContentType)
+        codename = models.CharField(max_length=100)
+        # ...
+        def natural_key(self):
+            return (self.codename,) + self.content_type.natural_key()
+
+The natural key for a ``Permission`` is a combination of the codename for the
+``Permission``, and the ``ContentType`` to which the ``Permission`` applies. This means
+that ``ContentType`` must be serialized before ``Permission``. To define this
+dependency, we add one extra line::
+
+    class Permission(models.Model):
+        # ...
+        def natural_key(self):
+            return (self.codename,) + self.content_type.natural_key()
+        natural_key.dependencies = ['contenttypes.contenttype']
+
+This definition ensures that ``ContentType`` models are serialized before
+``Permission`` models. In turn, any object referencing ``Permission`` will
+be serialized after both ``ContentType`` and ``Permission``.

tests/modeltests/fixtures/fixtures/fixture6.json

@@ -0,0 +1,41 @@
+[
+    {
+        "pk": "1",
+        "model": "fixtures.tag",
+        "fields": {
+            "name": "copyright",
+            "tagged_type": ["fixtures", "article"],
+            "tagged_id": "3"
+        }
+    },
+    {
+        "pk": "2",
+        "model": "fixtures.tag",
+        "fields": {
+            "name": "law",
+            "tagged_type": ["fixtures", "article"],
+            "tagged_id": "3"
+        }
+    },
+    {
+    	"pk": "1",
+    	"model": "fixtures.person",
+    	"fields": {
+    	    "name": "Django Reinhardt"
+    	}
+    },
+    {
+    	"pk": "2",
+    	"model": "fixtures.person",
+    	"fields": {
+    	    "name": "Stephane Grappelli"
+    	}
+    },
+        {
+    	"pk": "3",
+    	"model": "fixtures.person",
+    	"fields": {
+    	    "name": "Prince"
+    	}
+    }
+]

tests/modeltests/fixtures/fixtures/fixture7.xml

@@ -0,0 +1,27 @@
+<?xml version="1.0" encoding="utf-8"?>
+<django-objects version="1.0">
+    <object pk="2" model="fixtures.tag">
+        <field type="CharField" name="name">legal</field>
+        <field to="contenttypes.contenttype" name="tagged_type" rel="ManyToOneRel">
+            <natural>fixtures</natural>
+            <natural>article</natural>
+        </field>
+        <field type="PositiveIntegerField" name="tagged_id">3</field>
+    </object>
+    <object pk="3" model="fixtures.tag">
+        <field type="CharField" name="name">django</field>
+        <field to="contenttypes.contenttype" name="tagged_type" rel="ManyToOneRel">
+            <natural>fixtures</natural>
+            <natural>article</natural>
+        </field>
+        <field type="PositiveIntegerField" name="tagged_id">4</field>
+    </object>
+    <object pk="4" model="fixtures.tag">
+        <field type="CharField" name="name">world domination</field>
+        <field to="contenttypes.contenttype" name="tagged_type" rel="ManyToOneRel">
+            <natural>fixtures</natural>
+            <natural>article</natural>
+        </field>
+        <field type="PositiveIntegerField" name="tagged_id">4</field>
+    </object>
+</django-objects>

tests/modeltests/fixtures/fixtures/fixture8.json

@@ -0,0 +1,32 @@
+[
+    {
+    	"pk": "1",
+    	"model": "fixtures.visa",
+    	"fields": {
+    	    "person": ["Django Reinhardt"],
+    	    "permissions": [
+    	        ["add_user", "auth", "user"],
+    	        ["change_user", "auth", "user"],
+    	        ["delete_user", "auth", "user"]
+    	    ]
+    	}
+    },
+    {
+    	"pk": "2",
+    	"model": "fixtures.visa",
+    	"fields": {
+    	    "person": ["Stephane Grappelli"],
+    	    "permissions": [
+    	        ["add_user", "auth", "user"]
+    	    ]
+    	}
+    },
+        {
+    	"pk": "3",
+    	"model": "fixtures.visa",
+    	"fields": {
+    	    "person": ["Prince"],
+    	    "permissions": []
+    	}
+    }
+]

tests/modeltests/fixtures/fixtures/fixture9.xml

@@ -0,0 +1,48 @@
+<?xml version="1.0" encoding="utf-8"?>
+<django-objects version="1.0">
+    <object pk="2" model="fixtures.visa">
+        <field type="CharField" name="person">
+            <natural>Stephane Grappelli</natural>
+        </field>
+        <field to="auth.permission" name="permissions" rel="ManyToManyRel">
+            <object>
+                <natural>add_user</natural>
+                <natural>auth</natural>
+                <natural>user</natural>
+            </object>
+            <object>
+                <natural>delete_user</natural>
+                <natural>auth</natural>
+                <natural>user</natural>
+            </object>
+        </field>
+    </object>
+    <object pk="3" model="fixtures.person">
+        <field type="CharField" name="name">
+            <natural>Artist formerly known as &quot;Prince&quot;</natural>
+        </field>
+    </object>
+    <object pk="3" model="fixtures.visa">
+        <field type="CharField" name="person">
+            <natural>Artist formerly known as &quot;Prince&quot;</natural>
+        </field>
+        <field to="auth.permission" name="permissions" rel="ManyToManyRel">
+            <object>
+                <natural>change_user</natural>
+                <natural>auth</natural>
+                <natural>user</natural>
+            </object>
+        </field>
+    </object>
+    <object pk="1" model="fixtures.book">
+        <field type="CharField" name="name">Music for all ages</field>
+        <field to="fixtures.person" name="authors" rel="ManyToManyRel">
+            <object>
+                <natural>Django Reinhardt</natural>
+            </object>
+            <object>
+                <natural>Artist formerly known as &quot;Prince&quot;</natural>
+            </object>
+        </field>
+    </object>
+</django-objects>

tests/regressiontests/fixtures_regress/fixtures/forward_ref_lookup.json

@@ -0,0 +1,32 @@
+[
+    {
+        "pk": "4",
+        "model": "fixtures_regress.person",
+        "fields": {
+            "name": "Neal Stephenson"
+        }
+    },
+    {
+        "pk": "2",
+        "model": "fixtures_regress.store",
+        "fields": {
+            "name": "Amazon"
+        }
+    },
+    {
+        "pk": "3",
+        "model": "fixtures_regress.store",
+        "fields": {
+            "name": "Borders"
+        }
+    },
+    {
+        "pk": 1,
+        "model": "fixtures_regress.book",
+        "fields": {
+            "name": "Cryptonomicon",
+            "author": ["Neal Stephenson"],
+            "stores": [["Amazon"], ["Borders"]]
+        }
+    }
+]

tests/regressiontests/fixtures_regress/fixtures/non_natural_1.json

@@ -0,0 +1,25 @@
+[
+    {
+        "pk": 12,
+        "model": "fixtures_regress.person",
+        "fields": {
+            "name": "Greg Egan"
+        }
+    },
+    {
+        "pk": 11,
+        "model": "fixtures_regress.store",
+        "fields": {
+            "name": "Angus and Robertson"
+        }
+    },
+    {
+        "pk": 10,
+        "model": "fixtures_regress.book",
+        "fields": {
+            "name": "Permutation City",
+            "author": 12,
+            "stores": [11]
+        }
+    }
+]

tests/regressiontests/fixtures_regress/fixtures/non_natural_2.xml

@@ -0,0 +1,16 @@
+<?xml version="1.0" encoding="utf-8"?>
+<django-objects version="1.0">
+    <object pk="22" model="fixtures_regress.person">
+        <field type="CharField" name="name">Orson Scott Card</field>
+    </object>
+    <object pk="21" model="fixtures_regress.store">
+        <field type="CharField" name="name">Collins Bookstore</field>
+    </object>
+    <object pk="20" model="fixtures_regress.book">
+        <field type="CharField" name="name">Ender's Game</field>
+        <field to="fixtures_regress.person" name="author" rel="ManyToOneRel">22</field>
+        <field to="fixtures_regress.store" name="stores" rel="ManyToManyRel">
+            <object pk="21"/>
+        </field>
+    </object>
+</django-objects>

tests/regressiontests/fixtures_regress/models.py

@@ -13,7 +13,7 @@ class Animal(models.Model):
     specimens = models.Manager()
 
     def __unicode__(self):
-        return self.common_name
+        return self.name
 
 def animal_pre_save_check(signal, sender, instance, **kwargs):
     "A signal that is used to check the type of data loaded from fixtures"
@@ -69,10 +69,66 @@ class Article(models.Model):
 class Widget(models.Model):
     name = models.CharField(max_length=255)
 
+    class Meta:
+        ordering = ('name',)
+
+    def __unicode__(self):
+        return self.name
+
 class WidgetProxy(Widget):
     class Meta:
         proxy = True
 
+# Check for forward references in FKs and M2Ms with natural keys
+
+class TestManager(models.Manager):
+    def get_by_natural_key(self, key):
+        return self.get(name=key)
+
+class Store(models.Model):
+    objects = TestManager()
+    name = models.CharField(max_length=255)
+
+    class Meta:
+        ordering = ('name',)
+
+    def __unicode__(self):
+        return self.name
+
+    def natural_key(self):
+        return (self.name,)
+
+class Person(models.Model):
+    objects = TestManager()
+    name = models.CharField(max_length=255)
+
+    class Meta:
+        ordering = ('name',)
+
+    def __unicode__(self):
+        return self.name
+
+    # Person doesn't actually have a dependency on store, but we need to define
+    # one to test the behaviour of the dependency resolution algorithm.
+    def natural_key(self):
+        return (self.name,)
+    natural_key.dependencies = ['fixtures_regress.store']
+
+class Book(models.Model):
+    name = models.CharField(max_length=255)
+    author = models.ForeignKey(Person)
+    stores = models.ManyToManyField(Store)
+
+    class Meta:
+        ordering = ('name',)
+
+    def __unicode__(self):
+        return u'%s by %s (available at %s)' % (
+            self.name,
+            self.author.name,
+            ', '.join(s.name for s in self.stores.all())
+        )
+
 __test__ = {'API_TESTS':"""
 >>> from django.core import management
 
@@ -192,4 +248,121 @@ Weight = 1.2 (<type 'float'>)
 >>> management.call_command('dumpdata', 'fixtures_regress', format='json')
 [{"pk": 1, "model": "fixtures_regress.widget", "fields": {"name": "grommet"}}]
 
+###############################################
+# Check that natural key requirements are taken into account
+# when serializing models
+>>> management.call_command('loaddata', 'forward_ref_lookup.json', verbosity=0)
+
+>>> management.call_command('dumpdata', 'fixtures_regress.book', 'fixtures_regress.person', 'fixtures_regress.store', verbosity=0, use_natural_keys=True)
+[{"pk": 2, "model": "fixtures_regress.store", "fields": {"name": "Amazon"}}, {"pk": 3, "model": "fixtures_regress.store", "fields": {"name": "Borders"}}, {"pk": 4, "model": "fixtures_regress.person", "fields": {"name": "Neal Stephenson"}}, {"pk": 1, "model": "fixtures_regress.book", "fields": {"stores": [["Amazon"], ["Borders"]], "name": "Cryptonomicon", "author": ["Neal Stephenson"]}}]
+
+# Now lets check the dependency sorting explicitly
+
+# First Some models with pathological circular dependencies
+>>> class Circle1(models.Model):
+...     name = models.CharField(max_length=255)
+...     def natural_key(self):
+...         return self.name
+...     natural_key.dependencies = ['fixtures_regress.circle2']
+
+>>> class Circle2(models.Model):
+...     name = models.CharField(max_length=255)
+...     def natural_key(self):
+...         return self.name
+...     natural_key.dependencies = ['fixtures_regress.circle1']
+
+>>> class Circle3(models.Model):
+...     name = models.CharField(max_length=255)
+...     def natural_key(self):
+...         return self.name
+...     natural_key.dependencies = ['fixtures_regress.circle3']
+
+>>> class Circle4(models.Model):
+...     name = models.CharField(max_length=255)
+...     def natural_key(self):
+...         return self.name
+...     natural_key.dependencies = ['fixtures_regress.circle5']
+
+>>> class Circle5(models.Model):
+...     name = models.CharField(max_length=255)
+...     def natural_key(self):
+...         return self.name
+...     natural_key.dependencies = ['fixtures_regress.circle6']
+
+>>> class Circle6(models.Model):
+...     name = models.CharField(max_length=255)
+...     def natural_key(self):
+...         return self.name
+...     natural_key.dependencies = ['fixtures_regress.circle4']
+
+>>> class ExternalDependency(models.Model):
+...     name = models.CharField(max_length=255)
+...     def natural_key(self):
+...         return self.name
+...     natural_key.dependencies = ['fixtures_regress.book']
+
+# It doesn't matter what order you mention the models
+# Store *must* be serialized before then Person, and both
+# must be serialized before Book.
+>>> from django.core.management.commands.dumpdata import sort_dependencies
+>>> sort_dependencies([('fixtures_regress', [Book, Person, Store])])
+[<class 'regressiontests.fixtures_regress.models.Store'>, <class 'regressiontests.fixtures_regress.models.Person'>, <class 'regressiontests.fixtures_regress.models.Book'>]
+
+>>> sort_dependencies([('fixtures_regress', [Book, Store, Person])])
+[<class 'regressiontests.fixtures_regress.models.Store'>, <class 'regressiontests.fixtures_regress.models.Person'>, <class 'regressiontests.fixtures_regress.models.Book'>]
+
+>>> sort_dependencies([('fixtures_regress', [Store, Book, Person])])
+[<class 'regressiontests.fixtures_regress.models.Store'>, <class 'regressiontests.fixtures_regress.models.Person'>, <class 'regressiontests.fixtures_regress.models.Book'>]
+
+>>> sort_dependencies([('fixtures_regress', [Store, Person, Book])])
+[<class 'regressiontests.fixtures_regress.models.Store'>, <class 'regressiontests.fixtures_regress.models.Person'>, <class 'regressiontests.fixtures_regress.models.Book'>]
+
+>>> sort_dependencies([('fixtures_regress', [Person, Book, Store])])
+[<class 'regressiontests.fixtures_regress.models.Store'>, <class 'regressiontests.fixtures_regress.models.Person'>, <class 'regressiontests.fixtures_regress.models.Book'>]
+
+>>> sort_dependencies([('fixtures_regress', [Person, Store, Book])])
+[<class 'regressiontests.fixtures_regress.models.Store'>, <class 'regressiontests.fixtures_regress.models.Person'>, <class 'regressiontests.fixtures_regress.models.Book'>]
+
+# A dangling dependency - assume the user knows what they are doing.
+>>> sort_dependencies([('fixtures_regress', [Person, Circle1, Store, Book])])
+[<class 'regressiontests.fixtures_regress.models.Circle1'>, <class 'regressiontests.fixtures_regress.models.Store'>, <class 'regressiontests.fixtures_regress.models.Person'>, <class 'regressiontests.fixtures_regress.models.Book'>]
+
+# A tight circular dependency
+>>> sort_dependencies([('fixtures_regress', [Person, Circle2, Circle1, Store, Book])])
+Traceback (most recent call last):
+...
+CommandError: Can't resolve dependencies for fixtures_regress.Circle1, fixtures_regress.Circle2 in serialized app list.
+
+>>> sort_dependencies([('fixtures_regress', [Circle1, Book, Circle2])])
+Traceback (most recent call last):
+...
+CommandError: Can't resolve dependencies for fixtures_regress.Circle1, fixtures_regress.Circle2 in serialized app list.
+
+# A self referential dependency
+>>> sort_dependencies([('fixtures_regress', [Book, Circle3])])
+Traceback (most recent call last):
+...
+CommandError: Can't resolve dependencies for fixtures_regress.Circle3 in serialized app list.
+
+# A long circular dependency
+>>> sort_dependencies([('fixtures_regress', [Person, Circle2, Circle1, Circle3, Store, Book])])
+Traceback (most recent call last):
+...
+CommandError: Can't resolve dependencies for fixtures_regress.Circle1, fixtures_regress.Circle2, fixtures_regress.Circle3 in serialized app list.
+
+# A dependency on a normal, non-natural-key model
+>>> sort_dependencies([('fixtures_regress', [Person, ExternalDependency, Book])])
+[<class 'regressiontests.fixtures_regress.models.Person'>, <class 'regressiontests.fixtures_regress.models.Book'>, <class 'regressiontests.fixtures_regress.models.ExternalDependency'>]
+
+###############################################
+# Check that normal primary keys still work
+# on a model with natural key capabilities
+
+>>> management.call_command('loaddata', 'non_natural_1.json', verbosity=0)
+>>> management.call_command('loaddata', 'non_natural_2.xml', verbosity=0)
+
+>>> Book.objects.all()
+[<Book: Cryptonomicon by Neal Stephenson (available at Amazon, Borders)>, <Book: Ender's Game by Orson Scott Card (available at Collins Bookstore)>, <Book: Permutation City by Greg Egan (available at Angus and Robertson)>]
+
 """}
+