Refs #6846
This provides a way for pagination classes to add pagination properties (`count`, `next`, `results` etc.) to OpenAPI response schemas.
A new method `get_paginated_response_schema()` has been added to `BasePagination`. This method is intended to mirror `get_paginated_response()` (which takes a `list` and wraps it in a `dict`).
Hence, `get_paginated_response_schema()` takes an unpaginated response schema (of type `array`) and wraps that with a schema object of type `object` containing the relevant properties that the pagination class adds to responses.
The default implementation of `BasePagination.get_paginated_response_schema()` simply passes the schema through unmodified, for backwards compatibility.
This is the first part of #6846.
Previously, the response schema for list views was an object representing a single item. However, list views return a list of items, and hence it should be an array.
Further work will need to be done to support how pagination classes modify list responses.
There should be no change for views not determined to be list views.
* Add tests for raise_errors_on_nested_writes
* Fix dotted-source field checking on serializer write
The code was previously checking the validated data for the field's
attribute name, however, the data contain the first source attr.
Refs #6509
This enforces allow_empty=True when a ListSerializer is a child of another serializer and partial validation is being performed on the parent serializer.
This is because partial validation should allow fields to be omitted, but should not cause values that are invalid without partial validation to become valid.
This effectively reverts #4222. None of the tests added in that PR fail if the associated change is removed, so I‘m not sure what that PR was trying to fix.
* fixes#6751
* reverted condition
* save instance before setting m2m fields
* added comment why m2m fields are saved after instance
* removed blank line
* added test for the issue 6751
Use of PyYAML's yaml.load function without specifying the Loader
parameter has been deprecated, see https://msg.pyyaml.org/load.
Earlier versions of PyYAML already had the alternative safe_load
function, which limits the loader to a subset of YAML constructs, that
is enough for what we need here.
Fixes#6677
It might be useful for a serializer with many many fields which uses
read_only and write_only on a large percentage of the fields. But the
memory usage and statefulness it adds are not worth it for the common
case.
* Added regression tests (#6504)
Co-Authored-By: Tom Quinonero <tq@3yourmind.com>
* Fix CursorPagination when objects get deleted between calls (#6504)
Co-Authored-By: Tom Quinonero <tq@3yourmind.com>
On Python 3, the ugettext functions are a simple aliases of their non-u
counterparts (the 'u' represents Python 2 unicode type). Starting with
Django 3.0, the u versions will be deprecated.
https://docs.djangoproject.com/en/dev/releases/3.0/#id2
> django.utils.translation.ugettext(), ugettext_lazy(), ugettext_noop(),
> ungettext(), and ungettext_lazy() are deprecated in favor of the
> functions that they’re aliases for:
> django.utils.translation.gettext(), gettext_lazy(), gettext_noop(),
> ngettext(), and ngettext_lazy().
Thanks to Jon Dufresne (@jdufresne) for review.
Co-authored-by: Asif Saif Uddin <auvipy@gmail.com>
Co-authored-by: Rizwan Mansuri <Rizwan@webbyfox.com>
* Introspect ManyRelatedField data type recursively
For all `ManyRelatedField` objects, we were assuming that the inner type was always a `String`. While this may be true for the default output, a `ManyRelatedField` is a wrapper for a lot of other classes which includes more than just strings. This should allow us to document lists of things other than strings.
* Added test for schemas for many-to-many fields
This adds a test that makes sure we generate the schema for a many-to-many field such that it actually has the right type. For some reason we did not previously have any tests for schema generation that included them, so hopefully this will prevent any future issues from popping up.
This should serve as a regression test for the `items` field on to-many relationships, which was previously forced to a `String` even though in most cases it is a different inner type within the array.
As all source files import unicode_literals, type('') is always
equivalent to six.text_type (str on Python 3 and unicode on Python 2).
Removes the need to call type(), is more explicit, and will be easier to
catch places to change for when it is time to eventually drop Python 2.
see https://github.com/encode/django-rest-framework/issues/5660
Trying to register new routes on a router after having accessed the
router `urls` attribute leads to surprising results.
The route is added without error to the router's `registry` but the urls
are not updated, because they are cached in `_urls`.
This commit invalidates the cache after each new registration.
* Add test that verifies that HTML is correctly escaped in Browsable API views
* Fix `urlize_quoted_links` tag to avoid double escaping in autoescape mode
* Fix XSS in default DRF Browsable API template by re-enabling autoescape
* Fix issue1811: take limit_choices_to into account with FK
* Issue 1811: Add tests to illustrate issue
* Filter queryset only if limit_choices_to exists
* Move test_relations_with_limited_querysets file within test_relations_pk
* move limit_choices_to logic from relations.py to utils/field_mapping.py
* move limit_choices_to above other check to avoid conflicts
In some cases we end with an operation between two `OperandHolder`.
This didn't work as it didn't knew how to deal with | or &
This fixes by adding those operations.
* Release notes to 5174a26ec9
* Update version for v3.9.0
* Removed exclude_from_schema per deprecation policy.
* Updated list_route() and detail_route() deprecations.
* Weakened to PendingDeprecationWarning for `base_name`
cc @rpkilby.
* Add (beginning of) 3.9 release announcement.
@tomchristie: Input on OpenAPI and What’s Next very welcome! :)
* Add announcement section for Extra Actions in Browsable API
* Update release notes and add deprecation note for Django Guardian backend.
* Add release note for #6073
* Add release notes to dd19a44583
* Adding release notes
* Update 3.9 announcement
* Add Oct 18 release date
* added ability to add sections to custom action documentation
* added tests to cover docs sections in custom actions
* added method specific docs test for action mapping
* added docs for custom action documentation
* permissions: Allow permissions to be composed
Implement a system to compose permissions with and / or.
This is performed by returning an `OperationHolder` instance that keeps the
permission classes and type of composition (and / or).
When called it will return a AND/OR instance that will then delegate the
permission check to the operands.
* permissions: Add documentation about composed permissions
* Fix documentation typo in permissions
* Add suffix support for actions
Removes the newly introduced `action.name` in favor of leveraging the
View's `.get_view_name()` method, which supports both name and suffix.
* Fix view description func docstrings
* Test action decorator name & suffix kwargs
* Adjust 'extra action' docs
On all supported Pythons, the io.BytesIO is always a stream
implementation using an in-memory bytes buffer.
Makes code slightly more forward compatible by reducing use of the six
module and promotes more forward compatible practices in the docs.
* Make admin detail link have small width
* Disable admin detail link when no URL
* Add 'AdminRenderer.get_result_url'
Attempts to reverse the result's detail view URL.
* Add failing test for extra action schemas
* Add ViewInspector setter to store instances
* Fix schema disabling for extra actions
* Add docs note about disabling schemas for actions
* View suffix already set by initializer
* Add 'name' and 'description' attributes to ViewSet
ViewSets may now provide their `name` and `description` attributes
directly, instead of relying on view introspection to derive them.
These attributes may also be provided with the view's initkwargs.
The ViewSet `name` and `suffix` initkwargs are mutually exclusive.
The `action` decorator now provides the `name` and `description` to
the view's initkwargs. By default, these values are derived from the
method name and its docstring. The `name` may be overridden by providing
it as an argument to the decorator.
The `get_view_name` and `get_view_description` hooks now provide the
view instance to the handler, instead of the view class. The default
implementations of these handlers now respect the `name`/`description`.
* Add 'extra actions' to ViewSet & browsable APIs
* Update simple router tests
Removed old test logic around link/action decorators from `v2.3`. Also
simplified the test by making the results explicit instead of computed.
* Add method mapping to ViewSet actions
* Document extra action method mapping
* Revert "Non-required fields with 'allow_null=True' should not imply a default value (#5639)"
This reverts commit 905a5579df.
Closes#5708
* Add test for allow_null + required=False
Ref #5708: allow_null should imply default=None, even for non-required fields.
* Re-order allow_null and default in field docs
default is prior to allow_null. allow_null implies an outgoing default=None.
* Adjust allow_null note.
* Always exclude read_only fields from _writable_fields
* Remove `read_only` from `CreateOnlyDefault` example.
In this context (without mentioning `save`) now slightly misleading.
The tests look for the "admin" app in the list of apps. If not present, running `runtests.py` gives the following error:
```
LookupError: No installed app with label 'admin'.
```
Adding admin to `INSTALLED_APPS` fixes it.
DictFields were incorrectly being output as String in the schema.
This pull request outputs an Object instead and adds a unit test.
Update s/detail_route/action/ after rebase
By default all subclasses of RelatedField are output as string fields in
the schema, which works well for StringRelatedField, SlugRelatedField or
HyperlinkedRelatedField.
Handle the common case of a PrimaryKeyRelatedField pointing to an
AutoField.
This adds `__eq__` to handle `code` in comparisons.
When comparing an ErrorDetail to a string (missing `code` there) the
ErrorDetail's `code` is ignored, but otherwise it is taken into account.
Calling dict.keys() is unnecessary. The two are functionally equivalent
on modern Pythons.
Inspired by Lennart Regebro's talk "Prehistoric Patterns in Python" from
PyCon 2017.
https://www.youtube.com/watch?v=V5-JH23Vk0I
* Add failing test for #5672
* Add get_original_route to complement get_regex_pattern
* [WIP] Fix path handling
* needs more tests
* maybe needs some refactoring
* Add django 2 variant for all tests and fix trailing slash bug
* Add more combinations to mixed path test
* Extract method for `manual_fields` processing
Allows reuse of logic to replace Field instances in a field list by `Field.name`.
Adds a utility function for the logic plus a wrapper method on `AutoSchema`.
Closes#5632
* Manual fields suggestions (#2)
* Use OrderedDict in inspectors
* Move empty check to 'update_fields()'
* Make 'update_fields()' an AutoSchema staticmethod
* Add 'AutoSchema.get_manual_fields()'
* Conform '.get_manual_fields()' to other methods
* Add test for update_fields
* Make sure `manual_fields` is a list.
(As documented to be)
* Add docs for new AutoSchema methods.
* `get_manual_fields`
* `update_fields`
* Add release notes for PR.
* Update assertion style in user logout test
* Apply middlewares to django request object
* Fix test for request auth hiding AttributeErrors
* Re-raise/wrap auth attribute errors
* Fix test for py2k
* Add docs for WrappedAttributeError
* SchemaJSRenderer renders invalid Javascript
Under Py3 the base64.b64encode() method returns a binary object, which gets rendered as `b'...'` in schema.js. This results in the output becoming:
var coreJSON = window.atob('b'eyJf...'');
which is invalid Javascript. Because base64 only uses ASCII characters it is safe to decode('ascii') it. Under Py2 this will result in a unicode object, which is fine. Under Py3 it results in a string, which is also fine. This solves the problem and results in a working schema.js output.
* Add regression test for #5608
* Add regression test for #5608
* Apparently the linter on Travis wants the imports in a different order than on my box...
* Remove trailing whitespace from lines
* Remove trailing nad leading whitespace from files
Allows for cleaner diffs in future changes. For editors that
automatically clean up whitespace on save, will avoid unrelated line
changes in diffs.
Set literals are available on all supported Python versions. They are
idiomatic and always faster:
$ python3 -m timeit '{}'
10000000 loops, best of 3: 0.0357 usec per loop
$ python3 -m timeit 'dict()'
10000000 loops, best of 3: 0.104 usec per loop
$ python3 -m timeit '{1, 2, 3}'
10000000 loops, best of 3: 0.0754 usec per loop
$ python3 -m timeit 'set([1, 2, 3])'
1000000 loops, best of 3: 0.228 usec per loop
Running the tests with bytes warning enabled shows some bytes/str
mixups. Fix them all.
Some examples of mixing usage:
str(b'foo') -- calling str() on bytes
b'foo' == 'foo' -- compare str with bytes
'foo' + b'bar' -- concatenating str and bytes
* Adding rounding parameter to DecimalField.
* Using standard `assert` instead of `self.fail()`.
* add testcase and PEP8 multilines fix
* flake8 fixes
* Use decimal module constants in tests.
* Add docs note for `rounding` parameter.
When running tests with warnings enabled, appear as:
DeprecationWarning: invalid escape sequence \d
Starting with Python 3.6, invalid escape sequences are deprecated. In a
future Python versions they will be a syntax error. For more details, see:
https://docs.python.org/3/whatsnew/3.6.html#deprecated-python-behavior
> A backslash-character pair that is not a valid escape sequence now
> generates a DeprecationWarning. Although this will eventually become a
> SyntaxError, that will not be for several Python releases.
Closes#5528.
Viewset custom actions (@detail_route etc) OPTIONS (and HEAD) methods were not being excluded from Schema Generations.
This PR adds a test reproducing the reported error and adjusts `EndpointEnumerator.get_allowed_methods()` to filter ViewSet actions in the same way as other `APIView`s
* substitute '@@' code block delimiter with triple back-tick, more consistent with other markdown extensions
* remove development print and allow spaces between triple backtick and syntax name in codeblock
* update comparison content for markdown test
* Fixed issues with schema name collisions
* Fixed mutating issues in python 3
* Optimized solution
* Fixed isort
* Removed not needed cast
* Fix for key collision
* Added preferred key to preserve if available
* Add accidently removed test
* Update remaing `include` calls
Missed as part of #5481 cleanup.
* Provide app_name in include_docs_urls
* Update remaining get_regex_pattern usages
* Allow functools.partial in is_simple_callable check
* Identify code that needs to be pulled out of/removed from compat.py
* Extract modern code from get_names_and_managers in compat.py and remove compat code
* Extract modern code from is_authenticated() in compat.py and remove.
* Extract modern code from is_anonymous() in compat.py and remove
* Extract modern code from get_related_model() from compat.py and remove
* Extract modern code from value_from_object() in compat.py and remove
* Update postgres compat
JSONField now always available.
* Remove DecimalValidator compat
* Remove get_remote_field compat
* Remove template_render compat
Plus isort.
* Remove set_many compat
* Remove include compat
* add 'docstrings-with-pygments' feature without packages checks and tests
* move syntax_highlight doc filter in compatibility module and define it conditionally
* typo fixed
* add test for optional code highlight ('pygments' and 'markdown' packages must be installed)
Closes#5237
Generics/ModelViewset performs filtering on: list, retrieve, put, patch and destroy (plus method equivalents).
i.e. on list plus anything that calls `get_object`.
This PR makes schema generation follow that.
It adds `AutoSchema._allows_filters()` which can be overridden in subclasses.
I’ve made this initially “private” so we can make quick changes if needs be in a 3.7.1 etc.
## Description
The `choices` field for the `ChoiceField` class should be able to be edited after `ChoiceField.__init__` is called.
```
field = ChoiceField(choices=[1,2])
field.choices = [1] # Should no longer allow `2` as a choice
```
Currently, you must update `choices`, `grouped_choices`, and `choice_strings_to_values` to achieve this. This P/R keeps `grouped_choices` and `choice_strings_to_values` in sync whenever the `choices` are edited.
* Add tests for schema exclusions
* Move exclusion check to should_include_endpoint
* Update docs
* Switch to using `schema = None`
* Test PendingDeprecationWarnings
* Add note to release notes.
* s/deprecated/pending deprecation/
* Add PR link to release notes
* Correct typo in test class name
* Test 'exclude_from_schema' deprecation warning message (#1)
* Correct deprecation warning message
* Initial Refactor Step
* Add descriptor class
* call from generator
* proxy back to generator for implementation.
* Move `get_link` to descriptor
* Move `get_description` to descriptor
* Remove need for generator in get_description
* Move get_path_fields to descriptor
* Move `get_serializer_fields` to descriptor
* Move `get_pagination_fields` to descriptor
* Move `get_filter_fields` to descriptor
* Move `get_encoding` to descriptor.
* Pass just `url` from SchemaGenerator to descriptor
* Make `view` a property
Encapsulates check for a view instance.
* Adjust API Reference docs
* Add `ManualSchema` class
* Refactor to `ViewInspector` plus `AutoSchema`
The interface then is **just** `get_link()`
* Add `manual_fields` kwarg to AutoSchema
* Add schema decorator for FBVs
* Adjust comments
* Docs: Provide full params in example
Ref feedback b52e372f8f (r137254795)
* Add docstring for ViewInstpector.__get__ descriptor method.
Ref https://github.com/encode/django-rest-framework/pull/5354#discussion_r137265022
* Make `schemas` a package.
* Split generators, inspectors, views.
* Adjust imports
* Rename to EndpointEnumerator
* Adjust ManualSchema to take `fields`
… and `description`.
Allows `url` and `action` to remain dynamic
* Add package/module docstrings
* Add regression test for #2505. Thanks @pySilver!
* Add regression test for #5087
* Revert "Cached the field's root and context property."
This reverts commit 792005806b.
