django-api-forms 1.0.0rc7

Declarative Django request validation for RESTful APIs

Django API Forms

Django Forms approach in the processing of a RESTful HTTP request payload (especially for content type like JSON or MessagePack) without HTML front-end.

Motivation

The main idea was to create a simple and declarative way to specify the format of expecting requests with the ability to validate them. Firstly, I tried to use Django Forms to validate my API requests (I use pure Django in my APIs). I have encountered a problem with nesting my requests without a huge boilerplate. Also, the whole HTML thing was pretty useless in my RESTful APIs.

I wanted to:

• define my requests as object (Form),
• pass the request to my defined object (form = Form.create_from_request(request)),
• validate my request form.is_valid(),
• extract data form.clean_data property.

I wanted to keep:

• friendly declarative Django syntax, (DeclarativeFieldsMetaclass is beautiful),
• Validators,
• ValidationError,
• Form fields (In the end, I had to "replace" some of them).

So I have decided to create a simple Python package to cover all my expectations.

Installation

# Using pip
pip install django-api-forms

# Using poetry
poetry add django-api-forms

# Local installation
python -m pip install .

Optional:

# msgpack support (for requests with Content-Type: application/x-msgpack)
poetry add msgpack

# ImageField support
poetry add Pillow

Install application in your Django project by adding django_api_forms to yours INSTALLED_APPS:

INSTALLED_APPS = (
    'django.contrib.auth',
    'django.contrib.contenttypes',
    'django.contrib.sessions',
    'django.contrib.messages',
    'django_api_forms'
)

You can change the default behavior of population strategies or parsers using these settings (listed with default values). Keep in mind, that dictionaries are not replaced by your settings they are merged with defaults.

For more information about the parsers and the population strategies check the documentation.

DJANGO_API_FORMS_POPULATION_STRATEGIES = {
    'django_api_forms.fields.FormFieldList': 'django_api_forms.population_strategies.IgnoreStrategy',
    'django_api_forms.fields.FileField': 'django_api_forms.population_strategies.IgnoreStrategy',
    'django_api_forms.fields.ImageField': 'django_api_forms.population_strategies.IgnoreStrategy',
    'django_api_forms.fields.FormField': 'django_api_forms.population_strategies.IgnoreStrategy',
    'django.forms.models.ModelMultipleChoiceField': 'django_api_forms.population_strategies.IgnoreStrategy',
    'django.forms.models.ModelChoiceField': 'django_api_forms.population_strategies.ModelChoiceFieldStrategy'
}

DJANGO_API_FORMS_DEFAULT_POPULATION_STRATEGY = 'django_api_forms.population_strategies.BaseStrategy'

DJANGO_API_FORMS_PARSERS = {
    'application/json': 'json.loads',
    'application/x-msgpack': 'msgpack.loads'
}

Example

Simple nested JSON request

{
  "title": "Unknown Pleasures",
  "type": "vinyl",
  "artist": {
    "_name": "Joy Division",
    "genres": [
      "rock",
      "punk"
    ],
    "members": 4
  },
  "year": 1979,
  "songs": [
    {
      "title": "Disorder",
      "duration": "3:29"
    },
    {
      "title": "Day of the Lords",
      "duration": "4:48",
      "metadata": {
        "_section": {
          "type": "ID3v2",
          "offset": 0,
          "byteLength": 2048
        },
        "header": {
          "majorVersion": 3,
          "minorRevision": 0,
          "size": 2038
        }
      }
    }
  ],
  "metadata": {
    "created_at": "2019-10-21T18:57:03+0100",
    "updated_at": "2019-10-21T18:57:03+0100"
  }
}

Django API Forms equivalent + validation

from enum import Enum

from django.core.exceptions import ValidationError
from django.forms import fields

from django_api_forms import FieldList, FormField, FormFieldList, DictionaryField, EnumField, AnyField, Form


class AlbumType(Enum):
    CD = 'cd'
    VINYL = 'vinyl'


class ArtistForm(Form):
    class Meta:
        mapping = {
            '_name': 'name'
        }

    name = fields.CharField(required=True, max_length=100)
    genres = FieldList(field=fields.CharField(max_length=30))
    members = fields.IntegerField()


class SongForm(Form):
    title = fields.CharField(required=True, max_length=100)
    duration = fields.DurationField(required=False)
    metadata = AnyField(required=False)


class AlbumForm(Form):
    title = fields.CharField(max_length=100)
    year = fields.IntegerField()
    artist = FormField(form=ArtistForm)
    songs = FormFieldList(form=SongForm)
    type = EnumField(enum=AlbumType, required=True)
    metadata = DictionaryField(value_field=fields.DateTimeField())

    def clean_year(self):
        if self.cleaned_data['year'] == 1992:
            raise ValidationError("Year 1992 is forbidden!", 'forbidden-value')
        return self.cleaned_data['year']

    def clean(self):
        if (self.cleaned_data['year'] == 1998) and (self.cleaned_data['artist']['name'] == "Nirvana"):
            raise ValidationError("Sounds like a bullshit", code='time-traveling')
        if not self._request.user.is_authenticated():
            raise ValidationError("You can use request in form validation!")
        return self.cleaned_data



"""
Django view example
"""
def create_album(request):
    form = AlbumForm.create_from_request(request)
    if not form.is_valid():
        # Process your validation error
        print(form.errors)

    # Cleaned valid payload
    payload = form.cleaned_data
    print(payload)

If you want example with whole Django project, check out repository created by pawl django_api_forms_modelchoicefield_example, where he uses library with ModelChoiceField.

Running Tests

# install all dependencies
poetry install

# run code-style check
poetry run flake8 .

# run the tests
poetry run python runtests.py

Sponsorship

Navicat Premium is a super awesome database development tool for cool kids in the neighborhood that allows you to simultaneously connect to MySQL, MariaDB, MongoDB, SQL Server, Oracle, PostgreSQL, and SQLite databases from a single application. Compatible with cloud databases like Amazon RDS, Amazon Aurora, Amazon Redshift, Microsoft Azure, Oracle Cloud, Google Cloud and MongoDB Atlas. You can quickly and easily build, manage and maintain your databases.

Especially, I have to recommend their database design tool. Many thanks Navicat for supporting Open Source projects 🌈.

---

Made with ❤️ and ☕️ by Jakub Dubec, BACKBONE s.r.o. & contributors.