drf-madprops 1.2.0

DRF library to operate resource's properties as a dictionary

drf-madprops

DRF library of helpers to operate on lists of resource properties as dicts

Written and used by the folks at Yola to support our free website builder.

Overview

It's a typical case for relational DBs when some table (e.g. User) is extended via subordinate key-value table (e.g. UserPreference). This allows to dynamically add/delete fields to User (stored in UserPreference). Usually those property models have very simple structures, e.g:

class UserPreference(models.Model):
    id = models.AutoField()
    user = models.ForeignKey(User, related_name='preferences')
    name = models.CharField()
    value = models.CharField)

But it's not very convenient to expose/operate on them via standard DRF serializers. We'll get something like:

[
    {"id": "id1", "user": "user1", "name": "property1", "value": "value1"},
    {"id": "id2", "user": "user1", "name": "property2", "value": "value2"}
    ...
]

This library contains two base classes for property's serializers (for cases when properties are exposed as separate resource and as nested resource) which allows to retrieve/operate on parent resource properties as dicts. For example, instead of representation listed above, we'll get something like:

{
    "property1": "value1",
    "property2": "value2",
    ...
}

Usage

Additional meta options

• read_only_props: list of property names, which values cannot be changed via serializer.
• json_props: list of property names, which values are stored as JSON. Serializer will json.loads() those values when converting to representation and will json.dumps() them before saving.

As a nested serializer

from madprops.serializers import NestedPropertySerializer, PropertiesOwnerSerializer


class PreferenceSerializer(NestedPropertySerializer):
    class Meta:
        model = Preference
        read_only_props = ('user_token', 'tutorial_email_sent')
        json_props = ('packages',)


class UserSerializer(PropertiesOwnerSerializer):
    preferences = PreferenceSerializer(many=True, required=False)

As a serializer used for properties endpoint

from madprops.serializers import PropertySerializer


class PreferenceSerializer(PropertySerializer):
    class Meta:
        model = Preference
        read_only_props = ('user_token', 'tutorial_email_sent')


class PreferencePrivateViewSet(ModelViewSet):
    serializer_class = PreferenceSerializer
    paginate_by = None

Testing

Run the tests with:

tox

Or install development requirements and run tests with:

pip install -r requirements.txt
pytest tests