django-qserializer

Django QSerializer started as an internal Buser project to improve our serialization process.

Database queries and serialization are two separated steps, but really ORM coupled. Relationships must be fetched from database before serialization, but Django don't have an easy way to setup that.

Start with a custom manager SerializableManager.

from django.db import models
from django_qserializer import SerializableManager


class Company(models.Model):
    name = models.CharField(max_length=64)


class Bus(models.Model):
    objects = SerializableManager()
    company = models.ForeignKey(Company, on_delete=models.SET_NULL)

A basic serializer implementation would be:

from django_qserializer import BaseSerializer


class BusSerializer(BaseSerializer):
    select_related = ['company']

    def serialize_object(self, obj):
        return {
            'id': obj.id,
            'company': {
                'name': obj.company.name,
            }
        }

Add the serializer to your queryset as:

buses = Bus.objects.to_serialize(BusSerializer).all()

for bus in buses:
    # The serialize method is bound to BusSerializer.serialize_object.
    print(bus.serialize())

API

BaseSerializer.select_related

List of model fields to add to queryset with a select_related call.

BaseSerializer.prefetch_related

List of model fields to add to queryset with a prefetch_related call.

class BusSerializer(BaseSerializer):
    prefetch_related = ['company']

    def serialize_object(self, obj):
        return {
            'id': obj.id,
            'company': {
                'name': obj.company.name,
            }
        }

BaseSerializer.prepare_queryset

Callable to change the queryset. It is possible to implement select_related and prefetch_related attributes with it, but they work together with prepare_queryset.

class BusSerializer(BaseSerializer):
    select_related = ['company']

    def prepare_queryset(self, qs):
        return qs.annotate(state=Value('broken'))

BaseSerializer.prepare_objects

Prepare objects after they are loaded to memory. Add data in bulk to them, like fetching information from cache and attaching to loaded objects.

BaseSerializer.serialize_object

Required implementation. It converts the Django model to a serializable dict. Avoid slow calls here because it will cause N+1 issues.

BaseSerializer.serialize

Execute serialize_object for each model object.

Development

To run the project, it is necessary the following tools:

• Python 3.9 or higher

To create a virtual environment, run

python3  -m venv .venv

To activate the virtual environment, run

source .venv/bin/activate

To install the project's requirements in the virtual environment, run

pip install -e .\[test\]

To deactivate the virtual environment, run deactivate.

Run the commands of the following sections with the virtual environment active.

Quality

The quality metrics of the project are reproduced by the continuos integration (CI) pipeline of the project. CI configuration in .github/workflows/ci.yml file.

Tests and linter

To run tests, coverage report and linter, run

pytest

To see the html report, check htmlcov/index.html.

Tests and coverage configuration in setup.cfg file, at [tool:pytest] section.

Linter configuration in setup.cfg file, at [flake8] section.

License

This repository is licensed under the terms of MIT License.