django-graphene-filters 0.7.5

Advanced auto related filters for Graphene Django

Django Graphene Filters (Beta)

This package contains Advanced auto related filters for graphene-django.

This package takes inspiration from:

• https://pypi.org/project/graphene-django-filter
• https://pypi.org/project/djangorestframework-filters

Installation

# pip
pip install django-graphene-filters
# uv
uv add django-graphene-filters

Development Setup

# Install uv (if not already installed)
# https://docs.astral.sh/uv/getting-started/installation/

# Clone and install
git clone https://github.com/riodw/django-graphene-filters.git
cd django-graphene-filters
uv sync
uv sync --upgrade

Running

uv run python examples/cookbook/manage.py runserver

Seeding the example database

The cookbook example dynamically discovers all Faker providers at runtime and seeds the database accordingly. The command is idempotent — it ensures at least N objects exist per provider and only creates the shortfall.

# Ensure 5 objects per provider (default)
uv run python examples/cookbook/manage.py seed_data

# Ensure 50 objects per provider
uv run python examples/cookbook/manage.py seed_data 50

# Delete the first 10 objects (and their cascading values)
uv run python examples/cookbook/manage.py delete_data 10

# Delete all objects and values
uv run python examples/cookbook/manage.py delete_data all

# Wipe all four tables
uv run python examples/cookbook/manage.py delete_data everything

Test users

Create test users with individual Django view_* permissions for exercising get_queryset permission branches. Each set creates 6 users: 1 staff, 1 regular (no perms), and 4 per-model permission users. All share password admin. Superusers are never deleted.

# Create 1 set of test users (6 users)
uv run python examples/cookbook/manage.py create_users

# Create 3 sets (18 users)
uv run python examples/cookbook/manage.py create_users 3

# Delete all non-superusers
uv run python examples/cookbook/manage.py delete_users all

# Delete the first 5 non-superusers
uv run python examples/cookbook/manage.py delete_users 5

Testing

uv run coverage run -m pytest
uv run coverage report --fail-under=99
uv run coverage report --show-missing
# run on a single test file
uv run coverage run -m pytest tests/test_input_data_factories.py && uv run coverage report -m django_graphene_filters/input_data_factories.py
# Run Full Test Pipeline
https://github.com/riodw/django-graphene-filters/actions/workflows/django.yml

Formatting and Linting:

# pyproject.toml [tool.ruff]
uv run ruff format .
uv run ruff check --fix .

Updating Version:

• pyproject.toml:4
• django_graphene_filters/__init__.py:22
• tests/test_django_graphene_filters.py:8

Build

uv lock
rm dist/*
uv build

Publish

uv publish --token PASSWORD

Updating:

# Show outdated packages
uv pip list --outdated

# Add a dev dependency
uv add --group dev <package>

# Remove the virtual environment
rm -rf .venv

Local usage

1. go to the project you want to install the package
2. add django-graphene-filters to your pyproject.toml dependencies
3. point it at your local checkout:

# In your project's pyproject.toml
[tool.uv.sources]
django-graphene-filters = { path = "../django-graphene-filters", editable = true }

Then run:

uv sync

Permissions & Cascade Visibility

When a node's get_queryset hides rows (e.g. is_private=False for non-staff), FK fields pointing to hidden targets would normally cause "Cannot return null for non-nullable field" errors.

AdvancedDjangoObjectType solves this with sentinel nodes — redacted instances (pk=0) returned by get_node when the real row is hidden. Every node exposes an isRedacted: Boolean! field so clients can detect them.

Use apply_cascade_permissions inside get_queryset to proactively exclude rows whose FK targets are hidden:

from django_graphene_filters import AdvancedDjangoObjectType, apply_cascade_permissions

class ObjectNode(AdvancedDjangoObjectType):
    class Meta:
        model = Object
        interfaces = (Node,)
        fields = "__all__"

    @classmethod
    def get_queryset(cls, queryset, info):
        user = getattr(info.context, "user", None)
        if user and user.is_staff:
            return queryset
        return apply_cascade_permissions(cls, queryset.filter(is_private=False), info)

See the CHANGELOG for full details on the 0.4.0 permissions system.