Django GraphBox is a package for easy building GraphQL APIs with Django. This package is based on Graphene and Graphene-Django.
It provides a SchemaBuilder that can be used to build a GraphQL schema from Django models. It also provides a SessionManager that can be used to manage access to the GraphQL API.
The basic idea of this package is to provide a simple way to build a GraphQL API with Django without the need to write a lot of code. The SchemaBuilder can be used to build a GraphQL schema from Django models in a few lines of code. The SessionManager can be used to manage access to the GraphQL API in a few lines of code.
SchemaBuilder is designed to create the basic CRUD operations for a model. It can be used to create a GraphQL schema with the basic CRUD operations for a model. It also provides a way to configure the fields that will be used on the schema. The SessionManager is designed to manage access to the GraphQL API. It can be used to manage access to the GraphQL API with a user model. It also provides a way to configure the groups and permissions that will be used to manage access to the GraphQL API.
The operations that can be created with the SchemaBuilder are:
- create_field (Mutation)
- update_field (Mutation)
- delete_field (Mutation)
- get_field (Query) this operation will be called with the name of the model, for example, if the model is called MyModel then the operation will be called my_model and the arguments will be id.
- all_field (Query) this operation will be called with the name of the model, for example, if the model is called MyModel then the operation will be called all_my_model and the arguments will be filters, page, and others depending on the configuration when the model was added to the SchemaBuilder.
When the SessionManager is used, the operations related to authentication are:
- login (Mutation)
- social_login (Mutation)
- actual_user (Query)
This package don't limit to use this basic operations. Custom operations can be defined on classic style of Graphene and Graphene-Django and finally can be merged on the main schema as described on the Quickstart section of this documentation at 4. Create a main schema in a new file called schema.py on my_project folder. This file can be used to merge all queries and mutations from all apps builded with django_graphbox or just add your own queries and mutations.
See the API REFERENCE at https://90horasporsemana.com/graphbox/
$ pip install django-graphbox
Use this guide to get started with the GraphBox.
-
Create a new Django project.
$ django-admin startproject my_project $ cd my_project $ python manage.py startapp my_app -
Define your Django models in the my_app app.
from django.db import models class MyModel(models.Model): ...
$ python manage.py makemigrations my_app $ python manage.py migrate
-
Configure and Build your GraphQL schema with django_graphbox.builder.SchemaBuilder on a new file called schema.py on the my_app app.
from django_graphbox.builder import SchemaBuilder from my_app.models import MyModel builder = SchemaBuilder() builder.add_model(MyModel) query_class = builder.build_schema_query() mutation_class = builder.build_schema_mutation()
-
Create a main schema in a new file called schema.py on my_project folder. This file can be used to merge all queries and mutations from all apps builded with django_graphbox or just add your own queries and mutations.
import graphene from my_app.schema import query_class, mutation_class class Query(query_class, graphene.ObjectType): pass class Mutation(mutation_class, graphene.ObjectType): pass schema = graphene.Schema(query=Query, mutation=Mutation)
-
Add the schema on urls.py file.
from django.urls import path from graphene_file_upload.django import FileUploadGraphQLView from django.views.decorators.csrf import csrf_exempt from .schema import schema urlpatterns = [ path('graphql/', csrf_exempt(FileUploadGraphQLView.as_view(graphiql=True, schema=schema))), ]
-
Run your project.
$ python manage.py runserver
Django GraphBox implements a SessionManager that can be used to manage access to the GraphQL API. This Manager is based on JWT authentication, so you have to send on Bearer format the token in the Authorization header.
Follow the steps below to create a new user and Manage the access to the GraphQL API.
-
Create your User model.
from django.models import Model class User(Model): custom_uname = models.CharField(max_length=100) custom_pwd = models.CharField(max_length=100) custom_active = models.BooleanField(default=True) role = models.CharField(max_length=100)
Note that you can define your fields as you want, and you will be able to configure this fields in the SessionManager.
-
Configure groups and modify_permissions in [settings.py]{.title-ref} file.
ACCESS_GROUPS = { "GROUP_LEVEL_1": ["RULE_LEVEL1"], "GROUP_LEVEL_2": ["RULE_LEVEL1", "RULE_LEVEL_2",], "GROUP_LEVEL_3": ["RULE_LEVEL1" ,"RULE_LEVEL_2", "RULE_LEVEL_3",], }
This groups can be interpreted as: If an operation like [create_field]{.title-ref} is configured for allow to GROUP_LEVEL_2 then the user will be able to create a field only if he has the role RULE_LEVEL_1 or RULE_LEVEL_2.
MODIFY_PERMISSIONS = { "ROLE_LEVEL_1": ["ROLE_LEVEL_3", "ROLE_LEVEL_2", "ROLE_LEVEL_1"], "ROLE_LEVEL_2": ["ROLE_LEVEL_3", "ROLE_LEVEL_2",], "ROLE_LEVEL_3": ["ROLE_LEVEL_1",], }
This permissions are related with the operations of the user model used on SessionManager. A user with the permission ROLE_LEVEL_2 only can create, update and delete user instances with the permission ROLE_LEVEL_2 and ROLE_LEVEL_3.
-
Create a new instance of the SessionManager on your schema.py file on the my_app app and configure the user model.
from django_graphbox.session import Manager as SessionManager from my_app.models import User from django.conf import settings session_manager = SessionManager(User, rol_field_name='role', login_id_field_name='custom_uname', password_field_name='custom_pwd', active_field_name='custom_active', groups=settings.ACCESS_GROUPS, modify_permissions=settings.MODIFY_PERMISSIONS)
-
Configure and Build your GraphQL schema with django_graphbox.builder.SchemaBuilder on the file called schema.py on the my_app app.
from django_graphbox.builder import SchemaBuilder from my_app.models import MyModel # Add the SessionManager to the SchemaBuilder builder = SchemaBuilder(session_manager=session_manager) # Build your operations builder.add_model(MyModel, access_group="GROUP_LEVEL_2") # This operation will be available only for users with the permission ROLE_LEVEL_1 or ROLE_LEVEL_2 builder.add_model( User, exclude_fields=('custom_pwd',), # Exclude this field on the builded ModelType save_as_password=['custom_pwd',], # On create and update this field will be saved as a password access_group="GROUP_LEVEL_2", access_by_operation={'delete_field': 'GROUP_LEVEL_1'} ) # This operation will be available only for users with the permission ROLE_LEVEL_1 or ROLE_LEVEL_2 except delete_field operation only for users with the permission ROLE_LEVEL_1. query_class = builder.build_schema_query() mutation_class = builder.build_schema_mutation() # Build your session operations session_query, session_mutation = builder.build_session_schema()
-
Create a main schema in a new file called schema.py on my_project folder. This file can be used to merge all queries and mutations from all apps builded with django_graphbox or just add your own queries and mutations.
import graphene from my_app.schema import query_class, mutation_class, session_query, session_mutation class Query(query_class, session_query, graphene.ObjectType): pass class Mutation(mutation_class, session_mutation, graphene.ObjectType): pass schema = graphene.Schema(query=Query, mutation=Mutation)
-
Add the schema on urls.py file.
from django.urls import path from graphene_file_upload.django import FileUploadGraphQLView from django.views.decorators.csrf import csrf_exempt from .schema import schema urlpatterns = [ path('graphql/', csrf_exempt(FileUploadGraphQLView.as_view(graphiql=True, schema=schema))), ]
-
Run the server and try to access the GraphQL API. Session operations will be available called actualUser query and login mutation. Additionally you can see the operations will require a valid access token and will validate the user role and permissions as you configured.
Django GraphBox Builder allows you to add custom filters and validators to the GraphQL schema. This example assumes that you have two models called [User]{.title-ref} and [Favorite]{.title-ref} with the following fields:
class User(Model): custom_uname = models.CharField(max_length=100) custom_pwd = models.CharField(max_length=100) custom_active = models.BooleanField(default=True) role = models.CharField(max_length=100) class Favorite(Model): book_name = models.CharField(max_length=100) book_author = models.CharField(max_length=100) book_year = models.IntegerField() user = models.ForeignKey(User, on_delete=models.CASCADE)
-
You can add external filters for the Favorite query. External filters are parameters that will be provided by the client and will be used to filter the query. The filters are added to the [external_filters]{.title-ref} dictionary on the add_model method like this:
builder.add_model( Favorite, external_filters={ { "field_name": "book_name", # The field name on the Favorite model "param_name": "book_name", # The parameter name on the query "param_type": graphene.String(required=True), # The parameter graphene type } } )
-
You can add internal filters for the Favorite query. Internal filters are callables that will be resolved on the query execution with the parameters of the query resolver. The filters are added to the [internal_filters]{.title-ref} dictionary on the add_model method like this:
builder.add_model( Favorite, internal_filters={ "field_name": "user__id", # The field name on the Favorite model "resolver_filter": session_manager.actual_user_attr_getter(field_name='id'), # This function of session_manager will return a function that return the id of the actual user "on_return_none": "skip", # If the function returns None, the filter will be skipped. If you want apply the filter like user__id__is_null=True, you can set this parameter to "set__isnull". } )
This will build the query allFavorite filtered by the actual user.
-
Build operations with custom validators by operation for a customizable workflow. The validators callables need receive [info]{.title-ref}, [model_instance]{.title-ref}, [**kwargs]{.title-ref} and must return a boolean.
builder.add_model( Favorite, validators_by_operation={ 'create_field': { 'validators':( session_manager.actual_user_comparer(actual_user_field='id', operator='=', model_field='user__id'), # This function of session_manager will return a function that compare the id of the actual user with the id of the user field of the Favorite model session_manager.actual_user_comparer(actual_user_field='role', operator='=', default_value='ROLE_LEVEL_1'), # This function of session_manager will return a function that compare the role of the actual user with the default value ), 'connector': 'OR', # The connector between the validators. If you want to use AND, you can set this parameter to 'AND'. }, } ) The validators are evaluated recursively, this allows you to create complex validators replacing the callable function with other dict with the same structure.
-
Build operations with internal resolvers for some fields of the model. For example to set the actual user as the owner of the Favorite. The resolver callables need receive [info]{.title-ref}, [model_instance]{.title-ref}, [**kwargs]{.title-ref} and must return a value as the model field type.
builder.add_model( Favorite, internal_field_resolvers={ 'create_field': { 'user': session_manager.actual_user_attr_getter(field_name='id'), # This function of session_manager will return a function that return the id of the actual user }, 'update_field': { 'user': session_manager.actual_user_attr_getter(field_name='id'), # This function of session_manager will return a function that return the id of the actual user }, } )
Note that the ForeignKey fields need return the id of the related model.
-
Build operations based on modify_permissions. For this example we will configure the User operations for allow create, update and delete to the actual user only if this has permission.
builder.add_model( User, validators_by_operation={ 'create_field': { 'validators':( session_manager.build_access_level_validator(model_field='role'), # This function of session_manager will return a function that compare the role of the actual user with the role of the User instance on the create operation ), 'connector': 'OR', # The connector between the validators. If you want to use AND, you can set this parameter to 'AND'. }, 'update_field': { 'validators':( session_manager.build_access_level_validator(model_field='role'), # This function of session_manager will return a function that compare the role of the actual user with the role of the User instance on the update operation ), 'connector': 'OR', # The connector between the validators. If you want to use AND, you can set this parameter to 'AND'. }, 'delete_field': { 'validators':( session_manager.build_access_level_validator(model_field='role'), # This function of session_manager will return a function that compare the role of the actual user with the role of the User instance on the delete operation ), 'connector': 'OR', # The connector between the validators. If you want to use AND, you can set this parameter to 'AND'. }, } ) SessionManager.build_access_level_validator(model_field='role') will return a function that will validate if the user_instance.role exists on the list of MODIFY_PERMISSIONS[actual_user.role].
This package has support for social login with Google and Facebook.
-
Create a project on Google Cloud Platform.
-
Create a OAuth 2.0 client ID on the Credentials section of the Google Cloud Platform.
-
Add the client ID and client secret on the settings.py file.
GOOGLE_CLIENT_ID='YOUR_CLIENT_ID'- Setup SessionManager to use social login:
session_manager = SessionManager(
Usuario,
use_social_session=True, # This will enable the social login
session_key='public-schema', # This is the key of the session
active_field_name='activo', # This is the name of the active field of the model
login_id_field_name='email', # This is the name of the field that will be used to login
rol_field_name='tipo', # This is the name of the field that will be used to set the role of the user
name_field_name='nombres', # This is the name of the field that will be used to set the name of the user (this will be set from response of the social login)
photo_field_name='foto', # This is the name of the field that will be used to set the photo of the user (this will be set from response of the social login)
social_id_field='social_id', # This is the name of the field that will be used to set the social id of the user (this will be set from response of the social login)
social_origin_field='social_origin', # This is the name of the field that will be used to set the social origin of the user (this will be send on social login mutation to identify the social login origin and can be 'FACEBOOK' or 'GOOGLE')
security_key=settings.SECRET_KEY, # Key used to generate the token
groups=PUBLIC_ACCESS_GROUPS, # Explained on the previous section
modify_permissions=PUBLIC_MODIFY_PERMISSIONS # Explained on the previous section
)- Use the session_manager to create the schema builder:
builder = SchemaBuilder(session_manager) # Explained on the previous section-
Create an app on Facebook Developers.
-
Configure the app to use Facebook Login.
-
Add FACEBOOK_APP_ID in settings.py file:
FACEBOOK_APP_ID='YOUR_APP_ID'- Setup SessionManager like the Google example. The social authentication will be the same, and the only difference is the social_origin_field that will be 'FACEBOOK' when the socialLogin mutation is called.
Django Graphbox has some extra features that can be used, but the documentation is not complete yet. If you want to use this features, you can see the code of the package to know how to use it or see the API Reference.
Some of the extra features are:
- Pagination for queries (infinitely scrolling and paginated)
- Moodle authentication (login with moodle credentials and get the user data from moodle)
- Internal filters (filters that can be used to filter the data of the query based on a callable resolver)
- Validators by operation (validators that can be used to validate the data of the mutation based on a callable resolver)
- Internal field resolvers (resolvers that can be used to resolve the data of the field based on a callable resolver)
- Callbacks by operation (callbacks that can be used to execute a callable resolver after the mutation is executed)
- Automatic integration with Django Auditor Logs (A package that can be used to log the data changes on the models, maintained by the same author of this package)
- Version 1.0.0 to 1.1.5 was a package developed for a specific project, and the code was not published on GitHub. The code was refactored and published on GitHub on version 1.2.0.
- Version 1.2.3 add support to set custom attributes on the model Type and set custom ordering field for the queries.
- Version 1.2.4 Fix custom attributes on the model Type.
- Version 1.2.5 Add support to select the operations to build for the model. You can select between field_by_id, list_field, create_field, update_field and delete_field operations. By default all operations are selected.