guardrail is a Python library for managing object-level permissions that's designed to integrate with arbitrary databases and web frameworks. guardrail is inspired by django-guardian and currently supports the SQLAlchemy, Peewee, Pony, and Django ORMs.
Homepage: https://guardrail.readthedocs.org/
pip install guardrail
guardrail supports Python >= 2.7 or >= 3.3 and pypy.
Define your models as usual, using the registry.agent and registry.target decorators to set up permissions relationships:
import peewee as pw
from guardrail.core import registry
from guardrail.ext.peewee import PeeweePermissionSchemaFactory
database = pw.SqliteDatabase(':memory:')
class Base(pw.Model):
class Meta:
database = database
@registry.agent
class User(Base):
name = pw.CharField()
@registry.target
class Post(Base):
name = pw.CharField()
@registry.target
class Comment(Base):
name = pw.CharField()
factory = PeeweePermissionSchemaFactory((Base, ))
registry.make_schemas(factory)
database.connect()
database.create_tables([User, Post, Comment], safe=True)
database.create_tables(registry.permissions, safe=True)Then use the permission manager to perform CRUD operations on permissions between any agent and target models:
from guardrail.ext.peewee import PeeweePermissionManager
manager = PeeweePermissionManager()
user = User.create(name='fred')
post = Post.create(name='news of the world')
comment = Comment.create(name='dragon attack')
manager.add_permission(user, post, 'edit')
manager.add_permission(user, comment, 'delete')
manager.has_permission(user, post, 'edit') # True
manager.remove_permission(user, comment, 'delete')
manager.has_permission(user, comment, 'delete') # FalseMIT licensed. See the LICENSE file for details.