django-anon

Anonymize production data so it can be safely used in not-so-safe environments

Install | Read Documentation | PyPI | Contribute

django-anon will help you anonymize your production database so it can be shared among developers, helping to reproduce bugs and make performance improvements in a production-like environment.

Features

🚀	Really fast data anonymization and database operations using bulk updates to operate over huge tables
🍰	Flexible to use your own anonymization functions or external libraries like Faker
🐩	Elegant solution following consolidated patterns from projects like Django and Factory Boy
🔨	Powerful. It can be used on any projects, not only Django, not only Python. Really!

Table of Contents

• Installation
• Supported versions
• License
• Usage
  • Built-in functions
  • Lazy attributes
  • The clean method
  • Defining a custom QuerySet
  • High-quality fake data
• Changelog
• Contributing

Installation

pip install django-anon

Supported versions

• Python 3.7+
• Django 3.0+

License

MIT

Usage

Use anon.BaseAnonymizer to define your anonymizer classes:

import anon

from your_app.models import Person

class PersonAnonymizer(anon.BaseAnonymizer):
   email = anon.fake_email

   # You can use static values instead of callables
   is_admin = False

   class Meta:
      model = Person

# run anonymizer: be cautious, this will affect your current database!
PersonAnonymizer().run()

Built-in functions

import anon

anon.fake_word(min_size=_min_word_size, max_size=20)
anon.fake_text(max_size=255, max_diff_allowed=5, separator=' ')
anon.fake_small_text(max_size=50)
anon.fake_name(max_size=15)
anon.fake_username(max_size=10, separator='')
anon.fake_email(max_size=40, suffix='@example.com')
anon.fake_url(max_size=50, scheme='http://', suffix='.com')
anon.fake_phone_number(format='999-999-9999')

Lazy attributes

Lazy attributes can be defined as inline lambdas or methods, as shown below, using the anon.lazy_attribute function/decorator.

import anon

from your_app.models import Person

class PersonAnonymizer(anon.BaseAnonymizer):
   name = anon.lazy_attribute(lambda o: 'x' * len(o.name))

   @lazy_attribute
   def date_of_birth(self):
      # keep year and month
      return self.date_of_birth.replace(day=1)

   class Meta:
      model = Person

The clean method

import anon

class UserAnonymizer(anon.BaseAnonymizer):
   class Meta:
      model = User

   def clean(self, obj):
      obj.set_password('test')
      obj.save()

Defining a custom QuerySet

A custom QuerySet can be used to select the rows that should be anonymized:

import anon

from your_app.models import Person

class PersonAnonymizer(anon.BaseAnonymizer):
   email = anon.fake_email

   class Meta:
      model = Person

   def get_queryset(self):
      # keep admins unmodified
      return Person.objects.exclude(is_admin=True)

High-quality fake data

In order to be really fast, django-anon uses it's own algorithm to generate fake data. It is really fast, but the generated data is not pretty. If you need something prettier in terms of data, we suggest using Faker, which can be used out-of-the-box as the below:

import anon

from faker import Faker
from your_app.models import Address

faker = Faker()

class PersonAnonymizer(anon.BaseAnonymizer):
   postalcode = faker.postalcode

   class Meta:
      model = Address

Changelog

Check out CHANGELOG.rst for release notes

Contributing

Check out CONTRIBUTING.rst for information about getting involved

---

Icon made by Eucalyp from www.flaticon.com