preply/dj_anonymizer

Anonymize django database. Big-data anonymizer. Preply team.

dj_anonymizer

dj_anonymizer is a utility designed to anonymize production databases with various types of mock data, specifically designed for use within Django projects.

Project works with Django 4.2 or higher and Python 3.8 or higher.

Documentation

https://dj-anonymizer.readthedocs.io/en/latest/

dj_anonymizer's Issues

Django 2.2 and django-bulk-update dependency

Since Django 2.2 bulk-update() method will be available out the box for any Model.
Need to remove django-bulk-update dependency for Django>=2.2

Check model_fields list before start anonymization process

Need to add the check for comparing specified_fields and model_fields before start execution anonymization cause anonymization process can raise an exception in the middle of it in case of different list of existing and specified fields

Possibility to truncate tables for register_clean

By default register_clean is doing DELETE FROM "table_name" here
Need to add possibility do it via TRUNCATE TABLE "table_name"

Possibility to execute anonymizer on specified table

it may be done in a way like described below

python manage.py anonymize_db --only django.contrib.auth.models.User

Handle cases when table does not exist in database

If dump is made not from latest database snapshot there is a possibility that a table exists in code but does not exist in the database.
Now in such cases anonymization execution fails
Think it can be mitigated with an additional flag for anonymizer and/or warning during execution

ProtectedError on clean step if model is referenced throught a protected Foreign key

In cases like

class Musician(models.Model):
    instrument = models.CharField(max_length=100)

class Album(models.Model):
    artist = models.ForeignKey(
        'app.Musician',
        on_delete=models.PROTECT
    )

Order matters for cleaning models

For example

from dj_anonymizer import register_models

register_models.register_clean([
    Musician, Album
])

Will raise

django.db.models.deletion.ProtectedError: ("Cannot delete some instances of model 'Musician' because they are referenced through a protected foreign key: ...

And in case

from dj_anonymizer import register_models

register_models.register_clean([
    Album, Musician
])

All will be good

Add new anonym_field - password field

For anonymized DB will be great to have same password for all users
It will help easilly login via any user by knowing username/password

now we suport two type of fileds:

• anonym_field.function
• anonym_field.string

and it's possible to do via anonym_field.function
but easier way is to support anonym_field.password out the box