django-encrypted-json-fields 1.0.4

A set of fields that wrap standard Django fields, including JSONField, with encryption provided by the python cryptography library

Django Encrypted Model Fields (including JSONField)

About

This is a fork of https://gitlab.com/lansharkconsulting/django/django-encrypted-model-fields, which in turn was a fork of https://github.com/foundertherapy/django-cryptographic-fields.

It has been renamed, and updated to properly support Python3 and the latest versions of Django.

django-encrypted-json-fields is set of fields that wrap standard Django fields with encryption provided by the python cryptography library. These fields are much more compatible with a 12-factor design since they take their encryption key from the settings file instead of a file on disk used by keyczar.

While keyczar is an excellent tool to use for encryption, it's not compatible with Python 3, and it requires, for hosts like Heroku, that you either check your key file into your git repository for deployment, or implement manual post-deployment processing to write the key stored in an environment variable into a file that keyczar can read.

JSONField support

django-encrypted-json-fields extends the origin project django-encrypted-model-fields by adding a specific support for JSONFields, with the following features:

• the encrypted data remains a valid JSON, so you can inherit from django.db.models.JSONField and all validations will still work
• if the data contains dictionaries, the keys are preserved so that the overall structure remains intact
• that is: we only encrypt the values

Implementation notes

I opted to encrypt the repr() of the values, then apply eval() later only (after decrypting).

This is usefull to reconstruct both the value and the type; since JSON manages only a few simple types, this naive solution just fits the bill.

The crypter

All functions responsible for encryption/decryption (see below) require a crypter, which can be obtained in a few ways:

• default crypter: assign a key or a list of keys to the EJF_ENCRYPTION_KEYS setting, and a default crypter will be build for you
• assigning a callable to the EJF_ENCRYPTION_KEYS setting, which in turn will return a list of keys as above
• invoke build_crypter(keys) explicitly, and pass the resulting object around

For the latter, the use case I had in mind was the need to keep the data in play text on the server, and export encrypted data for a remote client, sharing a common key.

Deferred get_crypter()

Since EJF_ENCRYPTION_KEYS setting now accepts a callable, which might very well need to retrieve some data from the Django models, I had to postpone the call to get_crypter() until all apps have been loaded.

As a side effect, now you can always and safely call the generate_encryption_key management command (see below)

App settings

EJF_ENCRYPTION_KEYS

either a key, a list of keys, or a callable returning the list of keys to
be used for building the default crypter

EJF_DISABLE_ENCRYPTION

skip encryption when saving the model (save data unencrypted)

Helpers

All function used internally when saving and reading Django models can also be invoked explicitly to apply encryption/decryption to arbitrary strings or JSON values.

A possible use case consists in serializing encrypted data to be sent to a remote client.

Function	Purpone
generate_random_encryption_key()	generate a key
build_crypter(keys)	given a list of keys (or a key) builds the corresponding crypter
is_encrypted(s: Union[str, bytes]) -> bool	Check if the given string (or bytes) is the result of an encryption
encrypt_str(s: str, crypter=None, force=False) -> bytes	Encrypts the given string applying either the supplied crypter or, in None, the default crypter
decrypt_bytes(t: bytes, crypter=None, force=False) -> str	Decrypts the given bytes and returns a string
encrypt_values(data, crypter=None, force=False, json_skip_keys=None)	Applyes encryption to a JSON-serializable object
decrypt_values(data, crypter=None, force=False)	reverses encrypt_values()

• force means: proceed even when encryption is disabled in project's settings

Managment commands

Some management commands are supplied; run with --help for detailed informations:

• generate_encryption_key
• encrypt_str
• decrypt_str
• encrypt_all_tables
• decrypt_all_tables

Getting Started

$ pip install django-encrypted-json-fields

Add "encrypted_json_fields" to your INSTALLED_APPS setting like this:

    INSTALLED_APPS = (
        ...
        'encrypted_json_fields',
    )

django-encrypted-json-fields expects the encryption key to be specified using FIELD_ENCRYPTION_KEY in your project's settings.py file. For example, to load it from the local environment:

    import os

    FIELD_ENCRYPTION_KEY = os.environ.get('FIELD_ENCRYPTION_KEY', '')

To use an encrypted field in a Django model, use one of the fields from the encrypted_json_fields module:

    from encrypted_json_fields.fields import EncryptedCharField

    class EncryptedFieldModel(models.Model):
        encrypted_char_field = EncryptedCharField(max_length=100)

For fields that require max_length to be specified, the Encrypted variants of those fields will automatically increase the size of the database field to hold the encrypted form of the content. For example, a 3 character CharField will automatically specify a database field size of 100 characters when EncryptedCharField(max_length=3) is specified.

Due to the nature of the encrypted data, filtering by values contained in encrypted fields won't work properly. Sorting is also not supported.

Running Tests

Does the code actually work?

Running the unit tests from this app:

python manage.py test -v 2

or

./runtests.py

or

coverage run --source='.' runtests.py
coverage report

Sample project

A sample project using this app for demonstration purposes is available here:

https://github.com/morlandi/test_django_encrypted_json_fields

Credits

• https://gitlab.com/lansharkconsulting/django/django-encrypted-model-fields has been shared by Scott Sharkey
• https://github.com/foundertherapy/django-cryptographic-fields has been shared by Dana Spiegel

django-encrypted-model-fields Changelog

• 1.0.4 - fix fetch_raw_field_value() for sqlite
• 1.0.3 - prevent double encryption
• 1.0.2 - "force" flag added
• 1.0.0 - helpers available as standalone funcs
  • major refactoring
  • more unit tests
• 0.7.2 - More hacks to optionally disable encryption of JSON fields
• 0.7.1 - Overridable crypter
  • Optionally disable encryption with FIELD_SKIP_ENCRYPTION
• 0.7.0 - JSONField support
  • deferred get_crypter()
  • FIELD_ENCRYPTION_KEY setting now accepts a callable
• 0.6.5 - RE-Correct links to repository and homepage
• 0.6.4 - Correct links to repository and homepage
• 0.6.3 - Converted to poetry for build system
• 0.6.2 - Bump support to any python < 4
  • Make Generating an Encryption Key more prominent
• 0.6.1 - Add Support for Python up to 3.11
• 0.6.0 - Major Update
  • Drop Python 2 Support
  • Drop support for Django < 2.2
  • Add support for Django 2.2, 3.0, 3.1, and 3.2
  • Drop support for Python < 3.6
  • Add Support for Python 3.6, 3.7, 3.8, and 3.9
  • Move travis-ci testing to gitlab
  • Remove Support for EncryptedNullBooleanField (deprecated by Django 4.0)
  • Update Test App to Django 2.2 standard
  • Include 2 of Oleg Pesok's fixes: for Timezone-aware datetimes, and the cached validator crash
• 0.5.8 - Move to GITLAB repository
• 0.5.7 - rearrange requirements and update dependencies
  • fix issue with output of generate_encryption_key command
  • define default value for FIELD_ENCRYPTION_KEY
  • restore support for Django 1.9, 1.10, and 1.11
• 0.5.6 - Fixed issue with generating encryption keys (thanks Dave Alan)
  • dropped support for Django 1.8, 1.9 and 1.10
  • fix classifier to include Django 2.0
• 0.5.5 - Fixed README
• 0.5.4 - drop python 3.4 support
  • add django 2.0 support
  • expand testing to cover python 2.7, 3.5 and 3.6 for django 1.9, 1.10, 1.11 and 2.0
  • fix broken test
• 0.5.3 - Fork from foundertherapy/django-cryptographic-fields
  • update dependencies
  • convert to run under python 3.x
  • rename package so it can be uploaded to pypi
  • add support for tox/travis-ci testing