Problem with django-docutils (and thus abandoned here):
When calling the filter, the app used ``return parts['fragments']``
which cut the first headlines from the text input.
The correct call is ``return parts['html_body']``, as implemented
in the new self-made filter.

On staging, everything is started by running ~/start_celery.sh - should
work now for testing.

Note that as of now, it requires 3 services (to do with Celery) to run:
celery -A SciPost_v1 worker --loglevel=info -E
celery -A SciPost_v1 beat --loglevel=info --scheduler django_celery_beat.schedulers:DatabaseScheduler
And optionally (for using the 'flower' console for monitoring tasks -
might not use this after all):
celery -A SciPost_v1 flower

Additionally, the 'rabbitMQ' broker needs to be installed - working on
it on the staging at the moment.

I will add this in some proper config/readme later.

Passwords will be hashed with the Argon2 library.
https://docs.djangoproject.com/en/1.10/topics/auth/passwords/#using-argon2-with-django

Also, new passwords will go through 'common-passwords-checks'. This will prevent
access to accounts using one of those common brute force password hackers around,
in case of a breach.