The current process to migrate the data suffers from side effects,
especially when the database needs to drop, start from scratch and add
some migrations through Django. Django handles really badly an existing
database and it's impossible to reproduce a migration process from
scratch when django migrations had been add up to the current database
migration process.

This commit resolve this by reversing the process from migrating to old
database to the new one and telling django to take this new database as
it is to telling Django to create the database from his models and
then inserting the data into the new database. By doing so, any
migration applied through Django is taken into account at any stage
during the process.

During the rewriting of the migration process I discovered several
issues that needed to be addressed.

- deletion of the token management library
- add of a token field in the allgo user table
- add of foreignkeys instead of onetoone fiels in the webapp to avoid a
uniqueness issue on fields that were supposed to have several times the
same ID.
- Update of the default values of job's fields

Signed-off-by: Matthieu Berjon <matthieu.berjon@inria.fr>

Further cleaning & integration in the django container

See merge request !28

The script is now well-structured (with functions!) and can be used
to reset a single container, without requiring a bootstrap of the
whole environment.

    usage: $0 [CONTAINER ...]

    The bootstrap script initialises the environment and the selected
    containers (or by default all the containers).

    If a container is already initialised, then the script asks for
    confirmation before purging its data (to bootstrap it again).

This commit also merges the 'prepare.sh' and 'init-container' scripts
into 'bootstrap'.

(to store the secret_key)

mysql-client is needed because the import script uses mysqldump

(thus we'll have security updates)

- install django statics into /var/www at image build time
- django now on port 8000 (django's default) instead of 4000
- rename @allgo as @Django in nginx's config to decrease confusion
- publish port 8080 in docker-compose

- django is now launch in the foregroud, and its stdout/stderr is now
  the container stdout/stderr

- nginx is now run as a background process, without supervision. A
  HEALTHCHECK CMD is added to monitor the status of nginx (so as to
  a nagios alert in case an hypothetical nginx crash)

- add ALLGO_HTTP_SERVER to select the http server. Default is
 'gunicorn', docker-compose.yml sets it to 'django' for development.

and define them in config.env with a default value

it is unused and cause extra delayed when loading the page
(because they load some files from the cloud, which does not
 respond very well)

- env variables are defined in 'config/env.py', this includes:
  - their default value
  - their documentation

- the loaded variables are stored as attributes of this module
  (eg: config.env.ALLGO_DEBUG)

- all variables are guaranteed to be set once the module is loaded
  thus, the config.settings just has to use env.ALLGO_XXXXX

- the sphinx documentation of the variables is generated directly
  from the source code

- the plumbing is done by the config.env_loader module (while loads
  the variables, and patches the docstring of the config.env module)

+ensure that ENV is defined

(to suppress a warning at runtime)

my setup seems to require it (did not investigate further, and this
will be void if we switch to nginx+uwsgi)

(rather than everytime the container is run)

it's easier to read and easier to comment

# Conflicts:
#	django/Dockerfile
#	docker-compose.yml

Signed-off-by: Matthieu Berjon <matthieu.berjon@inria.fr>

Signed-off-by: Matthieu Berjon <matthieu.berjon@inria.fr>

Signed-off-by: Matthieu Berjon <matthieu.berjon@inria.fr>

Signed-off-by: Matthieu Berjon <matthieu.berjon@inria.fr>

I updated the confirmation page at registration to fit the stylesheet.

Signed-off-by: Matthieu Berjon <matthieu.berjon@inria.fr>

I updated the script to fix several issues such as ID keys that weren't
in auto increment mode.

Signed-off-by: Matthieu Berjon <matthieu.berjon@inria.fr>

I added a superuser for development purposes. The password should be
changed before going to production.

Signed-off-by: Matthieu Berjon <matthieu.berjon@inria.fr>

I updated the Oauth connection using OAuth2 and Gitlab. It's no possible
to register and authorize the connection with Gitlab. The identification
still doesn't work and need to be investigated.
I adapted as well the login and register forms.

Signed-off-by: Matthieu Berjon <matthieu.berjon@inria.fr>

I added a link in the footer related to the gitlab issue tracker for any
issue or request a user can have to declare.

Signed-off-by: Matthieu Berjon <matthieu.berjon@inria.fr>