GitLab

Do not update holding limit if there exist user-specific base quota
for the particular resource.

The following indicators are returned for each holding:
"limit": the quota limit (as before)
"usage": the current usage (pessimistically, includes any pending
         allocations)
"pending": the difference of max and min allocation, indicates
           that there are pending commissions on this holding

also include services for which no existing presentation data exist.
Admin can use ASTAKOS_SERVICE_META setting to customize the service
presentation metadata.

Update view to use the introduced astakos quotas api.

Using this parameter, the service can query the quotas of a specific
user.

Rename it to fields.py and update all migrations that reference the field.

Remove quantity, import_limit, export_limit.
Update user-set-initial-quota management command.

Several fields of models Project and ProjectMembership were related
to the external syncing to quotaholder and have become obsolete since
the inclusion of quotaholder in astakos.

Returns quotas for all resources associated with the service that makes
the call.

- Remove presentation information from Service model (order, icon)
- To access the list of services filled with both presentation and api
  data use Service.catalog(). The method fills a dict with service
  identifiers and service metadata through the following steps
    - Initialize a service dict using
      astakos.im.presentation.service_defaults
    - Extend (merge and overwrite) it with the correspoding entry in
      astakos.im.presentation.SERVICES if exists.
    - Merge once again if user has set service metadata in
      ASTAKOS_SERVICE_META setting. This way user can change only the
      desired service parameters.

- An example of ASTAKOS_SERVICE_META containing keys which

    ASTAKOS_SERVICE_META = {
        'myservice': {
            'name': 'service_identifier',
            'url': 'https://service.url/service/ui/',
            'verbose_name': 'My service name',
            'order': 1,

            # cloudbar specific parameters
            'cloudbar': {
                'show': True,
                'title': 'My service name in cloudbar'
                # defaults to verbose_name
            },

            # dashboard specific parameters
            'dashboard': {
                'show': True,
                # we use different ordering in dashboard
                'order': 100,
                'description': 'My service description'
            }
        }
    }

- Updated dashboard to use Service.catalog() instead of hardcoded
  html.
- Updated service-* commands to keep up with model changes

- Remove foreign key dependency between Service/Resource model. Replace
  it with a service CharField to keep track of the service identifier
  that the resource was registered from.
- Remove all references to resource service name in resource lookups

When changing the default base quota (uplimit) for a preexisting resource,
do a bulk update of the related holdings.
When importing a new resource, create the new related holdings.

Prompt the admin for the uplimit in the management command, if a config
file is not given.

Register a service's resources giving option --json <file>.
The json file is expected to represent a dictionary with keys
"service" and "resources".

Currently the default base quota (uplimit) for each resource is read
from a configuration json file, given with --conf <file>. This file
represents a dictionary with numeric values for each resource name.

Upon importing the resources, the quotaholder has to be updated due to
the modified resource limits.

Also remove unused ResourceMetadata model.

UI-related presentation data go to presentation.py.
Remove load_service_resources call and related astakos-init
management command.

move astakos.im.synnefo_settings to astakos.synnefo_settings and make
the appropriate changes in setup.py

An unevaluated query ends up being used as an argument of intended list
type to another query, which fails in the case of an EmptyQuerySet, which
is not iterable.

landing instead of user profile

It is not necessary, since the service token in used for authorization.
The api_method decorator is modified so that it passes to the method
the service data through the request object.