stack trace on deleting index not in database

[codekiln]

In 0.8.0, if there is an index only-in-es created in elasticsearch that doesn't exist in the database, it can be listed with ./manage.py es_list --es-only. In theory, ./manage.py es_drop only-in-es --exact should delete it, however, this currently errors out:

$ ./manage.py es_drop test_localdev_movies-1 --exact
[2018-11-18 01:03:48.353449]: While completing drop_index, encountered exception: 
 - message: 'NoneType' object has no attribute 'is_active' 
 - exception doc: Attribute not found. 
 - exception stack: Traceback (most recent call last):
  File "/Users/pnore/hbx/django-elastic-migrations/django_elastic_migrations/models.py", line 312, in start_action
    result = self.perform_action(dem_index, *args, **kwargs)
  File "/Users/pnore/hbx/django-elastic-migrations/django_elastic_migrations/models.py", line 726, in perform_action
    if self.index_version.is_active and not self.force:
AttributeError: 'NoneType' object has no attribute 'is_active'

Looking deeper, it seems as though the dem_index that is received in the drop index action doesn't have access to the full, exact name provided:

> dem_index.get_es_index_name()
None

This is a bug. A test should be written for this scenario and it should be fixed.