Annotations, aggregates and prefetching

gistfile1.txt

Two problems:

1) Individual annotations can't be filtered
2) `prefetch_related` doesn't allow for filtered relations

The `annotate()` method is currently used for calculating
aggregates for each result in a queryset, but it could be
used more generally as an API for attaching the result of
some calculation to each model in a queryset - whether it
be a single value (an aggregate calculation) or multiple
values (a list of models).

Curent functionality:

Book.objects.annotate(num_authors=Count('authors'))  # Attaches a count of each Book's authors to the model as `num_authors`

Possible extensions:

Book.objects.annotate(num_active_authors=Count('authors').filter(active=True))  # Attaches count of active authors
Book.objects.annotate(active_authors=ModelList('authors').filter(active=True))  # Attaches a list of active author instances

Is this possible?