Added support for aggregated fields in CSV output.

[nshafer]

If you have an aggregated field in your queryset, such as with:

Publisher.objects.annotate(num_books=Count('book'))

Trying to include this column 'num_books' in your CSV output results in a ValueError Exception with the error dict contains fields not in fieldnames: num_books. This fixes that by looking for column names for aggregates in the queryset.query, just like for extra columns.

[coveralls]

Coverage decreased (-0.94%) when pulling 4eee1c1 on Tekco:master into 3ac10ab on azavea:master.

[steventlamb]

Thanks for your contribution, this looks like a nice thing to add.

I've made another pull request on top of this that adds a unit test.