Refactor and unify annotation functions

[richardebeling]

Currently we have these functions (feel free to add more if I missed any):

1. Evaluation:annotate_with_participant_and_voter_counts: takes QuerySet, uses a subquery and calls .annotate, returns modified QuerySet
2. annotate_distributions_and_grades: takes iterable of Evaluations, modifies in place
3. annotate_evaluations_with_grade_document_counts: takes QuerySet, calls .annotate and returns modified QuerySet
4. get_evaluations_with_course_result_attributes: takes iterable of Evaluations, modifies and returns the same iterable
5. results.get_evaluations_with_prefetched_data: takes iterable or QuerySet. Wrapper for 2 that will prefetch data if given a QuerySet.
6. staff.get_evaluations_with_prefetched_data: Takes a semester, returns QuerySet annotated with a bunch of stuff, calls 1 and 3.
7. get_evaluations_of_course: Returns list, calls 4.

Since I'm already at it, I'll also add here that Semester has two class methods:

• get_all_with_published_unarchived_results
• get_all_with_unarchived_results

I think these functions should be refactored to use a uniform naming scheme. Maybe there are some possibilities for deduplication? Maybe we could always use QuerySets (perhaps not easily possible)?