This is an grading library for coursera that I wrote as a TA for Stanford's CS124: From Languages to Information Class. Essentially, this covers the intermediate nonsense required to communicate to coursera servers, leaving the administrator to write the actual grading logic. Naturally, the audience for this may be limited :)

An example grader that reports a fixed score is available in example_grader.py.

Your grading logic is registered as callback functions that accept the submission data object as a parameter. So, below, we are giving a score of 3 for part 1,2 and a score of 0 for part 3. The administrator would put their grading logic in place.

The simplest grader I could think of:

 from coursera_queue import QueueProcessor, SubmissionGrade
 QueueProcessor("queue_name", 
    "http://coursera-course.coursera.org/assignment/api/pending_submisison", 
    "http://coursera-course.coursera.org/assignment/api/score",
    "api_key", {
    ("1", "2"): (lambda s: SubmissionGrade(3.0, feedback="Great!")),
    "3": (lambda s: SubmissionGrade(0.0, feedback="Too bad!"))
 }).pollForever()