Yet another Racket autograder for Gradescope.
Forthcoming.
-
Create a Racket test/grade file. (See the example below for guidance.) It need not be in this directory.
-
Create any additional files you might need. Once again, they need not be in this directory.
-
Run
make-zip GRADE_FILE.rkt [other-files]
-
You should now have a zip file named
autograder-NAME.zip
. -
Go to the page for the assignment on Gradescope or create a new programming assignment (see below for details on creating a programming assignment).
-
The second checkbox should be "Configure Autograder". Select that.
-
Upload your autograder. Then click "Update Autograder" (or "Replace Autograder"). It should take a few minutes. Ideally, you'll see a bunch of stuff under "Build Output" and nothing under "Build Errors".
The "post-installing collections" step in Racket seems to take the longest.
At some point, it should say "Successfully built ....".
-
Click "Test Autograder".
-
Upload a Racket file and any other files the student might upload.
-
Wait for the Autograder to finish.
-
If things don't go as expected, you may have to debug via SSH. The documentation is at https://gradescope-autograders.readthedocs.io/en/latest/ssh/.
See examples/grade-example01.rkt
.
Assignment creation in Gradescope is supposed to be obvious. These notes are just to remind us of some common steps.
-
Click on "Assignments" on the left.
-
Click on "Create Assignment" (or "Duplicate Assignment" on the lower right.
-
Click on "Programming Assignment"
-
Click "Next" (duh)
-
Assign one (1) Autograder point.
-
Enable Manual Grading.
-
Fill in the remainder as appropriate. DO NOT enable a leaderboard.
-
We usually end up on a the Outline screen. Add a new question for the human grader. (I usually call it "Human Grader".) It's worth three points for MPs and 0 points for labs. Unfortunately, we can't add the rubric here.
-
Click on "Configure Autograder" and follow the steps above.
-
When the sample grading finishes, you may want to update the rubric under the "Manage Submissions" bubble. We usually use.
- 1 Some Issues/Needs Revision (R)
- 2 Meets Expectations (M)
- 3 Excellent (E)
- 4 Incomplete, incorrect, or no submissions (I)