canvas-hw: tools for offline homework grading in Canvas LMS

Bart Massey 2023

These are janky half-finished tools for downloading an assignment from Canvas that is presumed to be one ZIP file per submission, unpacking that locally, then allowing local grading. At the end, the grade results are uploaded to Canvas.

Thus is SpeedGrader avoided.

Instructions

One-Time Setup

• https://canvas.pdx.edu/profile/settings to generate Canvas API token
• Store Canvas API token in ~/.canvastoken
• Clone canvasgrader fork http://github.com/BartMassey-upstream/canvasgrader (for now) and run pip install .

Download

• Go to https://canvas.pdx.edu//students and copy-paste the JSON found there into students.json (gross, will fix later)

• Download a zipball of the HW from Canvas.

• Run unpack-hw.py --hw <assignment name> <zipball>

Grading

• For each directory in staged/:

  • Edit GRADING.txt. Make sure to replace the - with a score.
  • Move directory from staged/ to graded/

Upload

• Find course ID in course URL
• Find assignment ID in assignment URL
• Run upload-grades.py <courseid> <asgid>

Acknowledgments

Thanks to the authors of the Python canvasgrader package for making this all possible.

License

This work is licensed under the "MIT License". Please see the file LICENSE.txt in this distribution for license terms.