Env details

• Python 3.9.4

Set up project

• python -m venv myvenv
• source ./myvenv/bin/activate (On *nix)
  • .\myvenv\Scripts\activate.bat (On Windows)
• pip install -r requirements.txt

Initialize values

• Download chromedriver and set its path in constants.py
  • You may be required to upgrade the driver to a newer version from time to time (there'll be errors indicating that)
• Create a cache directory for storing preprocessing results and grades. Set its path in constant.py
• Copy handles.csv from the shared drive location and place it in the cache directory.
  • This csv file contains each student's names, unis, and each of their coding platform/website handles.
• Create a logs directory in the project root if you wish the logger to spool to a log file. Either way the grader will still print stuff to the console.

Running the grader

• Remember to activate the virtual env (see setup)
• Run the preprocessor first: python3 preprocessor.py -w <week_num>
  • This can take up to ~30min-1hr.
  • If someone's already done this, they can share the preprocessing cache files to avoid waiting.
• Run the grader next: python3 grader.py -w <week_num>
  • It will store grading events in /path/to/cache/dir/grading_events_<week_num>.log.
• Run the calculator/assimilator next: python3 calculate_points.py -w <week_num>
  • It will store the grades/points in /path/to/cache/dir/grades_<week_num>.csv
• After all grades in grades_<week_num>.csv are finalized i.e. final_points column's filled (perhaps after a manual comparison with code/screenshots submitted by students), download/export the gradebook from courseworks and then run the populate gradebook script: python3 populate_gradebook.py -w <week_num> -c /path/to/courseworks/exported/gradebook.csv -g /path/to/finalized/grades_<week_num>.csv -o /path/to/output/coursworks/import/gradebook.csv
  • The -o option is going to be the new file that the script will create which can be directly uploaded/imported on courseworks to update assignment scores.
• NOTE:
  • The grader can be run for a single person and/or a single platform using python3 grader.py -w <week_num> -u <uni> -p <platform_name_used_in_code>. This is very useful to crosscheck certain scores (if code/screenshots differ from what the grader calculated)

Current Shortcomings

• Leetcode Practice problem points cannot be calculated by auto-grader because Leetcode does not show all submissions made by a user.
• Topcoder Contest and practice problems points are not auto-calculated yet. It was too complicated too code and I thought we should only put efforts if we see enough student submissions for Topcoder.
• Kattis Practice problem points cannot be calculated by auto-grader because Kattis does not show submissions made by a user at all.