Simple TA Helper

Please study the slides for more details. Link: https://goo.gl/FEm6gR

• step 0: Download the classroom_roster.csv from Github Classroom.

  • example https://classroom.github.com/classrooms/61244606-ntu-machine-learning-spring-2020/roster#download-csv-modal

• step 1: Create student_list_csv.

  • usage:

  $ bash make_student_list.sh <roster_csv> <student_list_csv>
  

  • example:

  $ bash make_student_list.sh ../classroom_roster.csv ../student_list.csv
  

• step 2: Get Kaggle leaderboard scores. (Optional.) One can get the scores directly from the kaggle leaderboard or get the scores from all submissions.

  • usage

  $ # directly get scores from leaderboard
  $ python get_kaggle_score.py -c <competition> -l <student_list_csv> -o <output_csv>
  

  or

  $ # get scores from all team-submissions
  $ python get_kaggle_score_submission.py -c <competition> -l <student_list_csv> -o <output_csv> -p <cpus> -b <browser>
  

  • example:

  $ python get_kaggle_score.py -c ml2020spring-hw1 -l ../student_list.csv -o /tmp/kaggle_hw1_score.csv
  

  or

  $ python get_kaggle_score_submission.py -c ml2020spring-hw1 -l ../student_list.csv -o /tmp/kaggle_hw1_score.csv -p 8
  

  • Note: You should login as the host of the competition to run the second code. If it fails using chrome cookies, run chrome with --password-store=basic.

• step 3: Setup a environment for a specific homework.

  • edit config
  • run setup.sh

  bash setup.sh
  

• step 4: Run main.sh (FIXME)

  • requirements:
    • parallel
    • dos2unix
  • usage:

  $ bash main.sh <student_list_csv> <git_pull_or_not>
  

  • example:

  $ bash main.sh ../student_list.csv true