Skip to content

README.md

Latest commit

 

History

History
92 lines (79 loc) · 5.76 KB

README.md

File metadata and controls

92 lines (79 loc) · 5.76 KB
Raw

example-course - a worked dummy course for demos

A complete, ready-to-deploy dummy course that exercises the whole engine end to end: materials, two assignments, a roster, instructor/TA cards, and an auto-generated cohort website with a real schedule. Use it two ways:

  • Artifact A - the live reference demo. Deploy this dataset to the demo org pair below, then send faculty & instructors the links (the site + the Actions tab). They click a finished thing - no setup on their end. This is also the script for a live walkthrough.
  • Artifact B - self-serve. A faculty & instructors member follows the same steps to stand up their own course and feel the workflow. Same dataset, same runbook.

The canonical, engine-wide input reference is docs/required-input-schema.md. This file is the demo-specific concretisation of it.

The demo orgs

Tier Org Role
Course Hertie-DSL-Demo persistent control room - materials, assignment templates, the console
Cohort DSL-Demo-f2026 student-facing target - welcome, roster, released materials, the site

What's in this dataset

example-course/
  course-org/
    dsl-course.yml                  # course identity + course_admins (the SSOT) + display-only instructor/TA cards
    course-materials-f2026/
      lectures/01_week-1../05_week-5/  # 5 sessions (ordinal-prefixed dirs; slides.md + a code demo each)
      readings/01_week-1../05_week-5/  # 5 sessions of placeholder readings
      syllabus.md
    lecture-code-f2026/mlpkg/       # a growing package, disclosed module-by-module (release_code)
    assignment-1-f2026/             # individual (.py)
      main/                         #   -> main branch: README brief + starter.py
      solution/                     #   -> solution branch: solution/ + grading.yml + hidden tests/
    assignment-2-f2026/             # individual (notebook): main/ + solution/
    assignment-4-project-f2026/     # GROUP project: main/ + solution/ (grading.yml type: group)
  cohort-org/
    students.csv                    # 4 dummy students (handles blank - filled on onboard)
    teams.csv                       # team membership for the group project
    schedule.yml                    # the auto-release plan (materials_releases) + due dates + exams
    people.yml                      # this cohort's own instructors/TAs (real push access)
    grades/*.csv                    # per-assignment faculty & instructors grade tables (auto/manual/final)

Assignment layout: each assignment-*/ splits into main/ (-> the repo's main branch, what students get) and solution/ (-> the solution branch: the model solution, grading.yml, and the HIDDEN tests/ the faculty-side Grade assignment runs). Student repos never get solution/.

Deploy it (≈20 min)

Prereqs: the bot account is an owner of both demo orgs, and DSL_BOT_TOKEN (repo + admin:org + workflow) is available. See the token section.

  1. Create Hertie-DSL-Demo and DSL-Demo-f2026 in the GitHub web UI; add the bot as owner of each. (The only manual step - there is no org-creation API.)
  2. This repo → Actions → Bootstrap Course Org: org=Hertie-DSL-Demo, org_name=DSL Demo Course, course_name=Deep Learning (Demo), course_code=GRAD-DEMO, set_secret=true.
  3. Copy this dataset's course-org/dsl-course.yml into Hertie-DSL-Demo/.github/dsl-course.yml (web editor is fine). It declares course_admins (real, course-wide access) plus display-only instructor/TA cards (headshots + bio links) - so the site shows intended cards, not GitHub avatars. Real instructor/TA push access comes from the cohort's own people.yml instead (step 8).
  4. New materials repo (tag=f2026), then push the contents of course-org/course-materials-f2026/ into it (lectures/readings/syllabus).
  5. New assignment twice (number=1 then 2, tag=f2026), then push each course-org/assignment-N-f2026/ (README brief + starter.py) into the matching template.
  6. Refresh actions (populates dropdowns + propagates the repo secret).
  7. Bootstrap cohort: cohort_org=DSL-Demo-f2026.
  8. Replace the starter row in DSL-Demo-f2026/classroom-config/students.csv with cohort-org/students.csv, and copy this dataset's cohort-org/schedule.yml into DSL-Demo-f2026/classroom-config/schedule.yml - the release calendar (Scheduled release cron) and the schedule (real due/exam dates), so the site shows intended dates, not synthesised ones. Also copy cohort-org/people.yml into DSL-Demo-f2026/classroom-config/people.yml - this cohort's real instructor/TA push access (a instructors-f2026 team on the course org, scoped to this year's repos).
  9. Release materials for sessions 1-5 (the 01_week-1 ... 05_week-5 directories). Release assignment for assignment-1.
  10. Sync site (releases also trigger it).

What this stands up

  • The site: https://dsl-demo-f2026.github.io - course name, semester, instructor/TA cards, sessions 1-3 lectures linking the released files, two assignment briefs, and a schedule with the real dates from step 8 (Assignment 1 due 13 Oct, MidTerm 3 Nov, Final 15 Dec).
  • The console: Hertie-DSL-Demo/.github Actions tab - every faculty & instructors button.
  • Onboarding: open a Join issue in DSL-Demo-f2026/welcome, type a student ID from the roster (e.g. 220001), and watch the onboard action enrol you. (Only IDs whose row you can claim with a real GitHub account run fully end-to-end - the dummy rows have blank handles until someone joins.)