UlasKelas Backend

Final Project's Backend of Mobile Development SIG 2021 using Django Rest Framework.

Database Schmea

Database schema to as a guidance to data that will be used in this project

Getting Started

Move to backend directory

cd backend

Create virtual environment

python3 -m venv env

# Activate virtual environment
source env/bin/activate

# How to Deactivate
deactivate

Run app dev

1. duplicate ./UlasKelas/.env.sample and rename to .env

2. run postgres db

docker-compose up -d

3. activate env and install requirement

source env/bin/activate
pip install -r requirements.txt

4. run project

python manage.py runserver

Run app prod

docker-compose -f docker-compose-prod.yml up -d

if you make code changes, run this command

docker-compose -f docker-compose-prod.yml down && sudo docker-compose -f docker-compose-prod.yml build && docker-compose -f docker-compose-prod.yml up -d

Create or update database dev

before migrate db, make sure ulas-pg container running

1. change or add the related models

2. make migrations file && migrate database

python manage.py makemigrations
python manage.py migrate

Create or update database prod

1. change or add the related models

2. make migrations file make sure change pg host to localhost in settings.py before makemigrations

python manage.py makemigrations

change host to postgres again

3. migrate database before migrate db, make sure ulas-pg container running

docker exec -it ulas-server python manage.py migrate

Access db

docker exec -it ulas-pg bash
psql -U postgres

or use database management and input credentials provided in settings.py

Now you can login with superuser you just create on https://localhost:8000 and interact with API view OR call the API endpoint with cURL or Postman.

Push notification reminders

Register Firebase device tokens through POST /api/device-tokens. Production must set FIREBASE_ENABLED=true and provide either FIREBASE_CREDENTIALS_BASE64 or Application Default Credentials. Firebase's APNs integration must be configured for iOS delivery.

Configure the managed scheduler in the Asia/Jakarta timezone:

0 16 * * 5 python manage.py send_notification_reminders --type calculator
5 0 * * * python manage.py send_notification_reminders --type review

The review command safely skips every day except the calendar month's final day. Both commands use database deduplication and can be retried.

Sunjad Endpoint Used

Sunjad all courses mock servers https://3e081de5-8b4c-46ea-8736-99476c47204b.mock.pstmn.io/courses

Synchronize the cross-faculty course catalog

After applying migrations, populate all supported S1, D3, and D4 programs:

python manage.py migrate
python manage.py sync_courses --all

To refresh only one SSO/SunJad organization code:

python manage.py sync_courses --org-code 01.00.12.01

Configure the production managed scheduler to run the all-program command daily at 02:00 Asia/Jakarta (0 2 * * * when the scheduler supports an explicit timezone, or 0 19 * * * in UTC):

python manage.py sync_courses --all

Run the same command once immediately after the first deployment so the catalog is populated without waiting for the next schedule. The scheduler must retain stdout/stderr and alert on a non-zero exit status. A failed program is reported at the end without preventing the remaining programs from being synchronized. Programs that have no SunJad catalog yet are reported as unavailable without changing their existing course statuses.

SunJad requests use a 5-second connection timeout, a 20-second read timeout, and up to three attempts with exponential backoff for transient failures. Course mappings are deactivated only after a valid catalog snapshot; failed or untrusted responses never cause mass deactivation. Inactive courses stay in the database so existing reviews and calculator history remain accessible.

Before the first sync after deploying course-code validation, audit invalid legacy courses with the dry-run cleanup command:

python manage.py cleanup_invalid_course_codes

After reviewing the output and taking a database backup, apply the cleanup:

python manage.py cleanup_invalid_course_codes --apply

The cleanup deactivates invalid catalog mappings. Invalid courses are deleted only when they have no review, bookmark, calculator, course-semester, or TanyaTeman references; referenced courses are retained for manual remediation.

Import the latest SLCM IRS period locally

This proof of concept runs only as a local management command. Install the temporary browser used by Playwright:

pip install -r requirements-slcm.txt
python -m playwright install chromium

Run a preview for a local Teman Kuliah profile and calculator semester:

python manage.py import_slcm_irs \
    --username example.username \
    --semester 1 \
    --irs-url "https://slcm.ui.ac.id/path-to-irs" \
    --dry-run

The command opens an isolated browser at the supplied SLCM IRS page. Complete the SLCM login; no terminal confirmation is needed. After authentication, the command previews only the latest academic period that contains courses. Remove --dry-run to confirm the database import in the same browser session. Existing calculator courses are skipped, and SLCM codes missing from the local catalog are reported. The command does not save SLCM credentials, cookies, page HTML, or browser storage. Use --login-timeout to override the default five-minute login window.

SLCM autofill from the frontend

Configure SLCM_IRS_URL with the fixed IRS page and expose the remote browser using SLCM_BROWSER_PUBLIC_URL. The production Compose file includes the single-session Chromium/noVNC service used by the login popup.

Create an authenticated session with POST /api/slcm-autofill/sessions. The given_semester JSON field is optional: when omitted, the backend derives the student's current semester from their NPM entry year and the current UI academic period. Send a value such as {"given_semester":"1"} to override it manually. Open the returned popup_url, poll GET /api/slcm-autofill/sessions/{session_id}, display its preview once the status is ready, and finish with POST /api/slcm-autofill/sessions/{session_id}/confirm. Cancel an unfinished session with DELETE /api/slcm-autofill/sessions/{session_id}.

The remote browser defaults to a touch-enabled 430x932 kiosk viewport, so Chromium's tabs and address bar do not consume the mobile login area. noVNC scales that fixed portrait framebuffer to the available frontend view, including when the on-screen keyboard changes its height. Override the framebuffer with SLCM_BROWSER_SCREEN_WIDTH and SLCM_BROWSER_SCREEN_HEIGHT when needed. The frontend should keep polling after opening the popup and close its popup or browser view when the session becomes ready, failed, expired, or cancelled; the course preview remains in the regular application UI.

Changes to the Selenium screen environment require recreating the service; docker compose restart keeps the old container environment. Apply them with:

docker compose -f docker-compose-prod.yml up -d \
  --force-recreate slcm-browser server

---

Documentation from 2020

Authentication

Send POST request to /login/. This endpoint will redirect user to SSO Login page. If login success, login page will be directed to /token/ endpoint where you can retrieve token and username in its URL parameters.

Example:

{BACKEND_ROOT_URL}/token?token=f039021efcvery-long-tokencbe90717daded39&username=dummy.dumdum

This token will be used in authorization header as token authorization. Set Token {very-long-token} as value of Authorization on the headers of every endpoint that need authorization.

Example:

curl {BACKEND_ROOT_URL}/any-restricted-endpoint/
    -H "Accept: application/json"
    -H "Authorization: Token f039021efcvery-long-tokencbe90717daded39"