Quiz API: Django → Docker → Qt

Three software architecture labs built around one application: a quiz REST API on Django REST Framework, the same API in Docker with four deployment options, and a C++/Qt desktop client using the Singleton and Adapter patterns.

Status: coursework ("Software Systems Architecture", Moscow Polytechnic University, group 241-327, spring 2026), completed

Stack: Python · Django · Django REST Framework · PostgreSQL · Docker Compose · nginx · Caddy · C++17 · Qt 6 (Widgets, Network) · CMake

Labs

	What was built	Folder
1	REST API: `Quiz` model, serializer, `ModelViewSet`, Faker test data generator, PostgreSQL	`lab-1/`
2	The API in containers: gunicorn behind nginx, static files from a shared volume, own images in the `git.deev.su` registry	`lab-2/`
3	Desktop client for the API: five HTTP methods, table and text views, Singleton and Adapter	`lab-3/`

Deployment options in the second lab:

Folder	Images	PostgreSQL
`local/`	built from source	in a container
`web_pg/`	prebuilt `git.deev.su/deevev/quizlab`	in a container
`web_lite/`	prebuilt `git.deev.su/deevev/quizlab`	external
`caddy/`	built from source, Caddy instead of nginx, HTTPS with a self-signed certificate	in a container

How it works

flowchart LR
    Q[Qt client<br/>ApiClient + QuizJsonAdapter] -->|HTTP JSON| N[nginx or Caddy]
    N -->|/api/, /admin/| G[gunicorn + Django REST Framework]
    N -->|/static/| S[(static volume)]
    G --> P[(PostgreSQL)]

API: GET /api/quiz/, GET /api/quiz/<id>/, POST /api/quiz/, PUT /api/quiz/<id>/, DELETE /api/quiz/<id>/. Quiz fields: title, description, author, time limit in minutes, published flag, creation date.

In the client, ApiClient is a Meyers singleton: one QNetworkAccessManager for the whole app. QuizJsonAdapter implements IQuizAdapter and turns a QJsonObject into a Quiz, so the window works with a domain class rather than JSON.

Running

The API in Docker (built from source, PostgreSQL in a container, 100 quizzes generated on start):

cd lab-2/local
cp .env.example .env      # set the DB password and DJANGO_SECRET_KEY
docker compose up -d      # API at http://localhost/api/quiz/

The other options run the same way from their folders. For caddy/, run sh gen-cert.sh first; the site opens at https://localhost.

A prebuilt client is in the releases: quizlab-client-windows-x64.zip (unpack and run quizlab-client.exe) and quizlab-client-linux-x86_64.AppImage (chmod +x and run).

Building the client from source (needs Qt 6 and CMake):

cd lab-3
cmake -B build && cmake --build build
./build/lab-3             # API defaults to http://localhost:80; another address: QUIZ_API_URL=http://host:port

The first lab without Docker: PostgreSQL, then pip install -r lab-1/requirements.txt, python manage.py migrate, python manage.py runserver. The connection is set with POSTGRES_* variables; Django settings with DJANGO_SECRET_KEY, DJANGO_DEBUG, DJANGO_ALLOWED_HOSTS.

Screenshots

DRF browsable API	Client text view

Development

Each lab is a self-contained folder, so the Django project is repeated in lab-1/, lab-2/local/backend/ and lab-2/caddy/backend/. The http.restbook files are API requests for the REST Book extension in VS Code.

License

Coursework ("Software Systems Architecture", Moscow Polytechnic University, 2026). The code is open for study; there is no separate license.

Author

Egor Deev — GitHub · Telegram · egor@deev.space

_{⭐ If you find this project useful, give it a star!}

_{Made with ❤️ — deev.space}

4.3 KiB Raw Permalink Blame History