The main objective of webtask_scheduler project is to be a service that allows us to execute scheduled tasks. For simplicity, let’s assume that tasks can be triggered by accessing a web URL. You create a task by providing a URL and the desired time to run. When the specified time arrives, the service will call the URL.

I've reused the PeriodicTask Model to make use of it's features to create a model instance that will store task info. If I was about to build my Own model for our case in the project context, I would cloned some of it's fields (name, task, clocked). I found it fit our need in this project But in a bigger context we might need to add UUID field.

The usage of User app is to enable login and creating a superuser account with email as an identifier instead of username. So that reviewer can view the tasks list and filter though them.

I've Created the compose directory with local sub-directory, and inside it contains the start shell scripts for used services (flower, beat, and worker) in addition to Dockerfile The reason for this decision was to make it easy to deploy the app on different environment. All is needed is to create another directory with same directory structure and use it in the related compose file.

/timer Receives a JSON object containing hours, minutes, seconds, and a web url.

Return task ID and time left in seconds.

The good bractice for such a service is to create a model that inherits from PeriodicTask and Add replace the pk with UUID. This was skiped to avoid extra complexity and make project easy for review.

/timer/{timer_uuid} Receives the timer id in the URL, as the resource uuid.

Returns a JSON object with the amount of seconds left until the timer expires. If the timer already expired, returns 0.

poetry

I had a great opportunity to learn more about poetry with this task. I added poetry configurations in pyproject.toml and in the Dockerfile. However, since it was my first time using it, I encountered a couple of errors. Debugging these errors would have taken more time than I wanted to spend on the task, so I chose the safer option of using requirements.txt.

1. Go to webtask_scheduler directory

cd webtask_scheduler

2. Build

docker compose build

3. Run

docker compose up

4. Access Swagger Docs using the below url

0.0.0.0:8000/api/docs

For the ease of use the Authentication and Authorization was skiped. Assuming it's a public APIs to create/retrive tasks. But it depends on the app usage Authentication and Authorization can be added accordingly.

Use swagger docs url 0.0.0.0:8000/api/docs/#/scheduler/v1_scheduler_timer_create to test setting timer by providing hours, minutes, seconds and web_url values.

Validation Fields (hours, minutes, seconds) must be integer and greater than or equal zero.

Use swagger docs url 0.0.0.0:8000/api/docs/#/scheduler/v1_scheduler_timer_retrieve to test getting a timer with provided task ID or return 404 status code if ID was not found.

This step requires a superuser account.

PS: default login/register identifier is email instead of username

1. Create admin user

docker-compose exec django /entrypoint python manage.py createsuperuser

2. Enter email and password

3. Open 0.0.0.0:8000/admin/django_celery_beat/periodictask

Running the all project tests at once using pytest Use this command

docker-compose exec django /entrypoint pytest

In case of running single function Use this command

docker-compose exec django /entrypoint pytest -s -vv -k <test_function>

To avoid adding more complexity to the project Some of the production configurations was skiped. However in order to deploy and run this application into production environment you should:

• Sentry Usingsentry-sdk Add SENTRY_DSN to .envs/.production/.django (Use gitcrypt to encrypt production envs)

• Add production services with production envs
• Add Dockerfile for Django and Postgres
• Add start and entrypoint bash script in each one according to the project need
• Finally build the project then Push image to ECR or run it using compose up command.

Finally Add to templates directory the default error page 400, 403, 500 html files. Then Update config/urls.py with Error path and template.