zconnect-django-demo

zconnect-django-demo this is an example django application which uses zconnect-django, to see this as a part of a whole zconnect system please see https://github.com/zconnect-iot/demo-virtual-docker-compose

Starting from scratch

1. pip3 install --user pipenv (if you don't already have it)
2. pipenv install --dev
3. pipenv shell
4. ./start.sh

When done, run deactivate to close of the virtual env, or just close the terminal.

When coming back to do development, run pipenv shell

Seed data

There is some seed data in django_demo/seed.py which is used if the environment variable DJANGO_SEED_PROJECT is set when the server is started. This just has some basic devices and timeseries data for now, it's just there as an example and if you need anything extra then add it there.

Optionally you can set the environment variable DJANGO_SEED_PROJECT_ONCE as well to stop the server from re-seeding the database every time the code changes.

This data is also used in the fix_Demo_ts_data fixture, for example in the dashboard endpoint tests.

By default it will only seed the database 'once', which means if it finds the admin user already there it won't seed anything else. If you need to add more seed data, do docker volume prune to delete all the old volumes and then run the instructions above again

Running celerybeat

To run the scheduled celery tasks you need to run the worker and the beat with the following commands:

env DJANGO_SETTINGS_MODULE=django_demo.settings.development celery -A django_demo beat -l info
env DJANGO_SETTINGS_MODULE=django_demo.settings.development celery -A django_demo worker -l info --scheduler django_celery_beat.schedulers:DatabaseScheduler

However, the development settings set CELERY_TASK_ALWAYS_EAGER to True which means that when running these commands locally, the worker will never recieve the tasks. Instead, the beat will execute the tasks synchronously.

Documentation

1. Go into virtualenv with all dev requirements installed
2. Go into apps/zconnect-django/doc/source
3. Run make html

The documentation will be output in the build/html folder.

Tests

Running unit tests

(maybe pip install -r requirements.txt)

1. pipenv shell
2. py.test -c pytest.ini -k .py (-k deselects the tavern tests)

NB: The fixture fake_get_redis has autouse set to True, so any uses of get_redis in zconnect.tasks will return a Mocked version of redis, this is to avoid the need to run a redis-server during unit tests.

Linting

To recreate the linting that happens in GitLab CI, run ./lint.sh. Aim to do this before every push, so that you don't push, wait for tests and realise that you missed something trivial and have to repeat the process!

You can also run ./pylint.sh for just pylint.