Dash module¶

This is the main entry point for the frontend web interface. You should not interact that often with this technical dash.py, but suffices to say that you find there

The logic to initially create the tables inside the postgresql instance
The logic to connect to the app.
The logic to create initial user(s). Go re-read the ecodev-app intro documentation introduction to see one way to create the initial user(s).
The logic to define the app colors. Edit APP_COLORS as you see fit!
The logic to be in debug mode or nor for the dash app. Controlled by the DEBUG flag setup via env variables. Go re-read the ecodev-app intro
documentation introduction if needs be.
The logic to use either uvicorn or gunicorn. Controlled by the DASH_APP_SETTINGS.gunicorn_setup flag setup via env variables.

Warning

We found scarcely few documentation on how to use gunicorn with dash. For this to work implied for us

Using from flask import Flask and then create the Dash app with the Flask(__name__) instance

Instanciate the Dash app like so

if __name__ == '__main__':
    dash_app.run_server(host='0.0.0.0', port=80)

in the docker-compose.yml file, call
```
command: gunicorn app.dash_app:server --workers 4 --worker-class gevent  --bind 0.0.0.0:80
```
gevent is the best thing we found to work with workers for Dash. You can obviously change the number of workers as you see fit.

Warning

With gunicorn long (>30s dash callbacks with timeout. You should handle long callbacks in another fashion (which presumably will improve the UX of your application anyway 😋). You can call your backend and use a Fastapi

BackgroundTasks there, go nuts with Celery... Sky is the limit 😋.