Installation

The package requires version 3.2 or greater of Django, and a minimum Python version needed of 3.8.

Use pip to install the package, preferably to a local virtualenv:

pip install django_plotly_dash

Then, add django_plotly_dash to INSTALLED_APPS in the Django settings.py file:

INSTALLED_APPS = [
    ...
    'django_plotly_dash.apps.DjangoPlotlyDashConfig',
    ...
]

The project directory name django_plotly_dash can also be used on its own if preferred, but this will stop the use of readable application names in the Django admin interface.

Also, enable the use of frames within HTML documents by also adding to the settings.py file:

X_FRAME_OPTIONS = 'SAMEORIGIN'

Further, if the header and footer tags are in use then django_plotly_dash.middleware.BaseMiddleware should be added to MIDDLEWARE in the same file. This can be safely added now even if not used.

If assets are being served locally through the use of the global serve_locally or on a per-app basis, then django_plotly_dash.middleware.ExternalRedirectionMiddleware should be added, along with the whitenoise package whose middleware should also be added as per the instructions for that package. In addition, dpd_static_support should be added to the INSTALLED_APPS setting.

The application’s routes need to be registered within the routing structure by an appropriate include statement in a urls.py file:

urlpatterns = [
    ...
    path('django_plotly_dash/', include('django_plotly_dash.urls')),
]

The name within the URL is not important and can be changed.

For the final installation step, a migration is needed to update the database:

./manage.py migrate

It is important to ensure that any applications are registered using the DjangoDash class. This means that any python module containing the registration code has to be known to Django and loaded at the appropriate time.

Note

An easy way to register the Plotly app is to import it into views.py or urls.py as in the following example, which assumes the plotly_app module (plotly_app.py) is located in the same folder as views.py:

``from . import plotly_app``

Once your Plotly app is registered, plotly_app tag in the plotly_dash tag library can then be used to render it as a dash component. See Simple usage for a simple example.

Extra steps for live state

The live updating of application state uses the Django Channels project and a suitable message-passing backend. The included demonstration uses Redis:

pip install channels daphne redis django-redis channels-redis

A standard installation of the Redis package is required. Assuming the use of docker and the current production version:

docker pull redis:4
docker run -p 6379:6379 -d redis

The prepare_redis script in the root of the repository performs these steps.

This will launch a container running on the localhost. Following the channels documentation, as well as adding channels to the INSTALLED_APPS list, a CHANNEL_LAYERS entry in settings.py is also needed:

INSTALLED_APPS = [
    ...
    'django_plotly_dash.apps.DjangoPlotlyDashConfig',
    'channels',
    ...
    ]

CHANNEL_LAYERS = {
    'default': {
        'BACKEND': 'channels_redis.core.RedisChannelLayer',
        'CONFIG': {
            'hosts': [('127.0.0.1', 6379),],
        },
    },
}

The host and port entries in hosts should be adjusted to match the network location of the Redis instance.

Further configuration

Further configuration options can be specified through the optional PLOTLY_DASH settings variable. The available options are detailed in the configuration section.

This includes arranging for Dash assets to be served using the Django staticfiles functionality.

A checklist for using dash-bootstrap-components can be found in the bootstrap section.

Source code and demo

The source code repository contains a simple demo application.

To install and run it:

git clone https://github.com/GibbsConsulting/django-plotly-dash.git

cd django-plotly-dash

./make_env                # sets up a virtual environment
                          #   with direct use of the source
                          #   code for the package

./prepare_redis           # downloads a redis docker container
                          #   and launches it with default settings
                          #   *THIS STEP IS OPTIONAL*

./prepare_demo            # prepares and launches the demo
                          #   using the Django debug server
                          #   at http://localhost:8000

This will launch a simple Django application. A superuser account is also configured, with both username and password set to admin. If the prepare_redis step is skipped then the fourth demo page, exhibiting live updating, will not work.

More details on setting up a development environment, which is also sufficient for running the demo, can be found in the development section.

Note that the current demo, along with the codebase, is in a prerelease and very raw form. An overview can be found in the demonstration application section.`