ljunghanscode / example-dash-structure

What a dash app could look like

dash-practice

Project structure practice

Installation

After cloning/downloading the repository, simply install dash-practice as a package into your target virtual environment:

$ pip install PATH_TO_dash_practice

During development you will likely want to perform an editable install so that changes to the source code take immediate effect on the installed package.

$ pip install -e PATH_TO_dash_practice

Running Your App

This project comes with two convenience scripts for running your project in development and production environments, or you can use your own WSGI server to run the app.

Run Dev App

Installing this package into your virtualenv will result into the development executable being installed into your path when the virtualenv is activated. This command invokes your Dash app's run_server method, which in turn uses the Flask development server to run your app. The command is invoked as follows, with proj_slug being replaced by the value provided for this cookiecutter parameter.

$ run-project_slug-dev

The script takes a couple of arguments optional parameters, which you can discover with the --help flag. You may need to set the port using the --port parameter. If you need to expose your app outside your local machine, you will want to set --host 0.0.0.0.

Run Prod App

While convenient, the development webserver should not be used in production. Installing this package will also result in a production executable being installed in your virtualenv. This is a wrapper around the mod_wsgi-express command, which streamlines use of the mod_wsgi Apache module to run your your app. In addition to installing the mod_wsgi Python package, you will need to have installed Apache. See installation instructions in the mod_wsgi documentation. This script also takes a range of command line arguments, which can be discovered with the --help flag.

$ run-project_slug-prod

This script will also apply settings found in the module dash_practice.prod_settings (or a custom Python file supplied with the --settings flag) and which takes precedence over the same settings found in dash_practice.settings.

A notable advantage of using mod_wsgi over other WSGI servers is that we do not need to configure and run a web server separate to the WSGI server. When using other WSGI servers (such as Gunicorn or uWSGI), you do not want to expose them directly to web requests from the outside world for two reasons: 1) incoming requests will not be buffered, exposing you to potential denial of service attacks, and 2) you will be serving your static assets via Dash's Flask instance, which is slow. The production script uses mod_wsgi-express to spin up an Apache process (separate to any process already running and listening on port 80) that will buffer requests, passing them off to the worker processes running your app, and will also set up the Apache instance to serve your static assets much faster than would be the case through the Python worker processes.

Note: You will need to reinstall this package in order for changes to the run-dash_practice-prod script to take effect even if you installed its an editable install with (ie pip install -e).

Running with a different WSGI Server

You can easily run your app using a WSGI server of your choice (such as Gunicorn for example) with the dash_practice.wsgi entry point (defined in wsgi.py) like so:

$ gunicorn dash_practice.wsgi

Note: if you want to enable Dash's debug mode while running with a WSGI server, you'll need to export the DASH_DEBUG environment variable to true. See the Dev Tools section of the Dash Docs for more details.