The backend of our shared space reservation platform.

First, install pip. I don't think it matters whether pip2 or pip3 since we just need it to pull down virtualenv, then all other dependencies will live inside that virtual environment.

Install with

pip install virtualenv

I'm currently using virtualenv to make development consistent and manage dependencies. While not strictly required to develop, it's probably a Good Idea.

First, you'll need a python3 installation. On macOS this can be done via

brew install python3

(or various other ways). Next, in the root directory of the git repo, do

virtualenv --python=python3 env

to get an environment set up. Whenver you want to use the virtual environment (to develop), just do

source env/bin/activate

Inside the virtual environment, just run

pip install -r requirements.txt

If a new dependency is needed, it can be installed with pip and then pip freeze will update the dependencies (requirements.txt).

Before starting a local instance, environment variables must be set, and the SQL proxy must be started. (See Deployment section for additional details.)

source env_vars.sh
./start_sql_proxy

Like any django app, the server can be run with

python manage.py runserver

and a shell can be started to mess around with via

python manage.py shell

See the django docs for all the available options.

Production deployment of this application is configured for use with Google Cloud Platform. The Apollo Application is served by a App Engine Python 3 Flexible Runtime, static content is served from Cloud Storage, and structured data is managed by a PostgresSQL database on Cloud SQL.

For latency optimization for a Expo Demo in Chicago, the application is hosted in the us-central1 region, which is one state West in Iowa.

Working with GCP services requires the installation of the namesake SDK, and a proxy script for connecting to a hosted database.

• Google Cloud SDK
• Cloud SQL Proxy

Django is also used to serve frontend content found in the apollo-ui project via the django-webpack-loader module. Build the frontend via webpack, put the webpack-stats.json in the project root directory, and the bundles in /assets/bundles. If this is unclear, look at .gitignore for details.

Before deploying a new version of the application, static assets must be collected and uploaded. As with other django management commands, this will fail if environment variables have not been set.

python manage.py collectstatic
./sync_static.sh

Once static files are uploaded, the app itself can be deployed.

gcloud app deploy