rc-niceties

An application to manage niceties delivered by recursers at the end of their batch.

Application architecture

This application is pretty simple. It stores niceties as text and some metadata associated with a unique (author, recipient, batch) tuple. This is stored in a Postgres database. Faculty can view and edit all the niceties. Recursers can see all the nice things said about them, and all the things they have said in the past.

The backend is in Flask, and is just a REST API. The DB is Postgres, with SQLAlchemy as an ORM and Alembic to manage schema migrations. The frontend is in react, because it's 2016 and that's the current hotness.

Creating a development environment

I built this using Python 3.6.13, node.js 14.15.5 and Postgres 12.6.

1. Clone the repository into an appropriate location.

2. Set up your Python virtual environment by running pyvenv venv in that directory and running source venv/bin/activate to active it.

3. Install Python requirements with pip install -r requirements.txt. You may need to install some build prerequisites; on Debian-like systems, they include the packages python3-dev and libpq-dev.

4. Install the frontend requirements with npm install.

5. Install PostgreSQL and create an empty database, e.g. createdb rcniceties

6. Set the following environment variables, noting that the application will not run if any of these is not set:

   • FLASK_APP - the location of the app, e.g. backend:app
   • FLASK_SECRET_KEY_B64 - a base64-encoded random secret string, for example generated by running:

     from base64 import b64encode
     from os import urandom
     print(b64encode(urandom(24)))

   • DATABASE_URL - the database connection URL e.g. postgres://localhost/rcniceties
   • RC_OAUTH_ID - your Recurse Center OAuth application ID
   • RC_OAUTH_SECRET - your Recurse Center OAuth application secret
   • DEV - set to either TRUE or FALSE, depending on if this is a development or production environment
   • DEBUG_SHOW_ALL (optional) - set to TRUE to show every nicety in the DB on the Niceties For Me page (useful for debugging) or FALSE (default) for normal behavior

   A common way of setting up these environment variables is with a .env file in your project directory, containing export ENV_VAR=value on each line. This can be loaded by running source .env and will be automatically loaded by heroku local.

7. Optionally mock out the RC API by setting MOCK_OUT_RC_API = True in backend/__init__.py. This means you do not have to set RC_OAUTH_ID or RC_OAUTH_SECRET, but you'll only get sample data (contained in the mock/fixtures folder, and with request -> filename mapping in mock/rc.py). Alternatively, you'll need to set up an RC application with a redirect URI pointing to your local server (e.g. http://localhost:8000/login/authorized) or with the special value urn:ietf:wg:oauth:2.0:oob.

8. At the command prompt, run flask db upgrade to create the DB tables.

9. Run python, and inside it run:

   from backend import config
   config.set_to_default()

To run:

1. Compile the frontend static files by running npm run build.

2. Run the Flask application with gunicorn backend:app --log-file -.

Deploying

This is designed to be deployed to Heroku. To do this:

1. Enable the Python and node.js buildpacks for the application.

2. Set up a Postgres database for the application and run heroku pg:push [database-name] DATABASE_URL to copy your local database to Heroku.