cyphar / ncss-coffeerun

NCSS Coffeerun website

NCSS CoffeeRun by Maddy

Running Locally

• Create a virtual environment: virtualenv --python="$(which python3)" env
• Activate the virtual enviroment: source env/bin/activate
• install requirements: pip install -r requirements/dev.txt
• Create the database and tables and a default user with create_db.py python3 create_db.py
• Run on command line:
  • python3 run.py (for the web ui)
  • python3 coffeebot.py (for the slack bot)
• Open the browser at http://localhost:5000 (note: Use localhost, not 127.0.0.1)

Running tests locally

• Activate the virtual enviroment: source env/bin/activate
• Run the coffeespec tests: python test_coffeespecs.py

Running On The Website

• Push changes to the site using git push heroku master
• The program is started by this
• There's no DB migration, manually do any changes before you push the dependent changes

Making a new instance

Prereqs

1. Heroku CLI - https://devcenter.heroku.com/articles/heroku-cli
2. psql - idono, search your package manager (should come with postgres)
3. git clone [e-mail]:ncss/ncss-coffeerun.git
4. python -m venv .venv
5. source .venv/bin/activate (Every time you want to run/use this code)
6. pip install -r requirements/prod.txt (You need postgres, and you might have to mess with versions because of changes in Python - remove the version requirement and see what happens!)

Heroku

1. Create a new app: https://dashboard.heroku.com/new-app
2. heroku heroku git:remote -a <heroku-app-name> (add --remote=<blah> if you want to have a different name)
3. git push heroku master (replace heroku with custom remote name if you did that)

Slack

1. Create a new Slack app: https://api.slack.com/apps?new\_app=1
2. Add a bot user (and always show as online - but it probably doesn't matter)
3. Install app to workspace (under OAuth & Permissions menu)
4. Add redirect urls (under OAuth & Permissions)
   • https://<heroku-app-name>.herokuapp.com/login/authorized
   • https://<heroku-app-name>.herokuapp.com/team-auth-done/

Auth things

1. On Slack app settings page under Basic Information get the Client ID and Client Secret
2. On Heroku, go to the Settings tab and Reveal Config Vars
   • Add KEY=SLACK_OAUTH_CLIENT_ID, VALUE=<client_id>
   • Add KEY=SLACK_OAUTH_CLIENT_SECRET, VALUE=<client_secret>

Database things

1. export SLACK_OAUTH_CLIENT_ID='<client_id>'
2. export SLACK_OAUTH_CLIENT_SECRET='<client_secret>'
3. export DATABASE_URL=$(heroku config:get DATABASE_URL -a <heroku-app-name>
4. (In virtualenv) python create_db.py

More auth and database things

1. Go to https://.herokuapp.com/ and login with Slack
2. Go to https://.herokuapp.com/team-auth/ and add it to the right workspace and channel
   • "Access token stored in db" is Good!
3. Get the bot user details
   • Slack user id (This is probably terrible and totally wrong): New direct message to the bot user, pick out the part after /team/ in the URL you get when you hover over the @name.
   • Bot OAuth access token: In Slack app settings, OAuth & Permissions, copy the Bot token
4. Hack the database
   1. psql psql $(heroku config:get DATABASE_URL -a <heroku-app-name>
   2. update slack_team_access_token set coffee_bot_slack_user_id = '<user_id>', coffee_bot_slack_access_token = '<bot_token>', wants_slack_notifications=true;

Last steps

1. On Heroku, go to the Resources tab
2. Enable the worker python coffeebot.py
3. Add the bot user to the right channel somehow (I like to @ it)