Skip to content

AmericanAirlines/Hangar

Repository files navigation



Managing a hackathon? You've come to the right place! American Airlines loves to help devs learn new tech and we're passionate about being a big part of hackathons.

What is Hangar?

Hangar is a hackathon management platform that can help with everything from project registration, communicating prizes, listing schedule info, and provides judging.

Features

  • Event info: share info about the hack with your attendees and link to resources
  • Project registration: collect info from attendees about their hack
  • Prize info: communicate prizes of various types
  • Schedule: list your event schedule so attendees have access
  • Judging: judge projects either by fixed criteria or by "expo" style comparison judging

Table of Contents


Using Hangar

Interested in deploying an instance of Hangar to help with your hack? Follow the steps below to deploy and customize an instance of Hangar to use with your hack.

Deployment

Hangar is containerized via Docker; simply build the docker image and deploy to a cloud environment of your choice. See the sections within Setup to generate the necessary environment variables and make them accessible to your app at runtime.

Authentication

Authentication uses OAuth via Ping Federate OR Slack but it can be easily modified to use a new callback from a different OAuth provider. See the Ping Federate or Slack below for full details on how to setup and configure your SSO solution.

Feature Utilization

In Hangar's current state, you'll need to manually modify the Admin table to create a new record for your admin users. Once created, your admin users will have full access to create judging sessions and see their results.

Customization

In order to modify the content of the homepage and to make other modifications to the presentation of the app, simply modify the values in packages/shared/src/config.


Development

Prerequisites

  • Node 18+

  • Yarn v1

    This project is a monorepo, so Yarn was chosen for better dependency management

Setup

  1. Open the project via the workspace file (e.g., code hangar.code-workspace)

  2. Install dependencies

    yarn
  3. Copy packages/api/.env.sample to packages/api/.env.local

    cp packages/api/.env.sample packages/api/.env.local

    PostgreSQL

    If you don't have Postgres installed already, see the Installation and Use section below.

    After installing, create a DB with the name hangar (or use another name and override DATABASE_URL in your .env.local).

    Installation and Use

    macOS

    Using Postgres.app is recommended as the installation doesn't require a password and is generally easier to use that the traditional Postgres app below.

    Windows/macOS/Linux

    During the installation process (if you follow the steps on postgresql.org), you will be prompted to set a password - make sure to use something you'll remember.

    Viewing/Editing the DB

    If you'd like a visual way of viewing or editing your local database, try using TablePlus.

    Seeding the Database

    You can seed the database using the mikro-orm/cli tool.

    You can drop, create, migrate, and seed the database any time you need with this command (run from packages/api):

    • yarn mikro-orm migration:fresh --seed DatabaseSeeder

    ⚠️ NOTE: Running the above command will delete all data in your database

    Database migrations

    To run migrations locally, use yarn migrate from the root. If you ever want to replace your db contents with a fresh setup, run yarn db:fresh.

    If you need to make a new migration, simply run yarn workspace @hangar/database migration:create. There are also scripts for running the up and down commands: migration:up and migration:down respectively.

    If you need to incorporate a new migration from a recently merged feature branch, run yarn db:seed; db:seed will perform the migration:up action and seed the db, potentially including data relevant to the new migration.

    Reverting a migration locally

    After creating a migration locally that needs to be reverted, follow the process below to revert your local DB state:

    • Run yarn @hangar/database migration:down
    • Delete the new migration
    • Run yarn @hangar/database snapshot:reset
      • ⚠️ This will overwrite ANY changes made to your DB structure, not just the last migration
    • Begin the migration creation process again

    Updating Mikro-ORM

    The process to update all packages is a little painful because ALL Mikro-ORM dependencies across BOTH packages must be kept in sync. Run both commands below (after modifying the version to match whatever target needed)

    # API
    yarn workspace @hangar/api add @mikro-orm/core@^6.0 @mikro-orm/knex@^6.0 @mikro-orm/postgresql@^6.0 mikro-orm@^6.0
    
    # DB
    yarn workspace @hangar/database add @mikro-orm/cli@^6.0 @mikro-orm/core@^6.0 @mikro-orm/knex@^6.0 @mikro-orm/migrations@^6.0 @mikro-orm/postgresql@^6.0 @mikro-orm/seeder@^6.0 @mikro-orm/postgresql@^6.0 mikro-orm@^6.0

    Restoring a Production DB locally

    In order to debug an issue locally, it can be helpful to mirror the prod DB locally. When doing so, make sure to avoid actions that will trigger text messages because user data is REAL. Always re-seed the DB after fixing your issue to avoid sending erroneous texts.

    ⚠️ WARNING: Make sure to follow these steps closely. Making an error below has the potential to overwrite production data if your token allows for write access.

    1. Open TablePlus (no need to connect to a DB)
    2. Go to File > Backup (Windows steps TBD), choose your production database connection
    3. Select the prod database
    4. Make sure the Postgres version is PostgreSQL 14.0 and set the options to be --no-owner and --format=custom
    5. Click Start Backup
    6. (OPTIONAL) Drop all current data to ensure your local copy is an exact replica, otherwise some local data may persist
      1. Open a connection to your local database, select all tables and functions (command + a on macOS), right click, and choose "Delete"
      2. When prompted, check the option for Cascade and click OK
    7. Wait for the backup to finish and then go to File > Restore (Windows steps TBD), choose your local database connection
    8. Remove all flags from the left hand side (primarily, --single-transaction which will cause your restore to fail on conflict)
    9. Select your database on the right hand side
    10. Click Start restore

    Restarting Table Sequences

    If you add data manually to your database (through a tool like TablePlus) and do NOT let Postgres assign an id automatically, you will disrupt the sequence for that table that determines the next available id to assign. If that happens, perform the following queries to restart it.

    Find the appropriate sequence name:

    SELECT sequence_schema, sequence_name
    FROM information_schema.sequences
    ORDER BY sequence_name;

    Restart the sequence with a new id:

    ALTER SEQUENCE "YOUR_TABLE_SEQUENCE" RESTART WITH THE_NEXT_ID_TO_ASSIGN;

    (e.g., ALTER SEQUENCE "Provider_id_seq" RESTART WITH 11;)

    SSO

    Ping Federate

    Ping Federate is enabled by default for SSO. To configure it, simply add the following environment variables to your .env.local:

    • PINGFED_CLIENT_ID: Your client ID
    • PINGFED_CLIENT_SECRET: Your client secret
    • PINGFED_AUTH_BASE_URL: The URL which starts the authentication flow
    • PINGFED_TOKEN_BASE_URL: The URL from which the token containing user data can be retrieved

    Slack

    To configure the app to use Slack for SSO Authentication, modify the packages/shared/src/config/auth.ts and set slack as the auth method. Next, create a Slack app using the following manifest after selecting From an app manifest (NOTE: watch out for whitespace issues when copying):

    display_information:
      name: Hangar Dev (YOUR_FIRST_NAME)
      description: Dev instance of Hangar
      background_color: '#000000'
    features:
      bot_user:
        display_name: Hangar Dev (YOUR_FIRST_NAME)
        always_online: true
    oauth_config:
      redirect_urls:
        - [YOUR_NGROK_URL]/api/auth/callback/slack
      scopes:
        user:
          - email
          - openid
          - profile
        bot:
          - chat:write
          - chat:write.public
    settings:
      org_deploy_enabled: false
      socket_mode_enabled: false
      token_rotation_enabled: false

    After creating the app, use the sidebar and go to Install App and request to install the app to your workspace. Once approved, head back to this section and Install to workspace.

    After installation, go to OAuth & Permissions and copy the Bot User OAuth Token value (starting with xoxb-) and use that for your SLACK_BOT_TOKEN.

    You will also need to go to Basic Information and use Client ID and Client Secret as your SLACK_CLIENT_ID and SLACK_CLIENT_SECRET, respectively.

    For Slack OAuth to work, your bot needs to be configured with your redirect URL. Go to OAuth & Permissions if you need to update your Redirect URL (e.g., [YOUR_NGROK_URL]/api/auth/callback/slack).

Starting the App

  1. Start the development server

    yarn dev

    When you see this success message, open the url to load the site

    🚀 Listening at http://localhost:3000

    The server starts before Next.js finishes compiling, so the first time may take a little bit to load

  2. After your initial setup, you'll need to run your database migrations:

    yarn migrate
    
  3. Start developing

  4. For Slack integration, make sure to copy your ngrok address into BASE_URL config var (with no trailing slash!)

Containerization

Hangar is containerized to make deployment efficient and to prevent vendor-dependence for our cloud environment

Building and Running Docker Locally

  1. Installing relevant dependencies
  2. Duplicate .env.docker.sample as .env.docker and modify values as needed
  3. Download Docker Desktop and start it
    • Optionally modify Settings > Resources > Advanced to provide more resources to Docker and speed up your build commands
  4. Run yarn docker:build from the root to build your image
  5. Run yarn docker:run to start your container
  6. Visit localhost:3001 to use the app running in Docker
Environment Variables

packages/api/.env.local is used as the base for all environment variables but .env.docker will override any specified environment variables as needed.

Database User

On macOS Postgres user/pass is typically not required. If you do not use one in your API .env.local, you WILL need to specify one in .env.docker; the password is typically blank (no value needed) and your username is typically your machine user (use whoami in terminal).

Troubleshooting

If you are unable to start your Docker app, make sure the steps above have been followed, then:

  • Check the accuracy of your .env.docker and packages/api/.env.local values (the former will override the latter)
  • Make sure no quotes are used in either file; Docker does not remove them and your value will include them
  • You can use Docker Desktop to inspect the environment: Docker Desktop > Containers / Apps > hangar (environment variables can be found on the Inspect tab)