❗ This project has been abandoned. Use Laminar, instead of this AngularJS based approach. It was an experimental project and it served its purpose before I could bring it to completion (the purpose of learning). It doesn't compile because I left it in a refactoring state

Table of Contents

• What does it do?
• Tech stack
• How to run
• Personal notes

What does it do?

This project represents an anonymous poll solution. One can:

• Create polls with 3 types of questions (Number, Choice, OpenEnd)
• Invite voters via email using Gmail or MailChimp (TODO)
• Offers a results link
• Optionally provide weights to each individual invited voter (TODO)
• Get export via CSV (TODO)

All data lives in-memory and nothing is stored forever, since H2 is used as a main data store. On every server restart or every couple of days (cron job), the data is wiped clean.

Tech stack

Library	Description
cats-effect-3	❤️ Typelevel 🐐 ❤️ Functional goodies
munit-cats-effect-3	Testing
http4s	HTTP Server / Client
tapir	HTTP API server, client & documentation wrapper
scalajs	What is says on the tin
scalatags	HTML
scalacss	CSS
h2	In-memory database
doobie	SQL library
flywaydb	Database migration
ciris	Configuration
log4cats	Logging
circe	JSON
newtypes	Type-safe aliases

Plugin	Description
sbt-scalajs	Compiling, running and testing with Scala.js
sbt-scalajs-crossproject	Cross-platform compilation support for sbt Scala & ScalaJS
sbt-web-scalajs	fastOptJS / fullLinkJS
sbt-web-scalajs-bundler	npm / webpack
sbt-hepek	Generates files from Scala code (used for index.html at build step)

^I'm still in the dark when it comes to how people do Scala.js development; if any of these plugins are obsolete or no longer used, please open a PR

How to run

To run the backend, simply:

sbt backend/run

For a development build (http://localhost:8000):

sbt fastOptJS/webpack
cd frontend && python -m http.server

For a production-grade build (http://localhost:8000/index-prod.html):

sbt fullOptJS/webpack
cd frontend && python -m http.server

Personal notes

Subjected to random updates over time:

• DRY can be extended to frontend as well now, since there's no need to write any TS/JS definitions for types, objects or API calls (2022.07.02)
• It's a nuisance at first to follow the code path / structure, especially when it comes to shared code (common/shared directory) (2022.07.02)
• I've been working with the IntelliJ EAP version & Scala nightly builds for this project. Decent Scala 3 support is still far, far away 😭 (2022.07.02)
• About ciris:
  • No longer must the developer worry about .conf files & bugs that result from their merger (e.g. in tests or local setups) (2022.07.02)
  • A default working local config can be provided from the code itself, without the need for any initial setup, such as setting environment variables (2022.07.02)
• About tapir:
  • Since tapir can be leveraged for 3 different purposes (documentation, API definition, client libraries), this makes it an excellent choice for a Scala & Scala.js project (2022.07.02)
  • Being able to share the endpoint definition between backend & frontend just feels great (2022.07.02)
  • Had to inline a tapir code snippet (http4s) that was not built for Scala.js & also to strip away any fs2 mention since it greatly encumbers the size of the javascript files (output) (2022.07.02)
• About doobie & h2:
  • h2 can easily be replaced with PostgreSQL, since SQL compatible code was written (works with both) (2022.07.02)
  • Had to do a custom intervention for Meta[Array[A]] since it was not working (2022.07.02)
  • Psst. Pass the doobie, Scooby. (2022.07.02)