Website

This is the application for generating a static website for podcasts, which means output HTML/assets and can be served by just a web server. it is quite inspired by unix philosophy for deployment pipeline, and functional programming for stateless. it's currently used for https://kaigaiiju.ch.

Design Concept

• dependencies as little as possible

Try to make fewer external libraries and use a more standard solution of Ruby/Rails which is a lot of maintenance. for instance, not necessary to use RSpec.

• Try to follow Rails way architecture

more discussion can be found here on wiki

• Use Database modeling with a DDD perspective

use database constraints and modeling effectively, ERD can be found here.

• (experimental) CQRS

Write operation and Read Operation have different requirements, try not to mix it up.

Deployment pipeline with Github Action

if the latest commit message contains RELEASE_TRRIGER_MESSAGE then it will trigger the release workflow to build the website via kaigaiijuch/release.

 # e.g. with RELEASE_TRRIGER_MESSAGE=release!
 $ git commit --allow-empty -m 'release!'
 $ git push origin main

it will trigger a build and make a pull-request on kaigaiijuch/kaigaiijuch.github.io. the trigger defined at here

also there is fetch_rss trigger scheduled, it can trigger via curl. (you need token)

curl -L -X POST -H "Accept: application/vnd.github+json" -H "Authorization: Bearer ${GITHUB_TOKEN}" -H "X-GitHub-Api-Version: 2022-11-28" https://api.github.com/repos/kaigaiijuch/website/dispatches -d '{"event_type":"fetch_rss"}'

development requirement

• Ruby version: check .ruby_version
• sqlite3

setup

local

 $ bin/setup # with DATA_REPO=your_data_repo, it will clone the data repository
 $ bin/rails s
 $ open http://localhost:3000 # caution it's not HTTPS

docker

  $ docker-compose up
  $ open http://localhost:13000 # caution it's not HTTPS

Prepare data

fetch

 $ bin/data/fetch_all [rss_feed_url] [csv/directory] [answers.csv] [transcription/csv/directory]

This is idempotent operation, it fetches the data from the sources and store it in the data repository.

• rss_feed_url: Spotify for podcasters RSS feed is supported, it fetches the feeds data the data is stored default in FeedsSpotifyForPodcaster.

  important convention: the description of the episode should be ended by formatted as #123-a title where 123-a is the episode number, it can be alphanumeric.

• csv/directory: (this is a temporary solution) Import data from csv files in the directory, it is compatible with google sheets exported csv format, sample file: (TBD).

• answers.csv: (this is a temporary solution) Import answers from csv file, it is compatible with exported csv format of google forms linked spreadsheet. The format is #{question_number}: 【{category}】{question_original_title}, sample file: (TBD).

• transcription/csv/directory: (this is a temporary solution) Import transcriptions from csv file. The file name should be #{episode_number}.transcription.csv, sample file: (TBD).

commit data

 $ bin/data/commit

this will commit the data to the data repository.

static data

• data/episodes/#{episode_number}.chapters.txt - chapter data for the episode. the data format is HH:MM:SS.mmm title. sample file is here

development instructions

For adding new pages, be aware that it may needs to added for page list on bin/pages/list.rb and site map on config/sitemap.rb.

Configuration

check .env file and satisfy the requirements, they are used in config/application.rb.

TZ is used for setting timezone.

Database creation && Database initialization

  $ bin/rake db:create db:migrate db:seed

How to run the test suite and linter

 $ bin/rake

Deployment instructions

 $ bin/build

it will generate the static pages in public/ directory based on the path list in bin/pages/list.rb.

NOTE: after generate in public/ directory, the pages are served as static pages not by server. use bin/clean to remove static pages.

for github build workflow: it needs to set DATA_REPO_TOKEN as a secret.

Sitemap generation

 $ bin/rake sitemap:refresh