Skip to content

feat(configs): expand directory topics and add 12 verified live feed configs - #320

Open
gildesmarais wants to merge 3 commits into
masterfrom
feat/expand-directory-topics-and-configs
Open

feat(configs): expand directory topics and add 12 verified live feed configs#320
gildesmarais wants to merge 3 commits into
masterfrom
feat/expand-directory-topics-and-configs

Conversation

@gildesmarais

@gildesmarais gildesmarais commented Aug 30, 2026

Copy link
Copy Markdown
Member

What changed

  • 12 new feed configs under lib/html2rss/configs/:
    • aleph-alpha.com/news.yml, banrep.gov.co/noticias.yml, bostondynamics.com/news.yml, embl.org/news.yml, fdic.gov/press-releases.yml, figure.ai/news.yml, ipcc.ch/news.yml, press.un.org/press-releases.yml, replicate.com/changelog.yml, resend.com/changelog.yml, scale.com/blog.yml, together.ai/blog.yml
  • 17 existing configs retagged to use expanded directory.topics slugs already in the core validator (health, education, rights, transport):
    • University news: scienceeducation (aalto, anu, chalmers, ethz, imperial, kaist, ntu, tudelft, tue)
    • Health agencies: sciencehealth (who, ecdc); MSF adds health
    • Courts: add rights (curia, icj)
    • Transport agencies: add transport (dfs, faa, imo)

Configs-only PR — skill/docs changes from the original branch were dropped; master already has the curation workflow update (#322).

Why

Feed Directory browse needs configs tagged with the newer topic vocabulary, plus a batch of live-verified surfaces (AI vendors, changelogs, civic/finance/regulatory news) that were missing from the catalog.

Risk

  • Selector drift on JS-heavy or redesign-prone sites (replicate/resend changelogs, AI vendor blogs)
  • Topic retags are metadata-only — no selector or URL changes on existing configs
  • Low merge conflict risk with other in-flight config PRs if they touch the same retagged files

Review map

  1. New configs (12 files) — spot-check selectors and directory.title / summary on one changelog (replicate.com/changelog.yml), one civic (press.un.org/press-releases.yml), one finance (fdic.gov/press-releases.yml)
  2. Topic retags (17 files) — confirm directory.topics only; no accidental selector edits
  3. Dynamic fetch lane — new configs will be picked up by spec/html2rss/configs_dynamic_spec.rb on CI

Validation

Original branch verification (pre-rebase onto current master):

  • bin/validate_configs — all configs passed
  • bundle exec rspec spec/ — 1,574 examples, 0 failures
  • yamllint, rubocop, Prettier — clean

Not re-run after rebase onto 0d80e4b (configs-only rewrite). CI on this PR is the authoritative re-check.

@gildesmarais
gildesmarais force-pushed the feat/expand-directory-topics-and-configs branch from 5ba1645 to 505153f Compare August 30, 2026 20:34
@gildesmarais gildesmarais changed the title feat(configs): expand directory topics and add 112 curated global and flagship feed configs feat(configs): expand directory topics and add 12 verified live feed configs Aug 30, 2026
@gildesmarais
gildesmarais force-pushed the feat/expand-directory-topics-and-configs branch 2 times, most recently from d5c854a to 0248bd4 Compare August 30, 2026 21:43
@gildesmarais
gildesmarais force-pushed the feat/expand-directory-topics-and-configs branch from 0248bd4 to 7a0edd2 Compare September 1, 2026 21:42
The PR retags configs with health, education, rights, and transport topics
that landed in html2rss after the published 0.28.0 release.
Fix duplicated titles on Banrep and UN press releases, drop FDIC glider noise, and turn on enhance where MCP showed clean descriptions without page chrome.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant