Skip to content

podcast-sync-plan.md

Latest commit

 

History

History
99 lines (79 loc) · 4.96 KB

podcast-sync-plan.md

File metadata and controls

99 lines (79 loc) · 4.96 KB
Raw

Plan: Keep The PowerShell Podcast synced to the repo

Scope & key facts

  • content/podcast/ holds two shows: the archived PowerScripting Podcast (2007–2020, libsyn audio, original hosts) and The PowerShell Podcast (2022→, Podbean audio, Andrew Pla). They are told apart by audio host: mcdn.podbean.com ⇒ The PowerShell Podcast. This work touches only The PowerShell Podcast (the 220 modern, Podbean-hosted episodes).
  • The Podbean RSS feed (feed.podbean.com/powershellpodcast/feed.xml) is truncated to the 10 most recent items — it can keep the recent tail in sync but can never backfill history.
  • Episode number is embedded in the existing podcast_url filename (..._episode_220_...) for 217/220 modern files — a reliable local key.

Workstream 1 — Author rewrite (one-time, local)

For every modern (Podbean) episode:

  • author: → Andrew Pla (the byline shown on the page).
  • authors: → replace James Petty with Andrew Pla; insert Jordan Hammond when date ≤ 2023-12-25 (the "Farewell to Jordan" episode); keep all existing guests; dedupe; Andrew first.
  • James Petty remains the author of his 32 articles — only the podcast mis-attribution is corrected.

Workstream 2 — Frontmatter enrichment (one-time, local + playlist)

Add to each modern episode:

  • episode: — parsed from the podcast_url filename (217/220; 3 Summit/bar specials set by hand or left blank).
  • guid: — only where the live feed provides it (recent items). Old episodes never re-appear in the feed, so they need no guid.
  • youtube: — the youtu.be id. 76/220 already carry it in-body; the rest are matched against the YouTube playlist by episode number / title. Missing ⇒ no embed, template falls back to the icon.

Workstream 3 — Backfill the gap (one-time) — subsumed by WS4

Originally scoped as a YouTube/Podbean scrape of episodes that had rolled off the 10-item feed window. Not needed: the live feed turned out to carry the full archive (see ADR 0003 update), so the first WS4 run generated the entire gap (eps 221–234) directly. No scrape was performed.

Workstream 4 — Ongoing RSS sync (automated)

  • Runtime: Node + node-fetch, matching discourse-sync.yml / migrate-posts.cjs.
  • Trigger: scheduled GitHub Action (weekly, Tuesday — after the Monday release), workflow_dispatch enabled. Auto-commits to the branch; Netlify rebuilds on push. Must run at least every ~10 weeks or feed items roll off.
  • Implemented: .github/scripts/sync-podcast-feed.js (add-only) + .github/workflows/podcast-sync.yml (weekly Tuesday, workflow_dispatch, auto-commit). Run node .github/scripts/sync-podcast-feed.js --dry-run to preview.
  • Per feed item:
    1. Match existing files by enclosure URL, else guid, else episode number → skip if present. (All existing files carry podcast_url; none carried guid, so the enclosure URL is the universal key.)
    2. Filename YYYY-MM-DD-<slug>.md, date from pubDate, aliases: /YYYY/MM/<slug>/ (matches migrated convention).
    3. author: Andrew Pla; guests extracted conservatively (high-confidence only, never remove) — bio-section + with <Name> title patterns ported from scripts/update-podcast-authors.py.
    4. podcast_url = enclosure; episode = number parsed from the enclosure filename (the repo convention — itunes:episode runs one ahead, see ADR 0003); guid; youtube = youtu.be id from the notes. duration was omitted to keep the synced frontmatter identical to the existing files (no episode currently carries it).
    5. Body: content:encoded HTML → markdown, strip recurring boilerplate (Andrew's links, PDQ Discord, Summit promo, redundant YouTube line) via a maintained strip-list; keep episode-specific resource links.

Workstream 5 — Rendering (layout changes)

  • podcast/list.html & _default/single.html: render the YouTube thumbnail (https://img.youtube.com/vi/<id>/maxresdefault.jpg) hotlinked, falling back to the existing purple icon when no youtube id.
  • Single page: click-to-load YouTube embed facade (thumbnail poster → iframe on click) alongside the existing audio player.

Workstream 6 — Author profiles (one-time)

Scaffold profiles with tools/new-author.ps1 so the rewritten bylines resolve to real bio cards instead of bare taxonomy pages (per ADR 0002):

  • content/authors/andrew-pla/_index.md — host. Tagline + bio, links (e.g. andrewpla.tech/links), avatar.
  • content/authors/jordan-hammond/_index.md — early co-host.

Slugs must match Hugo's slug of the exact authors: string (Andrew Pla → andrew-pla, Jordan Hammond → jordan-hammond); new-author.ps1 enforces this.

Design record

The truncated-feed constraint and the incremental-sync-vs-backfill split are captured in ADR 0003. The two-podcast distinction is in CONTEXT.md.