feat(podcast): add automated RSS feed sync + backfill eps 221-234

Add an add-only sync that generates episode files from the Podbean feed,
plus a weekly scheduled Action that auto-commits new episodes (mirroring
discourse-sync.yml). The first run backfills the gap between the repo
(ep 220) and the feed (ep 234).

The feed turned out to carry the full archive rather than a 10-item
window, so idempotency is keyed on the enclosure URL (every existing
file has it) and the episode number is parsed from the Podbean filename,
since itunes:episode runs one ahead of the repo convention. This makes
the separate WS3 scrape unnecessary; ADR 0003 and the plan are updated
to record the corrected design. Guest extraction stays conservative
(high-confidence only) because the Action commits without review.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Qv5XTRCp9s9cH6SikDkrnS

Author: HeyItsGilbert

.github/scripts/sync-podcast-feed.js

@@ -0,0 +1,367 @@
+// .github/scripts/sync-podcast-feed.js
+// Incremental sync of The PowerShell Podcast (Podbean) into content/podcast/.
+//
+// Design: see docs/adr/0003-incremental-podcast-sync.md.
+// The feed currently carries the full archive, but this script is strictly
+// ADD-ONLY: it generates an episode file only when no existing file matches,
+// and never edits or deletes. That makes a full pass over the feed safe and
+// lets the first run backfill the whole gap between the repo and the feed.
+//
+// Idempotency key (in priority order): enclosure URL, then RSS guid, then
+// episode number. The enclosure URL is the universal key — every existing
+// modern file carries it as `podcast_url`.
+//
+// Run `node .github/scripts/sync-podcast-feed.js --dry-run` to preview.
+
+import fetch from 'node-fetch';
+import fs from 'fs';
+import path from 'path';
+
+const FEED_URL = 'https://feed.podbean.com/powershellpodcast/feed.xml';
+const PODCAST_DIR = path.join('content', 'podcast');
+const HOST = 'Andrew Pla';
+const TITLE_PREFIX = 'The PowerShell Podcast ';
+const DRY_RUN = process.argv.includes('--dry-run');
+
+// Recurring boilerplate links lifted/dropped from episode bodies. A body line
+// (typically a "Resource Links" bullet) is dropped if it contains any of these.
+// Keep this list tight — episode-specific links must survive. youtu.be is
+// dropped because the id is lifted into the `youtube` frontmatter field.
+const BOILERPLATE = [
+  /andrewpla\.tech/i,
+  /discord\.gg\/pdq/i,
+  /powershellsummit\.org/i,
+  /youtu\.be\//i,
+];
+
+// --- XML helpers -----------------------------------------------------------
+
+function decodeEntities(str) {
+  return str
+    .replace(/&#x([0-9a-fA-F]+);/g, (_, h) => String.fromCodePoint(parseInt(h, 16)))
+    .replace(/&#(\d+);/g, (_, d) => String.fromCodePoint(parseInt(d, 10)))
+    .replace(/'/g, "'")
+    .replace(/"/g, '"')
+    .replace(/&lt;/g, '<')
+    .replace(/&gt;/g, '>')
+    .replace(/&nbsp;/g, ' ')
+    .replace(/&amp;/g, '&');
+}
+
+// Read the text of a single child tag from an <item> block, unwrapping CDATA.
+function tag(item, name) {
+  const m = item.match(new RegExp(`<${name}(?:\\s[^>]*)?>([\\s\\S]*?)</${name}>`, 'i'));
+  if (!m) return '';
+  let v = m[1].trim();
+  const cdata = v.match(/^<!\[CDATA\[([\s\S]*?)\]\]>$/);
+  if (cdata) return cdata[1];
+  return decodeEntities(v);
+}
+
+function attr(item, tagName, attrName) {
+  const m = item.match(new RegExp(`<${tagName}\\b[^>]*\\b${attrName}=["']([^"']*)["']`, 'i'));
+  return m ? m[1] : '';
+}
+
+// --- Slug + date -----------------------------------------------------------
+
+// Mirrors Hugo's default slugify closely enough to match the existing modern
+// filenames: lowercase, drop apostrophes, collapse runs of other non-alnum
+// characters to single hyphens, trim hyphens.
+function slugify(s) {
+  return s
+    .toLowerCase()
+    .replace(/['‘’]/g, '')
+    .replace(/[^a-z0-9]+/g, '-')
+    .replace(/^-+|-+$/g, '');
+}
+
+function pad2(n) {
+  return String(n).padStart(2, '0');
+}
+
+// pubDate -> { date: 'YYYY-MM-DD', iso: 'YYYY-MM-DDTHH:MM:SS+00:00', y, m }
+function parseDate(pubDate) {
+  const d = new Date(pubDate);
+  const y = d.getUTCFullYear();
+  const m = pad2(d.getUTCMonth() + 1);
+  const day = pad2(d.getUTCDate());
+  const iso = `${y}-${m}-${day}T${pad2(d.getUTCHours())}:${pad2(d.getUTCMinutes())}:${pad2(d.getUTCSeconds())}+00:00`;
+  return { date: `${y}-${m}-${day}`, iso, y, m };
+}
+
+// --- Episode number from enclosure filename --------------------------------
+
+// The repo's episode-number convention is the number embedded in the Podbean
+// filename (`..._episode_NNN_...`), NOT itunes:episode (which runs one ahead).
+// Most filenames delimit the number (`episode_220_Morten`), but specials glue a
+// random suffix onto it (`episode_2298xv9d`). For the ambiguous case, pick the
+// digit-prefix closest to itunes:episode (the numbers track within ~1).
+function resolveEpisode(enclosureUrl, itunesEp) {
+  const clean = enclosureUrl.match(/episode[_-](\d+)[_-]/i);
+  if (clean) return parseInt(clean[1], 10);
+
+  const greedy = enclosureUrl.match(/episode[_-](\d+)/i);
+  if (!greedy) return null;
+  const digits = greedy[1];
+  if (!itunesEp) return parseInt(digits, 10);
+
+  let best = null;
+  for (let len = digits.length; len >= 1; len--) {
+    const cand = parseInt(digits.slice(0, len), 10);
+    if (Math.abs(cand - itunesEp) <= 2) { best = cand; break; }
+  }
+  return best != null ? best : parseInt(digits, 10);
+}
+
+// --- HTML body -> markdown -------------------------------------------------
+
+function htmlToMarkdown(html) {
+  let md = html;
+
+  // Lists
+  md = md.replace(/<\/?ul[^>]*>/gi, '\n');
+  md = md.replace(/<\/?ol[^>]*>/gi, '\n');
+  md = md.replace(/<li[^>]*>([\s\S]*?)<\/li>/gi, (_, t) => `- ${t.trim()}\n`);
+
+  // Links, emphasis
+  md = md.replace(/<a\b[^>]*href=["']([^"']*)["'][^>]*>([\s\S]*?)<\/a>/gi, '[$2]($1)');
+  md = md.replace(/<(strong|b)\b[^>]*>([\s\S]*?)<\/\1>/gi, '**$2**');
+  md = md.replace(/<(em|i)\b[^>]*>([\s\S]*?)<\/\1>/gi, '*$2*');
+
+  // Block + break tags
+  md = md.replace(/<br\s*\/?>/gi, '\n');
+  md = md.replace(/<\/p>/gi, '\n\n');
+  md = md.replace(/<p[^>]*>/gi, '');
+
+  // Strip anything left, decode, tidy
+  md = md.replace(/<\/?[a-zA-Z][^>]*>/g, '');
+  md = decodeEntities(md);
+
+  md = md
+    .split('\n')
+    .map((line) => line.replace(/ /g, ' ').replace(/[ \t]+$/g, '').trim())
+    .join('\n');
+  md = md.replace(/\n{3,}/g, '\n\n');
+  // Tighten lists: drop blank lines between consecutive bullets.
+  md = md.replace(/(^- .*)\n\n(?=- )/gm, '$1\n');
+  return md.trim();
+}
+
+function stripBoilerplate(md) {
+  const kept = md
+    .split('\n')
+    .filter((line) => !BOILERPLATE.some((re) => re.test(line)));
+
+  // Drop a trailing "Resource Links" header left empty after stripping.
+  while (kept.length) {
+    const last = kept[kept.length - 1].trim();
+    if (last === '' || /^resource links?:?$/i.test(last)) kept.pop();
+    else break;
+  }
+  return kept.join('\n').replace(/\n{3,}/g, '\n\n').trim();
+}
+
+// --- Conservative guest extraction -----------------------------------------
+// Ported from scripts/update-podcast-authors.py. High-confidence only; a missed
+// guest degrades to no byline, a wrong guest would be a bad byline (ADR 0003).
+
+const NOT_A_PERSON = /^(?:The|A|An|In|This|For|With|From|On|PowerShell|Microsoft|Windows|Azure|DevOps|Cloud|GitHub|AWS|Episode|Podcast|Session|Roundtable|Community|Tonight|Join|Listen|Watch|Learn|Get|Set|Bar|Summit|Meet|Introducing|Featuring|Welcome|Just)\b/i;
+
+function looksLikePerson(name) {
+  if (!name || name.length < 4) return false;
+  if (NOT_A_PERSON.test(name)) return false;
+  const words = name.split(/\s+/);
+  if (words.length < 2) return false;
+  if (words.filter((w) => /^[A-ZÀ-ɏ]/.test(w)).length < 2) return false;
+  // Reject emphasis words (ALSO, PLUS) but allow initials like "B."
+  if (words.some((w) => w.length > 1 && w === w.toUpperCase() && !w.endsWith('.'))) return false;
+  const low = name.toLowerCase();
+  return !['http', 'www.', 'episode', 'podcast', 'powershell', 'microsoft', 'summit'].some((b) => low.includes(b));
+}
+
+function cleanName(raw) {
+  return raw
+    .replace(/\[([^\]]+)\]\([^)]+\)/g, '$1')
+    .replace(/<[^>]+>/g, '')
+    .replace(/[*_`]/g, '')
+    .replace(/^["']|["']$/g, '')
+    .replace(/[.,;:!?]+$/g, '')
+    .replace(/^(?:MVPs?\s+|Dr\.\s+|Prof\.\s+)/i, '')
+    .trim();
+}
+
+// "...with Name", "...with Name1 and Name2"
+function guestsFromTitle(title) {
+  const m = title.match(/\bwith\s+(.+)$/i);
+  if (!m) return [];
+  let after = m[1]
+    .replace(/^(?:MVPs?\s+|special\s+guest\s+host\s+)/i, '')
+    .replace(/\s*[-–—].*$/, '')
+    .replace(/!.*$/, '');
+  const out = [];
+  for (let part of after.split(/\s+(?:and|&)\s+/i)) {
+    part = part.replace(/[,:(\[].*$/, '').trim();
+    const words = [];
+    for (const w of part.split(/\s+/).slice(0, 3)) {
+      const cw = w.replace(/[!?.]+$/, '');
+      if (cw && /^[A-ZÀ-ɏ]/.test(cw)) words.push(cw);
+      else break;
+    }
+    if (words.length >= 2) {
+      const name = cleanName(words.join(' '));
+      if (looksLikePerson(name) && !out.includes(name)) out.push(name);
+    }
+  }
+  return out;
+}
+
+// "Guest Bio:" / "Bio:" header followed by "Name is/works/..." paragraph.
+function guestsFromBio(body) {
+  const out = [];
+  const hdr = body.match(/^(?:Guest Bio|Bio)\s*:?\s*$/im) || body.match(/(?:Guest Bio|Bio)\s*:/i);
+  if (!hdr) return out;
+  const after = body.slice(body.indexOf(hdr[0]) + hdr[0].length);
+  for (const para of after.split(/\n\s*\n/).slice(0, 2)) {
+    const m = para
+      .trim()
+      .match(/^(?:Dr\.\s+|Prof\.\s+)?([A-ZÀ-ɏ][\w'À-ɏ-]+(?:\s+[A-Z]\.)?(?:\s+[A-ZÀ-ɏ][\w'À-ɏ-]+){1,3})(?=\s+(?:is|was|are|has|works|joined|lives|serves|brings|currently|helps|leads|comes|spent|built|started|founded|created|focuses|specializes|manages|develops|writes|teaches|runs|hosts|holds|received|earned|wrote|published|blogs)\b|\s*,)/);
+    if (m) {
+      const name = cleanName(m[1]);
+      if (looksLikePerson(name) && !out.includes(name)) out.push(name);
+    }
+  }
+  return out;
+}
+
+function extractGuests(title, body) {
+  const guests = [];
+  for (const g of [...guestsFromBio(body), ...guestsFromTitle(title)]) {
+    if (!guests.includes(g) && g !== HOST) guests.push(g);
+  }
+  return guests;
+}
+
+// --- YAML ------------------------------------------------------------------
+
+// Plain scalar unless the value would break YAML (matches the existing files,
+// which leave titles with parens/apostrophes/& unquoted).
+function yamlScalar(v) {
+  const s = String(v);
+  if (s === '' || /[:#]\s/.test(s) || /[:#]$/.test(s) || /^[\s>|*&!%@`"'\[\]{},]/.test(s)) {
+    return `"${s.replace(/\\/g, '\\\\').replace(/"/g, '\\"')}"`;
+  }
+  return s;
+}
+
+// --- Existing-file index ---------------------------------------------------
+
+function buildIndex() {
+  const idx = { urls: new Set(), guids: new Set(), episodes: new Set() };
+  for (const file of fs.readdirSync(PODCAST_DIR)) {
+    if (!file.endsWith('.md') || file === '_index.md') continue;
+    const c = fs.readFileSync(path.join(PODCAST_DIR, file), 'utf-8');
+    if (!c.includes('mcdn.podbean.com')) continue; // modern (Podbean) episodes only
+    const url = c.match(/^podcast_url:\s*"?([^"\r\n]+)"?/m);
+    if (url) idx.urls.add(url[1].trim());
+    const guid = c.match(/^guid:\s*"?([^"\r\n]+)"?/m);
+    if (guid) idx.guids.add(guid[1].trim());
+    const ep = c.match(/^episode:\s*(\d+)/m);
+    if (ep) idx.episodes.add(parseInt(ep[1], 10));
+  }
+  return idx;
+}
+
+// --- Build one episode file ------------------------------------------------
+
+function buildEpisode(item) {
+  const enclosure = attr(item, 'enclosure', 'url').trim();
+  if (!enclosure) return null;
+
+  const rawTitle = tag(item, 'title').trim();
+  const guidRaw = tag(item, 'guid').trim();
+  const pubDate = tag(item, 'pubDate').trim();
+  const itunesEp = parseInt(tag(item, 'itunes:episode'), 10) || null;
+  const contentHtml = tag(item, 'content:encoded') || tag(item, 'description');
+
+  const episode = resolveEpisode(enclosure, itunesEp);
+  const { date, iso, y, m } = parseDate(pubDate);
+  const slug = TITLE_PREFIX.toLowerCase().trim().replace(/\s+/g, '-') + '-' + slugify(rawTitle);
+  const filename = `${date}-${slug}.md`;
+
+  const ytMatch = contentHtml.match(/youtu\.be\/([A-Za-z0-9_-]{6,})/);
+  const youtube = ytMatch ? ytMatch[1] : null;
+
+  const body = stripBoilerplate(htmlToMarkdown(contentHtml));
+  const guests = extractGuests(rawTitle, body);
+  const authors = [HOST, ...guests];
+
+  const fm = [];
+  fm.push(`title: ${yamlScalar(TITLE_PREFIX + rawTitle)}`);
+  fm.push(`author: ${yamlScalar(HOST)}`);
+  fm.push('authors:');
+  for (const a of authors) fm.push(`  - ${yamlScalar(a)}`);
+  fm.push(`date: "${iso}"`);
+  fm.push(`podcast_url: "${enclosure}"`);
+  if (episode != null) fm.push(`episode: ${episode}`);
+  if (youtube) fm.push(`youtube: ${youtube}`);
+  if (guidRaw) fm.push(`guid: ${yamlScalar(guidRaw)}`);
+  fm.push('aliases:');
+  fm.push(`  - /${y}/${m}/${slug}/`);
+
+  const content = `---\n${fm.join('\n')}\n---\n\n${body}\n`;
+  return { filename, content, episode, enclosure, guid: guidRaw, title: rawTitle };
+}
+
+// --- Main ------------------------------------------------------------------
+
+async function main() {
+  console.log(`Fetching ${FEED_URL} ...`);
+  const res = await fetch(FEED_URL);
+  if (!res.ok) throw new Error(`Feed fetch failed: ${res.status} ${res.statusText}`);
+  const xml = await res.text();
+
+  const items = xml.match(/<item>[\s\S]*?<\/item>/g) || [];
+  console.log(`Feed items: ${items.length}`);
+
+  const idx = buildIndex();
+  console.log(`Existing modern episodes indexed: urls=${idx.urls.size} guids=${idx.guids.size} episodes=${idx.episodes.size}`);
+
+  let added = 0;
+  let skipped = 0;
+  const writes = [];
+
+  for (const item of items) {
+    const ep = buildEpisode(item);
+    if (!ep) { skipped++; continue; }
+
+    const exists =
+      idx.urls.has(ep.enclosure) ||
+      (ep.guid && idx.guids.has(ep.guid)) ||
+      (ep.episode != null && idx.episodes.has(ep.episode));
+    if (exists) { skipped++; continue; }
+
+    const dest = path.join(PODCAST_DIR, ep.filename);
+    if (fs.existsSync(dest)) { skipped++; continue; }
+
+    writes.push(ep);
+    // Reserve keys so a duplicate item in the same run can't double-write.
+    idx.urls.add(ep.enclosure);
+    if (ep.guid) idx.guids.add(ep.guid);
+    if (ep.episode != null) idx.episodes.add(ep.episode);
+  }
+
+  writes.sort((a, b) => a.filename.localeCompare(b.filename));
+  for (const ep of writes) {
+    console.log(`${DRY_RUN ? '[dry-run] ' : ''}+ ${ep.filename} (episode ${ep.episode ?? '?'})`);
+    if (!DRY_RUN) fs.writeFileSync(path.join(PODCAST_DIR, ep.filename), ep.content, 'utf-8');
+    added++;
+  }
+
+  console.log(`\n${DRY_RUN ? 'Would add' : 'Added'}: ${added}   Skipped (already present): ${skipped}`);
+}
+
+main().catch((err) => {
+  console.error('Sync failed:', err);
+  process.exit(1);
+});