Add RSS feed loader for episode page generation

tests/fixtures/feed_sample.xml

@@ -0,0 +1,5 @@
+<?xml version="1.0" encoding="utf-8"?>
+<rss xmlns:itunes="http://www.itunes.com/dtds/podcast-1.0.dtd" xmlns:podcast="https://podcastindex.org/namespace/1.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:content="http://purl.org/rss/1.0/modules/content/" version="2.0"><channel><atom:link href="https://podcast.macneilmediagroup.com/@LukeAtTheRoost/feed.xml" rel="self" type="application/rss+xml"></atom:link><atom:link href="https://pubsubhubbub.appspot.com/" rel="hub" type="application/rss+xml"></atom:link><atom:link href="https://pubsubhubbub.superfeedr.com/" rel="hub" type="application/rss+xml"></atom:link><atom:link href="https://websubhub.com/hub" rel="hub" type="application/rss+xml"></atom:link><atom:link href="https://switchboard.p3k.io/" rel="hub" type="application/rss+xml"></atom:link><lastBuildDate>Tue, 04 Aug 2026 09:23:29 +0000</lastBuildDate><generator>Castopod - https://castopod.org/</generator><docs>https://cyber.harvard.edu/rss/rss.html</docs><podcast:guid>46cfb731-b4a1-55e2-80f3-cadb1c6c18fa</podcast:guid><title>Luke at the Roost</title><description><![CDATA[<p>A late-night call-in radio show broadcast from a desert hermit’s RV, featuring a mix of real callers and AI-generated callers talking to Luke about life, love, and everything in between. Call in live: 208-439-LUKE (208-439-5853).</p>
+<p>Website: <a href="https://lukeattheroost.com">https://lukeattheroost.com</a></p>
+]]></description><podcast:medium>podcast</podcast:medium><itunes:image href="https://podcast.macneilmediagroup.com/media/podcasts/LukeAtTheRoost/cover_feed.png"></itunes:image><language>en</language><podcast:location geo="geo:50.8050519,-0.4525129" osm="W996958777">Smugglers&amp;#039; Roost</podcast:location><podcast:locked owner="luke@macneilmediagroup.com">no</podcast:locked><podcast:social priority="1" platform="castopod" protocol="activitypub" accountId="@LukeAtTheRoost@podcast.macneilmediagroup.com" accountUrl="https://podcast.macneilmediagroup.com/@LukeAtTheRoost"></podcast:social><itunes:category text="Comedy"></itunes:category><category>Comedy</category><itunes:explicit>true</itunes:explicit><itunes:author>MacNeil Media Group, LLC</itunes:author><link>https://podcast.macneilmediagroup.com/@LukeAtTheRoost</link><itunes:owner><itunes:name>Luke MacNeil</itunes:name><itunes:email>luke@macneilmediagroup.com</itunes:email></itunes:owner><itunes:type>episodic</itunes:type><copyright>MacNeil Media Group, LLC</copyright><image><url>https://podcast.macneilmediagroup.com/media/podcasts/LukeAtTheRoost/cover_feed.png</url><title>Luke at the Roost</title><link>https://podcast.macneilmediagroup.com/@LukeAtTheRoost</link></image><item><title>Episode 58: Rayfield's Nephew, the Marfa Lights, and Why Nobody Believes Concho</title><enclosure url="https://op3.dev/e,pg=46cfb731-b4a1-55e2-80f3-cadb1c6c18fa/podcast.macneilmediagroup.com/audio/@LukeAtTheRoost/episode-58-rayfield-s-nephew-the-marfa-lights-and-why-nobody-believes-concho.mp3" length="114501589" type="audio/mpeg"></enclosure><guid>https://podcast.macneilmediagroup.com/@LukeAtTheRoost/episodes/episode-58-rayfield-s-nephew-the-marfa-lights-and-why-nobody-believes-concho</guid><pubDate>Tue, 04 Aug 2026 09:22:38 +0000</pubDate><description><![CDATA[<p>Luke fields calls about stolen catalytic converters, unexplained phenomena over Mitchell Flat, a sheriff's deputy's moral dilemma, and a fence builder convinced someone is mapping West Texas water sources. Plus: Donald Judd's aluminum boxes change everything for a skeptical rancher named Fern.</p>]]></description><itunes:duration>4770</itunes:duration><link>https://podcast.macneilmediagroup.com/@LukeAtTheRoost/episodes/episode-58-rayfield-s-nephew-the-marfa-lights-and-why-nobody-believes-concho</link><itunes:image href="https://podcast.macneilmediagroup.com/media/podcasts/LukeAtTheRoost/cover_feed.png"></itunes:image><itunes:explicit>true</itunes:explicit><itunes:episode>58</itunes:episode><itunes:episodeType>full</itunes:episodeType><podcast:episode>58</podcast:episode><podcast:comments uri="https://podcast.macneilmediagroup.com/@LukeAtTheRoost/episodes/episode-58-rayfield-s-nephew-the-marfa-lights-and-why-nobody-believes-concho/comments" contentType="application/podcast-activity+json"></podcast:comments><podcast:transcript url="https://podcast.macneilmediagroup.com/media/podcasts/LukeAtTheRoost/episode-58-rayfield-s-nephew-the-marfa-lights-and-why-nobody-believes-concho.srt" type="application/x-subrip" rel="captions" language="en"></podcast:transcript><podcast:chapters url="https://podcast.macneilmediagroup.com/media/podcasts/LukeAtTheRoost/episode-58-rayfield-s-nephew-the-marfa-lights-and-why-nobody-believes-concho-chapters.json" type="application/json+chapters"></podcast:chapters></item><item><title>Episode 57: Trace's Box of Family Secrets</title><enclosure url="https://op3.dev/e,pg=46cfb731-b4a1-55e2-80f3-cadb1c6c18fa/podcast.macneilmediagroup.com/audio/@LukeAtTheRoost/episode-57-trace-s-box-of-family-secrets.mp3" length="97583646" type="audio/mpeg"></enclosure><guid>https://podcast.macneilmediagroup.com/@LukeAtTheRoost/episodes/episode-57-trace-s-box-of-family-secrets</guid><pubDate>Tue, 02 Jun 2026 11:32:58 +0000</pubDate><description><![CDATA[<p>Luke hosts another late-night call-in show from Alpine, Texas, featuring intense conversations about personal dilemmas. Callers include Silas struggling with a community move, a landman wrestling with an ethical choice, and Trace discovering hidden letters with a potentially life-changing secret. What happens when the past suddenly becomes present?</p>
+]]></description><itunes:duration>4066</itunes:duration><link>https://podcast.macneilmediagroup.com/@LukeAtTheRoost/episodes/episode-57-trace-s-box-of-family-secrets</link><itunes:image href="https://podcast.macneilmediagroup.com/media/podcasts/LukeAtTheRoost/cover_feed.png"></itunes:image><itunes:explicit>true</itunes:explicit><itunes:episode>57</itunes:episode><itunes:episodeType>full</itunes:episodeType><podcast:episode>57</podcast:episode><podcast:comments uri="https://podcast.macneilmediagroup.com/@LukeAtTheRoost/episodes/episode-57-trace-s-box-of-family-secrets/comments" contentType="application/podcast-activity+json"></podcast:comments><podcast:socialInteract uri="https://podcast.macneilmediagroup.com/@LukeAtTheRoost/posts/418bda4c-56df-47c3-ae29-6e3eb4b70a50" priority="1" platform="castopod" protocol="activitypub" accountId="@LukeAtTheRoost@podcast.macneilmediagroup.com" pubDate="2026-06-02T11:32:58+0000"></podcast:socialInteract><podcast:transcript url="https://podcast.macneilmediagroup.com/media/podcasts/LukeAtTheRoost/episode-57-trace-s-box-of-family-secrets.srt" type="application/x-subrip" rel="captions" language="en"></podcast:transcript><podcast:chapters url="https://podcast.macneilmediagroup.com/media/podcasts/LukeAtTheRoost/episode-57-trace-s-box-of-family-secrets-chapters.json" type="application/json+chapters"></podcast:chapters></item></channel></rss>

tests/test_feed_loader.py

@@ -0,0 +1,42 @@
+from pathlib import Path
+
+import pytest
+
+from website_gen.feed import parse_feed
+
+FIXTURE = Path(__file__).parent / "fixtures" / "feed_sample.xml"
+
+
+@pytest.fixture
+def feed_xml():
+    return FIXTURE.read_text(encoding="utf-8")
+
+
+def test_parses_core_fields(feed_xml):
+    eps = parse_feed(feed_xml)
+    ep = next(e for e in eps if e.number == 58)
+    assert ep.slug == "episode-58-rayfield-s-nephew-the-marfa-lights-and-why-nobody-believes-concho"
+    assert ep.title.startswith("Episode 58: Rayfield's Nephew")
+    assert ep.duration_seconds == 4770
+    assert ep.audio_url.startswith("https://")
+
+
+def test_description_is_stripped_of_cdata_and_html(feed_xml):
+    ep = parse_feed(feed_xml)[0]
+    assert "<![CDATA[" not in ep.description
+    assert "<p>" not in ep.description
+
+
+def test_slug_comes_from_link_and_drops_trailing_slash(feed_xml):
+    for ep in parse_feed(feed_xml):
+        assert not ep.slug.endswith("/")
+        assert "/" not in ep.slug
+
+
+def test_pubdate_parses_to_iso_date(feed_xml):
+    ep = next(e for e in parse_feed(feed_xml) if e.number == 58)
+    assert ep.published_iso.startswith("2026-08-04")
+
+
+def test_returns_all_items_in_fixture(feed_xml):
+    assert len(parse_feed(feed_xml)) == 2

website_gen/feed.py

@@ -0,0 +1,75 @@
+import html
+import re
+import xml.etree.ElementTree as ET
+from dataclasses import dataclass
+from email.utils import parsedate_to_datetime
+
+ITUNES_NS = {"itunes": "http://www.itunes.com/dtds/podcast-1.0.dtd"}
+
+TAG_RE = re.compile(r"<[^>]+>")
+WHITESPACE_RE = re.compile(r"\s+")
+
+
+@dataclass
+class Episode:
+    number: int | None
+    slug: str
+    title: str
+    description: str
+    published_iso: str
+    duration_seconds: int | None
+    audio_url: str
+
+
+def parse_feed(xml_text: str) -> list[Episode]:
+    """Turn RSS feed XML into Episode records, skipping items with no usable link."""
+    channel = ET.fromstring(xml_text).find("channel")
+    if channel is None:
+        return []
+
+    episodes: list[Episode] = []
+    for item in channel.findall("item"):
+        slug = _slug_from_link(item.findtext("link"))
+        if not slug:
+            continue
+        enclosure = item.find("enclosure")
+        episodes.append(
+            Episode(
+                number=_as_int(item.findtext("itunes:episode", namespaces=ITUNES_NS)),
+                slug=slug,
+                title=(item.findtext("title") or "").strip(),
+                description=_plain_text(item.findtext("description")),
+                published_iso=_iso_date(item.findtext("pubDate")),
+                duration_seconds=_as_int(item.findtext("itunes:duration", namespaces=ITUNES_NS)),
+                audio_url=(enclosure.get("url") or "") if enclosure is not None else "",
+            )
+        )
+    return episodes
+
+
+def _slug_from_link(link: str | None) -> str:
+    slug = (link or "").strip().rstrip("/")
+    if "/episodes/" not in slug:
+        return ""
+    return slug.rsplit("/episodes/", 1)[-1]
+
+
+def _plain_text(raw: str | None) -> str:
+    text = TAG_RE.sub(" ", raw or "")
+    return WHITESPACE_RE.sub(" ", html.unescape(text)).strip()
+
+
+def _iso_date(raw: str | None) -> str:
+    if not raw:
+        return ""
+    try:
+        return parsedate_to_datetime(raw).isoformat()
+    except (TypeError, ValueError):
+        return ""
+
+
+def _as_int(raw: str | None) -> int | None:
+    try:
+        return int((raw or "").strip())
+    except ValueError:
+        return None