Sync bundled man page catalog from man.sr.ht (#7)

The More tab's official man-page list was a hardcoded array, updated by
hand. Move it to a checked-in Hutch/Resources/man-pages.json (bundled via
the synchronized group) loaded by a new ManPageCatalog, with the previous
list kept as a built-in fallback.

Add scripts/sync_man_pages.py, which re-derives the catalog from the
man.sr.ht landing page, and a weekly scheduled workflow that runs it and
opens a pull request when the list diverges upstream. The script refuses to
write a suspiciously short list so a markup change can't gut the catalog.

This first sync also picks up chat.sr.ht, which upstream added since the
list was last hand-edited. Tests cover catalog loading, the fallback, and
JSON decoding.

1 files changed, 63 insertions, 0 deletions

diff --git a/scripts/sync_man_pages.py b/scripts/sync_man_pages.py
new file mode 100755
index 0000000..c59197a
--- /dev/null
+++ b/scripts/sync_man_pages.py
@@ -0,0 +1,63 @@
+#!/usr/bin/env python3
+"""Regenerate the bundled sr.ht man page catalog from man.sr.ht.
+
+Fetches the man.sr.ht landing page, extracts the per-service "User Manual"
+links, and writes them to ``Hutch/Resources/man-pages.json``. The scheduled
+GitHub workflow that runs this opens a pull request whenever the result differs
+from the committed copy, so the in-app list stays in sync with upstream without
+hand edits.
+
+Run locally with ``python3 scripts/sync_man_pages.py``; exits non-zero (without
+writing) if upstream markup changed enough that too few pages were found, so a
+bad scrape can never wipe the bundled list.
+"""
+import json
+import re
+import sys
+import urllib.request
+from pathlib import Path
+
+INDEX_URL = "https://man.sr.ht/"
+OUTPUT = Path(__file__).resolve().parent.parent / "Hutch" / "Resources" / "man-pages.json"
+# The suite has ~12 service manuals; a scrape returning far fewer means the page
+# structure changed and we should fail loudly rather than commit a gutted list.
+MINIMUM_EXPECTED = 8
+
+
+def fetch(url: str) -> str:
+    request = urllib.request.Request(url, headers={"User-Agent": "hutch-man-page-sync"})
+    with urllib.request.urlopen(request, timeout=30) as response:
+        return response.read().decode("utf-8")
+
+
+def build_catalog(html: str) -> list[dict[str, str]]:
+    """Extract official man-page links, deduplicated and sorted by title."""
+    entries: dict[str, str] = {}
+    for href in re.findall(r'href="([^"]+)"', html):
+        service = re.fullmatch(r"/([a-z0-9][a-z0-9.-]*\.sr\.ht)/?", href)
+        if service:
+            title = service.group(1)
+            entries[title] = f"https://man.sr.ht/{title}/"
+        elif re.fullmatch(r"sr\.ht/?", href):
+            entries["sr.ht"] = "https://man.sr.ht/sr.ht/"
+        elif re.fullmatch(r"https://srht\.site/?", href):
+            entries["srht.site"] = "https://srht.site/"
+    return [{"title": title, "url": entries[title]} for title in sorted(entries)]
+
+
+def main() -> int:
+    catalog = build_catalog(fetch(INDEX_URL))
+    if len(catalog) < MINIMUM_EXPECTED:
+        print(
+            f"Refusing to write catalog with only {len(catalog)} entries; "
+            "man.sr.ht markup may have changed.",
+            file=sys.stderr,
+        )
+        return 1
+    OUTPUT.write_text(json.dumps(catalog, indent=2, ensure_ascii=False) + "\n")
+    print(f"Wrote {len(catalog)} man page(s) to {OUTPUT}")
+    return 0
+
+
+if __name__ == "__main__":
+    sys.exit(main())