seo: drop tag archives from sitemap, add hreflang and BlogPosting schema

[foxis]

Three technical SEO fixes, each measured against a local build of Hugo 0.163.3 (the version pages.yml pins) before and after. No content changes — titles and copy are handled separately.

1. Tag archives were 66% of the sitemap

The EN sitemap carried 1392 URLs, of which 916 were tag term pages against 449 real content URLs. LT carried 346. On a site with ~93 referring domains, that spends most of Googlebot's crawl allowance re-reading near-empty archives instead of finding new posts.

blog/layouts/sitemap.xml now filters term and taxonomy pages belonging to the tags taxonomy. Categories (22) and series (5) are curated, so they stay listed.

Tag pages remain crawlable and indexable — they are only delisted from the sitemap, so any tag page that already earns traffic keeps it. (/tags/diffraction-grating/ currently ranks, so a blanket noindex was deliberately not used.)

Sitemap	Before	After
EN total	1392	476
— tag pages	916	0
— content	449	449
— categories	22	22
— series	5	5
LT total	346	99

Verified no /tags/ URL leaks through. The xhtml:link hreflang alternates that Hugo's embedded template emits are preserved deliberately — they were the site's only hreflang signal, and dropping them while rewriting the template would have been a silent regression.

2. No hreflang in the HTML head

Every post ships EN and LT, but alternates were declared only inside the sitemap. Head tags are what Google treats as authoritative for a page reached by internal link, backlink or share.

Added reciprocal alternates plus x-default, with the default language resolved from hugo.Sites.Default rather than hardcoded to en. Verified both language versions emit an identical, reciprocal set:

<link rel="alternate" hreflang="en" href="https://itohi.com/fpv/edgetx-cockpit-voice-rebuild/">
<link rel="alternate" hreflang="x-default" href="https://itohi.com/fpv/edgetx-cockpit-voice-rebuild/">
<link rel="alternate" hreflang="lt" href="https://itohi.com/lt/fpv/edgetx-cockpit-voice-rebuild/">

3. No Article schema anywhere

Every URL on the site emitted only Person + WebSite. No page ever declared a headline, publish date, modified date, author or section — Google had to infer all of it from rendered HTML. That also left the site ineligible for article rich results and gave no explicit freshness signal for a series that gets revised.

Added BlogPosting for article pages with datePublished, dateModified, wordCount, articleSection, keywords, and series as CreativeWorkSeries. Excluded from taxonomy pages (verified: tag pages emit no BlogPosting).

Built with dict + jsonify rather than hand-written JSON so quotes and non-ASCII escape correctly instead of producing invalid JSON-LD. Verified the LT output parses and Lithuanian diacritics land as valid \u escapes.

Notes

• Used .Language.Locale instead of the .Language.LanguageCode the embedded template uses, since that was deprecated in Hugo 0.158. Build is warning-clean; before this change the build emitted three deprecation warnings.
• Page counts unchanged (EN 2403 / LT 632), so nothing was dropped from the build itself.
• Expect Search Console "Discovered — currently not indexed" to fall and crawl requests to shift toward posts over the following few weeks. Indexed tag pages stay indexed; this changes crawl priority, not the index.

[cursor]

Bugbot is not enabled for your account, so this pull request was not reviewed.

Enable Bugbot in the Cursor dashboard to get automatic reviews on future PRs.