feat(seo): OG/Twitter tags, sitemap.xml, and Dataset structured data enhancements

[solrevdev]

Context

This PR continues a series of SEO improvements to the winget-search static GitHub Pages site at https://solrevdev.com/winget-search/.

Google Search Console status at time of PR

Both missing-field issues are now in validation started state in Search Console (Google is re-crawling to confirm the fixes from the previous commits):

• ✅ Missing field 'license' — validation started (fix landed in 323a3b6)
• ✅ Missing field 'creator' — validation started (fix landed in 323a3b6)

The screenshot below shows the Data sets report in Google Search Console confirming validation has started for both issues:

Google Search Console → solrevdev.com → Enhancements → Data sets

• 0 Invalid, No critical issues
• 1 Valid
• "Improve item appearance": Missing field 'license' (Started) + Missing field 'creator' (Started)

These were non-critical warnings. The fixes have been live since 323a3b6 and the CI daily run at 02:00 UTC auto-stamps dateModified so Google always sees a fresh signal.

---

What this PR adds

1. Open Graph meta tags (index.html)

Adds og:type, og:site_name, og:title, og:description, og:url — improves how the page renders when shared on social platforms and is a general crawl signal for search engines.

2. Twitter Card meta tags (index.html)

Adds twitter:card (summary), twitter:site / twitter:creator (@solrevdev), title, description.

3. Canonical link + sitemap link (index.html)

• <link rel="canonical" href="https://solrevdev.com/winget-search/"> — prevents duplicate content issues.
• <link rel="sitemap" type="application/xml" href="sitemap.xml"> — advertises the sitemap in-page.

4. sitemap.xml (new file)

Covers both pages:

• https://solrevdev.com/winget-search/ — changefreq: daily, priority: 1.0
• https://solrevdev.com/winget-search/agent-access.html — changefreq: weekly, priority: 0.5

CI stamps <lastmod> on every build so it stays current.

Next step after merge: submit https://solrevdev.com/winget-search/sitemap.xml in Google Search Console → Sitemaps.

5. Dataset JSON-LD enhancements (index.html)

Three new fields added to the existing @type: Dataset block:

Field	Value	Why
sameAs	https://github.com/solrevdev/winget-search	Links the canonical GitHub source for the project
numberOfItems	0 (placeholder; CI stamps actual count)	Concrete size signal for Google Dataset Search
variableMeasured	["packageId", "name", "version", "publisher"]	Describes the fields in packages.json for richer Dataset results

6. CI workflow: numberOfItems + sitemap lastmod stamping

The build step now:

# Stamp numberOfItems with actual package count
PACKAGE_COUNT=$(jq '.metadata.total // length' packages.json)
sed -i "s/\"numberOfItems\": [0-9]*/\"numberOfItems\": ${PACKAGE_COUNT}/" deploy/index.html

# Stamp sitemap lastmod
sed -i "s/<lastmod>[^<]*<\/lastmod>/<lastmod>$(date -u +%Y-%m-%d)<\/lastmod>/g" deploy/sitemap.xml

---

Full Dataset JSON-LD state after merge

{
  "@context": "https://schema.org",
  "@type": "Dataset",
  "name": "Winget Package Search catalog",
  "description": "Current package metadata extracted from microsoft/winget-pkgs.",
  "url": "https://solrevdev.com/winget-search/",
  "isBasedOn": "https://github.com/microsoft/winget-pkgs",
  "keywords": ["winget", "Windows Package Manager", "package search", "software catalog", "Windows software", "package manager"],
  "inLanguage": "en",
  "temporalCoverage": "2025-05-29/..",
  "datePublished": "2025-05-29",
  "dateModified": "<stamped-by-CI>",
  "creator": { "@type": "Person", "name": "John Smith", "url": "https://solrevdev.com/about/" },
  "license": "https://spdx.org/licenses/MIT.html",
  "sameAs": "https://github.com/solrevdev/winget-search",
  "numberOfItems": "<stamped-by-CI>",
  "variableMeasured": ["packageId", "name", "version", "publisher"],
  "distribution": {
    "@type": "DataDownload",
    "encodingFormat": "application/json",
    "contentUrl": "https://solrevdev.com/winget-search/packages.json"
  }
}

---

Remaining SEO opportunities (future work)

• og:image — a social preview image would complete the OG card
• Submit sitemap.xml to Google Search Console manually (Sitemaps tab)
• robots.txt with Sitemap: directive (none exists yet)
• Consider a citation field in the Dataset block if the data is referenced externally

---

Test plan

• Merge and confirm CI build passes + pages-build-deployment succeeds
• Verify live structured data: curl -s https://solrevdev.com/winget-search/ | grep -A5 'numberOfItems'
• Verify numberOfItems is non-zero (CI-stamped from packages.json count)
• Verify sitemap at https://solrevdev.com/winget-search/sitemap.xml has today's lastmod
• Check OG tags render correctly with https://developers.facebook.com/tools/debug/
• Check Dataset structured data with https://search.google.com/test/rich-results
• Submit sitemap.xml in Google Search Console → Sitemaps

🤖 Generated with Claude Code

[solrevdev]

Implemented the og:image work in commit 50931e2 (feat(seo): add social preview image).

Preview from this PR branch:

What changed:

• Added og-image.svg as the editable 1200x630 source asset.
• Generated og-image.png as the social-scraper target. PNG is used for og:image because it is broadly supported by social platforms.
• Updated index.html with og:image, dimensions, image type, alt text, and Twitter summary_large_image metadata.
• Updated the GitHub Pages build workflow to publish both og-image.png and og-image.svg.

Validation performed locally:

• xmllint --noout og-image.svg passes.
• rsvg-convert --width 1200 --height 630 --output og-image.png og-image.svg generated the PNG.
• file og-image.png reports PNG image data, 1200 x 630.
• Local static serving returned HTTP 200 with Content-type: image/png for /og-image.png.
• Playwright CLI opened the image successfully as og-image.png (1200×630).

After merge and Pages deployment, the expected live image URL is:
https://solrevdev.com/winget-search/og-image.png

Suggested post-deploy checks:

• Open https://solrevdev.com/winget-search/og-image.png directly.
• Re-scrape https://solrevdev.com/winget-search/ in the Facebook Sharing Debugger and any Twitter/X card validator available to confirm the card refreshes.

[solrevdev]

Follow-up fix: the first preview URL rendered as a Git LFS pointer because this repo tracks *.png through LFS by default.

Fixed in commit 949eec2 (fix(seo): store og image outside lfs):

• Added a narrow .gitattributes exception for /og-image.png.
• Re-added og-image.png as a normal Git blob, not an LFS pointer.

Verified the same raw URL now serves the image correctly:

• content-type: image/png
• content-length: 94651
• first bytes are the PNG header (89504e47...)

Preview should now render here: