Sync Website from LaTeX CV via _pages/cv.md

Summary

• Make _pages/cv.md the only CV page on the website and remove the dormant cv-json path and its supporting code.
• Use {resume-main} as the source of truth, with split ownership:
  • resume-main/automation/ owns publication parsing from publication.tex, publication.bib, and patent.bib.
  • homepage-test/scripts/ owns the final site sync into markdown pages and collections.
• Keep the current site structure and public scope: update existing pages only (/, /research/, /talks/, /teaching/, /projects/, /cv/).

Local Check

1. python .\scripts\sync_from_resume.py --resume-repo ..\resume-main
2. bundle exec jekyll build --config _config.yml,_config_local.yml
3. scripts\serve.bat
4. Visit http://localhost:5000/, /research/, /talks/, /teaching/, /projects/, and /cv/