- Create pdf_text_to_html.js: JS port of the Ruby heuristic parser
(ALL_CAPS→h2, numbered headings→h3, bullets→ul/li, body→p dir=auto)
- Add pdf_view, text_view, document_view_options keys to i18n.js (en)
- Update document.vue: tab switcher shown when all pages have extracted
text; PDF View renders the existing page images; Text View renders
heuristic HTML in a prose container with per-page sections
- ArrowLeft/ArrowRight keyboard navigation between tabs with focus management
- Tab is hidden entirely for scanned/image-only PDFs (hasFullText gate)
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>