feat: per-extraction OCR language override

Allow users to specify a Tesseract language string (e.g. jpn+jpn_vert)
on a per-extraction basis, overriding the global OCR language setting.

- Add payload column to ai_jobs table (migration) to carry per-call data
- Thread ocrLanguages payload through enqueueJob → processNextJob → extractItemText
- New GET /api/ai-settings/ocr endpoint (requireAuth) returns { ocrMode, ocrLanguages }
- ImageLightbox fetches OCR settings and shows a language input next to the
  Extract Text button when mode is hybrid or tesseract (hidden for llm-only)
- MixedView fetches OCR settings and passes them down to EntryTile; kebab
  Extract Text on images shows an inline language prompt before dispatching the job

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

src/components/DoomScrollView.tsx

@@ -336,7 +336,7 @@ export default function DoomScrollView({ items, videoContext = 'mixed', onClose,
       {/* Text overlay */}
       {showTextOverlay && displayText && (
         <div
-          className="absolute bottom-16 left-4 right-4 z-20 rounded-xl p-4"
+          className="absolute bottom-4 left-4 right-4 z-20 rounded-xl p-4 max-w-fit"
           style={{ backgroundColor: 'rgba(0,0,0,0.75)' }}
           onClick={(e) => e.stopPropagation()}
         >