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/lib/ai-jobs.ts

@@ -34,6 +34,7 @@ interface AiJobRow {
   started_at: number | null
   completed_at: number | null
   item_title: string | null
+  payload: string | null
 }
 
 function rowToJob(row: AiJobRow): AiJob {
@@ -75,6 +76,7 @@ export function enqueueJob(
   jobType: AiJobType,
   libraryId: string,
   sourceLanguage?: string,
+  payload?: Record<string, string>,
 ): string {
   const db = getDb()
 
@@ -96,9 +98,9 @@ export function enqueueJob(
   const metadata = jobType === 'translate' && sourceLanguage ? sourceLanguage : null
 
   db.prepare(
-    `INSERT INTO ai_jobs (id, item_key, library_id, job_type, status, error, attempt, max_retries, created_at, item_title)
-     VALUES (?, ?, ?, ?, 'queued', ?, 0, ?, ?, ?)`
-  ).run(id, itemKey, libraryId, jobType, metadata, maxRetries, Date.now(), title)
+    `INSERT INTO ai_jobs (id, item_key, library_id, job_type, status, error, attempt, max_retries, created_at, item_title, payload)
+     VALUES (?, ?, ?, ?, 'queued', ?, 0, ?, ?, ?, ?)`
+  ).run(id, itemKey, libraryId, jobType, metadata, maxRetries, Date.now(), title, payload ? JSON.stringify(payload) : null)
 
   // Wake the processor
   wakeProcessor()
@@ -251,6 +253,8 @@ async function processNextJob(): Promise<boolean> {
 
   // Extract sourceLanguage for translate jobs (stored in error field at enqueue)
   const sourceLanguage = row.job_type === 'translate' ? row.error : null
+  // Parse job payload (carries per-call overrides, e.g. ocrLanguages for extract jobs)
+  const jobPayload = row.payload ? (JSON.parse(row.payload) as Record<string, string>) : null
 
   db.prepare(
     "UPDATE ai_jobs SET status = 'running', started_at = ?, error = NULL WHERE id = ?"
@@ -265,7 +269,7 @@ async function processNextJob(): Promise<boolean> {
         await generateItemDescription(row.item_key)
         break
       case 'extract':
-        await extractItemText(row.item_key)
+        await extractItemText(row.item_key, jobPayload?.ocrLanguages)
         break
       case 'translate':
         await translateItemText(row.item_key, sourceLanguage || undefined)