Merge pull request 'image-viewer-improvements' (#27) from image-viewer-improvements into main

Reviewed-on: http://gitea.lan/gpatti/MediaLore/pulls/27

src/app/api/ai-tagging/extract-text/route.ts

@@ -3,14 +3,14 @@ import { requireLibraryAccess } from '@/lib/auth'
 import { enqueueJob } from '@/lib/ai-jobs'
 
 export async function POST(request: NextRequest) {
-  let body: { itemKey?: string; ocrLanguages?: string }
+  let body: { itemKey?: string; ocrLanguages?: string; ocrMode?: string }
   try {
     body = await request.json()
   } catch {
     return NextResponse.json({ error: 'Invalid JSON body' }, { status: 400 })
   }
 
-  const { itemKey, ocrLanguages } = body
+  const { itemKey, ocrLanguages, ocrMode } = body
   if (!itemKey || typeof itemKey !== 'string') {
     return NextResponse.json({ error: 'itemKey is required' }, { status: 400 })
   }
@@ -19,12 +19,15 @@ export async function POST(request: NextRequest) {
   const auth = await requireLibraryAccess(request, libraryId)
   if (auth instanceof NextResponse) return auth
 
+  const payload: Record<string, string> = {}
+  if (ocrLanguages) payload.ocrLanguages = ocrLanguages
+  if (ocrMode) payload.ocrMode = ocrMode
   const jobId = enqueueJob(
     itemKey,
     'extract',
     libraryId,
     undefined,
-    ocrLanguages ? { ocrLanguages } : undefined,
+    Object.keys(payload).length ? payload : undefined,
   )
   return NextResponse.json({ jobId }, { status: 202 })
 }

src/app/api/ai-tagging/fields/route.ts

@@ -1,6 +1,6 @@
 import { NextRequest, NextResponse } from 'next/server'
 import { requireLibraryAccess } from '@/lib/auth'
-import { getAiFields, updateExtractedText } from '@/lib/ai-tagger'
+import { getAiFields, updateExtractedText, updateAiDescription } from '@/lib/ai-tagger'
 
 export async function GET(request: NextRequest) {
   const { searchParams } = request.nextUrl
@@ -19,25 +19,37 @@ export async function GET(request: NextRequest) {
 }
 
 export async function PATCH(request: NextRequest) {
-  let body: { itemKey?: string; extractedText?: string }
+  let body: { itemKey?: string; extractedText?: string; aiDescription?: string }
   try {
     body = await request.json()
   } catch {
     return NextResponse.json({ error: 'Invalid JSON body' }, { status: 400 })
   }
 
-  const { itemKey, extractedText } = body
+  const { itemKey, extractedText, aiDescription } = body
   if (!itemKey || typeof itemKey !== 'string') {
     return NextResponse.json({ error: 'itemKey is required' }, { status: 400 })
   }
-  if (typeof extractedText !== 'string') {
-    return NextResponse.json({ error: 'extractedText is required' }, { status: 400 })
+  if (extractedText === undefined && aiDescription === undefined) {
+    return NextResponse.json({ error: 'extractedText or aiDescription is required' }, { status: 400 })
   }
 
   const libraryId = itemKey.split(':')[0]
   const auth = await requireLibraryAccess(request, libraryId)
   if (auth instanceof NextResponse) return auth
 
-  updateExtractedText(itemKey, extractedText)
+  if (extractedText !== undefined) {
+    if (typeof extractedText !== 'string') {
+      return NextResponse.json({ error: 'extractedText must be a string' }, { status: 400 })
+    }
+    updateExtractedText(itemKey, extractedText)
+  }
+  if (aiDescription !== undefined) {
+    if (typeof aiDescription !== 'string') {
+      return NextResponse.json({ error: 'aiDescription must be a string' }, { status: 400 })
+    }
+    updateAiDescription(itemKey, aiDescription)
+  }
+
   return NextResponse.json({ ok: true })
 }

src/app/api/ai-tagging/translate-bulk/route.ts

@@ -0,0 +1,36 @@
+import { NextRequest, NextResponse } from 'next/server'
+import { requireLibraryAccess } from '@/lib/auth'
+import { enqueueJob } from '@/lib/ai-jobs'
+import { getDb } from '@/lib/db'
+
+export async function POST(request: NextRequest) {
+  let body: { libraryId?: string; path?: string }
+  try {
+    body = await request.json()
+  } catch {
+    return NextResponse.json({ error: 'Invalid JSON body' }, { status: 400 })
+  }
+
+  const { libraryId, path: dirPath } = body
+  if (!libraryId || typeof libraryId !== 'string') {
+    return NextResponse.json({ error: 'libraryId is required' }, { status: 400 })
+  }
+
+  const auth = await requireLibraryAccess(request, libraryId)
+  if (auth instanceof NextResponse) return auth
+
+  const db = getDb()
+  const prefix = dirPath
+    ? `${libraryId}:mixed_file:${encodeURIComponent(dirPath + '/')}`
+    : `${libraryId}:mixed_file:`
+
+  // Only enqueue translate jobs for items that already have extracted text
+  const items = db
+    .prepare(
+      'SELECT item_key FROM media_items WHERE item_key LIKE ? AND item_type = ? AND extracted_text IS NOT NULL'
+    )
+    .all(`${prefix}%`, 'mixed_file') as { item_key: string }[]
+
+  const jobIds = items.map(({ item_key }) => enqueueJob(item_key, 'translate', libraryId))
+  return NextResponse.json({ jobIds, queued: jobIds.length }, { status: 202 })
+}

src/app/api/browse/route.ts

@@ -33,18 +33,37 @@ export async function GET(request: NextRequest) {
     ? scanDirectoryRecursive(root, libraryId, subpath)
     : scanDirectory(root, libraryId, subpath)
 
-  // Annotate image entries with whether they have extracted text
+  // Annotate image files with hasExtractedText, and directories if any descendant has extracted text
   const db = getDb()
   const rows = db
     .prepare('SELECT item_key FROM media_items WHERE library_id = ? AND extracted_text IS NOT NULL')
     .all(libraryId) as { item_key: string }[]
   const withText = new Set(rows.map((r) => r.item_key))
 
+  // Build a set of all ancestor directory relative paths that contain at least one item with text
+  // e.g. item_key "lib:mixed_file:manga%2Fch1%2Fp1.jpg" → ancestors "manga", "manga/ch1"
+  const dirsWithText = new Set<string>()
+  const keyPrefix = `${libraryId}:mixed_file:`
+  for (const key of withText) {
+    const decoded = decodeURIComponent(key.slice(keyPrefix.length))
+    const parts = decoded.split('/')
+    for (let i = 1; i < parts.length; i++) {
+      dirsWithText.add(parts.slice(0, i).join('/'))
+    }
+  }
+
   listing.entries = listing.entries.map((e) => {
-    if (e.type !== 'file' || e.mediaType !== 'image') return e
-    const relPath = subpath ? path.join(subpath, e.name) : e.name
-    const itemKey = `${libraryId}:mixed_file:${encodeURIComponent(relPath)}`
-    return { ...e, hasExtractedText: withText.has(itemKey) }
+    if (e.type === 'file') {
+      if (e.mediaType !== 'image') return e
+      const relPath = subpath ? path.join(subpath, e.name) : e.name
+      const itemKey = `${libraryId}:mixed_file:${encodeURIComponent(relPath)}`
+      return { ...e, hasExtractedText: withText.has(itemKey) }
+    }
+    if (e.type === 'directory') {
+      const dirRel = subpath ? `${subpath}/${e.name}` : e.name
+      if (dirsWithText.has(dirRel)) return { ...e, hasExtractedText: true }
+    }
+    return e
   })
 
   return NextResponse.json(listing)

src/components/mixed/MixedView.tsx

@@ -27,6 +27,7 @@ export default function MixedView({ libraryId, initialPath }: Props) {
   const [loading, setLoading] = useState(true)
   const [error, setError] = useState<string | null>(null)
   const [modal, setModal] = useState<ModalState>(null)
+  const [modalShowTags, setModalShowTags] = useState(false)
   const [tagPanel, setTagPanel] = useState<TagPanelState>(null)
   const [search, setSearch] = useState('')
   const [selectedTagIds, setSelectedTagIds] = useState<Set<string>>(new Set())
@@ -469,15 +470,28 @@ export default function MixedView({ libraryId, initialPath }: Props) {
                     }
                   }}
                   onTranslate={async (e) => {
-                    const itemKey = itemKeyFor(e)
-                    const res = await fetch('/api/ai-tagging/translate', {
-                      method: 'POST',
-                      headers: { 'Content-Type': 'application/json' },
-                      body: JSON.stringify({ itemKey }),
-                    })
-                    if (!res.ok) {
-                      const data = await res.json().catch(() => ({}))
-                      throw new Error((data as { error?: string }).error ?? 'Translation failed')
+                    if (e.type === 'directory') {
+                      const dirRel = filtersActive ? e.name : (currentPath ? `${currentPath}/${e.name}` : e.name)
+                      const res = await fetch('/api/ai-tagging/translate-bulk', {
+                        method: 'POST',
+                        headers: { 'Content-Type': 'application/json' },
+                        body: JSON.stringify({ libraryId, path: dirRel }),
+                      })
+                      if (!res.ok) {
+                        const data = await res.json().catch(() => ({}))
+                        throw new Error((data as { error?: string }).error ?? 'Translation failed')
+                      }
+                    } else {
+                      const itemKey = itemKeyFor(e)
+                      const res = await fetch('/api/ai-tagging/translate', {
+                        method: 'POST',
+                        headers: { 'Content-Type': 'application/json' },
+                        body: JSON.stringify({ itemKey }),
+                      })
+                      if (!res.ok) {
+                        const data = await res.json().catch(() => ({}))
+                        throw new Error((data as { error?: string }).error ?? 'Translation failed')
+                      }
                     }
                   }}
                   onDelete={(e) => {
@@ -520,9 +534,11 @@ export default function MixedView({ libraryId, initialPath }: Props) {
           name={modal.name}
           itemKey={modal.itemKey}
           onTagsChanged={() => { setFilterRefreshKey((k) => k + 1); fetchAssignments() }}
-          onClose={() => setModal(null)}
+          onClose={() => { setModal(null); setModalShowTags(false) }}
           onPrev={modal.mediaIndex > 0 ? () => navigateModal(-1) : undefined}
           onNext={modal.mediaIndex < mediaEntries.length - 1 ? () => navigateModal(1) : undefined}
+          showTags={modalShowTags}
+          onShowTagsChange={setModalShowTags}
           onAiTag={modal.itemKey ? async () => {
             const res = await fetch('/api/ai-tagging', {
               method: 'POST',
@@ -544,9 +560,11 @@ export default function MixedView({ libraryId, initialPath }: Props) {
           name={modal.name}
           itemKey={modal.itemKey}
           onTagsChanged={() => { setFilterRefreshKey((k) => k + 1); fetchAssignments() }}
-          onClose={() => setModal(null)}
+          onClose={() => { setModal(null); setModalShowTags(false) }}
           onPrev={modal.mediaIndex > 0 ? () => navigateModal(-1) : undefined}
           onNext={modal.mediaIndex < mediaEntries.length - 1 ? () => navigateModal(1) : undefined}
+          showTags={modalShowTags}
+          onShowTagsChange={setModalShowTags}
           onAiTag={async () => {
             const res = await fetch('/api/ai-tagging', {
               method: 'POST',
@@ -746,7 +764,7 @@ function EntryTile({ entry, onOpen, onTag, onDelete, onRename, onAiTag, onExtrac
       </button>
 
       {/* Kebab menu — bottom-right, shown on hover */}
-      {(onDelete || onRename || (onAiTag && entry.mediaType === 'image') || (onExtractText && entry.mediaType === 'image') || (onExtractText && entry.type === 'directory') || (onDescribe && (entry.mediaType === 'image' || entry.mediaType === 'video' || entry.type === 'directory'))) && (
+      {(onDelete || onRename || (onAiTag && entry.mediaType === 'image') || (onExtractText && entry.mediaType === 'image') || (onExtractText && entry.type === 'directory') || (onDescribe && (entry.mediaType === 'image' || entry.mediaType === 'video' || entry.type === 'directory')) || (onTranslate && (entry.mediaType === 'image' || entry.type === 'directory') && entry.hasExtractedText)) && (
         <div className="absolute bottom-2 right-2 opacity-0 group-hover:opacity-100 transition-opacity hidden group-hover:block z-10" ref={menuRef}>
           <button
             onClick={(e) => { e.stopPropagation(); setMenuOpen((o) => !o); setConfirming(false); setAiTagError(null); setDescribeError(null) }}
@@ -917,7 +935,7 @@ function EntryTile({ entry, onOpen, onTag, onDelete, onRename, onAiTag, onExtrac
                   🔍 Extract Text for Folder
                 </button>
               )}
-              {onTranslate && entry.mediaType === 'image' && entry.hasExtractedText && (
+              {onTranslate && (entry.mediaType === 'image' || entry.type === 'directory') && entry.hasExtractedText && (
                 <button
                   onClick={(e) => {
                     e.stopPropagation()
@@ -934,7 +952,7 @@ function EntryTile({ entry, onOpen, onTag, onDelete, onRename, onAiTag, onExtrac
                   onMouseEnter={(e) => ((e.currentTarget as HTMLElement).style.backgroundColor = 'var(--border)')}
                   onMouseLeave={(e) => ((e.currentTarget as HTMLElement).style.backgroundColor = 'transparent')}
                 >
-                  🌐 Translate
+                  {entry.type === 'directory' ? '🌐 Translate Folder' : '🌐 Translate'}
                 </button>
               )}
               {onRename && (

src/components/mixed/VideoPlayerModal.tsx

@@ -14,15 +14,20 @@ interface Props {
   onTagsChanged?: () => void
   onAiTag?: () => Promise<void>
   context?: 'mixed' | 'movies' | 'tv'
+  showTags?: boolean
+  onShowTagsChange?: (v: boolean) => void
 }
 
-export default function VideoPlayerModal({ url, name, onClose, onPrev, onNext, itemKey, onTagsChanged, onAiTag, context = 'mixed' }: Props) {
+export default function VideoPlayerModal({ url, name, onClose, onPrev, onNext, itemKey, onTagsChanged, onAiTag, context = 'mixed', showTags: showTagsProp, onShowTagsChange }: Props) {
   const settings = useUserSettings()
   const autoPlay = context === 'mixed' ? settings.mixedAutoplay : context === 'movies' ? settings.moviesAutoplay : settings.tvAutoplay
   const loop     = context === 'mixed' ? settings.mixedLoop     : context === 'movies' ? settings.moviesLoop     : settings.tvLoop
   const muted    = context === 'mixed' ? settings.mixedMuted    : context === 'movies' ? settings.moviesMuted    : settings.tvMuted
+
   const overlayRef = useRef<HTMLDivElement>(null)
-  const [showTags, setShowTags] = useState(false)
+  const [showTagsLocal, setShowTagsLocal] = useState(false)
+  const showTags = showTagsProp ?? showTagsLocal
+  const setShowTags = onShowTagsChange ?? setShowTagsLocal
   const [aiTagging, setAiTagging] = useState(false)
   const [aiTagError, setAiTagError] = useState<string | null>(null)
   const [tagRefreshKey, setTagRefreshKey] = useState(0)
@@ -45,90 +50,72 @@ export default function VideoPlayerModal({ url, name, onClose, onPrev, onNext, i
     if (e.target === overlayRef.current) onClose()
   }
 
+  const handleAiTag = async () => {
+    if (!onAiTag) return
+    setAiTagging(true)
+    setAiTagError(null)
+    try {
+      await onAiTag()
+      setTagRefreshKey((k) => k + 1)
+      onTagsChanged?.()
+    } catch (err) {
+      setAiTagError(err instanceof Error ? err.message : 'AI tagging failed')
+      setTimeout(() => setAiTagError(null), 4000)
+    } finally {
+      setAiTagging(false)
+    }
+  }
+
   const smallBtn = 'w-7 h-7 rounded-full flex items-center justify-center transition-colors flex-shrink-0'
 
   return (
     <div
       ref={overlayRef}
-      className="fixed inset-0 z-50 flex flex-col items-center p-4 gap-3 overflow-hidden max-h-screen"
-      style={{ backgroundColor: 'rgba(0,0,0,0.9)', height: '100vh', maxHeight: '100vh' }}
+      className="fixed inset-0 z-50 overflow-hidden"
+      style={{ backgroundColor: 'rgba(0,0,0,0.9)', height: '100vh' }}
       onClick={handleOverlayClick}
     >
-      {/* Toolbar — collapses to just filename when tag panel is open */}
-      <div className={`flex items-center justify-between w-full flex-shrink-0 ${showTags ? '' : 'max-w-4xl'}`}>
-        <span className="text-sm truncate max-w-[80%]" style={{ color: 'var(--text-secondary)' }}>
-          {name}
-        </span>
-        {!showTags && (
-          <div className="flex items-center gap-2 flex-shrink-0">
-            {itemKey && (
+      {/* Outer flex — row on md+, col on mobile when panel open */}
+      <div className={`flex h-full w-full ${showTags ? 'flex-col md:flex-row' : 'flex-row'}`}>
+
+        {/* ── Video column ── */}
+        <div className="flex flex-col flex-1 min-h-0 min-w-0 relative">
+
+          {/* Toolbar — scoped to this column's width */}
+          <div className="flex items-center justify-between px-3 py-2 flex-shrink-0" onClick={(e) => e.stopPropagation()}>
+            <span className="text-sm truncate mr-2" style={{ color: 'var(--text-secondary)' }}>
+              {name}
+            </span>
+            <div className="flex items-center gap-1.5 flex-shrink-0">
+              {itemKey && !showTags && (
+                <button
+                  onClick={() => setShowTags(true)}
+                  className={smallBtn}
+                  style={{ backgroundColor: 'var(--surface)', color: 'var(--text-primary)' }}
+                  onMouseEnter={(e) => (e.currentTarget as HTMLElement).style.backgroundColor = 'var(--surface-hover)'}
+                  onMouseLeave={(e) => (e.currentTarget as HTMLElement).style.backgroundColor = 'var(--surface)'}
+                  aria-label="Show tags"
+                  title="Tags"
+                >
+                  🏷
+                </button>
+              )}
               <button
-                onClick={(e) => { e.stopPropagation(); setShowTags(true) }}
-                className="w-8 h-8 rounded-full flex items-center justify-center text-sm transition-colors"
+                onClick={onClose}
+                className={smallBtn}
                 style={{ backgroundColor: 'var(--surface)', color: 'var(--text-primary)' }}
                 onMouseEnter={(e) => (e.currentTarget as HTMLElement).style.backgroundColor = 'var(--surface-hover)'}
                 onMouseLeave={(e) => (e.currentTarget as HTMLElement).style.backgroundColor = 'var(--surface)'}
-                aria-label="Show tags"
-                title="Tags"
+                aria-label="Close"
+                title="Close"
               >
-                🏷
+                ✕
               </button>
-            )}
-            {onAiTag && (
-              <button
-                onClick={async (e) => {
-                  e.stopPropagation()
-                  setAiTagging(true)
-                  setAiTagError(null)
-                  try {
-                    await onAiTag()
-                    setTagRefreshKey((k) => k + 1)
-                    onTagsChanged?.()
-                  } catch (err) {
-                    setAiTagError(err instanceof Error ? err.message : 'AI tagging failed')
-                    setTimeout(() => setAiTagError(null), 4000)
-                  } finally {
-                    setAiTagging(false)
-                  }
-                }}
-                disabled={aiTagging}
-                className="w-8 h-8 rounded-full flex items-center justify-center text-sm transition-colors disabled:opacity-50"
-                style={{
-                  backgroundColor: aiTagError ? '#7f1d1d' : 'var(--surface)',
-                  color: aiTagError ? '#fca5a5' : 'var(--text-primary)',
-                }}
-                onMouseEnter={(e) => {
-                  if (!aiTagging && !aiTagError) (e.currentTarget as HTMLElement).style.backgroundColor = 'var(--surface-hover)'
-                }}
-                onMouseLeave={(e) => {
-                  if (!aiTagError) (e.currentTarget as HTMLElement).style.backgroundColor = 'var(--surface)'
-                }}
-                aria-label="AI Tag this video"
-                title={aiTagError ?? (aiTagging ? 'Tagging…' : 'AI Tag')}
-              >
-                {aiTagging ? (
-                  <span className="animate-spin" style={{ display: 'inline-block' }}>⟳</span>
-                ) : '✨'}
-              </button>
-            )}
-            <button
-              onClick={onClose}
-              className="w-8 h-8 rounded-full flex items-center justify-center text-sm flex-shrink-0 transition-colors"
-              style={{ backgroundColor: 'var(--surface)', color: 'var(--text-primary)' }}
-              onMouseEnter={(e) => (e.currentTarget as HTMLElement).style.backgroundColor = 'var(--surface-hover)'}
-              onMouseLeave={(e) => (e.currentTarget as HTMLElement).style.backgroundColor = 'var(--surface)'}
-              aria-label="Close"
-            >
-              ✕
-            </button>
+            </div>
           </div>
-        )}
-      </div>
 
-      {showTags ? (
-        <div className="flex gap-4 w-full flex-1 min-h-0 items-start overflow-hidden">
-          {/* Video */}
-          <div className="flex-1 min-w-0 min-h-0 flex items-center justify-center max-h-full relative">
+          {/* Video area — single element, never remounts on panel toggle */}
+          <div className="relative flex-1 min-h-0" onClick={(e) => e.stopPropagation()}>
             <video
               key={url}
               src={url}
@@ -136,40 +123,45 @@ export default function VideoPlayerModal({ url, name, onClose, onPrev, onNext, i
               autoPlay={autoPlay}
               muted={muted}
               loop={loop}
-              className="w-full h-full object-contain rounded-lg"
+              playsInline
+              className="w-full h-full object-contain"
               style={{ backgroundColor: '#000' }}
-              onClick={(e) => e.stopPropagation()}
             />
-            {onPrev && (
-              <button
-                onClick={(e) => { e.stopPropagation(); onPrev() }}
-                className="absolute left-2 top-1/2 -translate-y-1/2 w-12 h-12 rounded-full flex items-center justify-center text-lg transition-opacity hover:opacity-100 opacity-70"
-                style={{ backgroundColor: 'rgba(0,0,0,0.4)', color: '#fff' }}
-                aria-label="Previous"
-              >
-                ‹
-              </button>
-            )}
-            {onNext && (
-              <button
-                onClick={(e) => { e.stopPropagation(); onNext() }}
-                className="absolute right-2 top-1/2 -translate-y-1/2 w-12 h-12 rounded-full flex items-center justify-center text-lg transition-opacity hover:opacity-100 opacity-70"
-                style={{ backgroundColor: 'rgba(0,0,0,0.4)', color: '#fff' }}
-                aria-label="Next"
-              >
-                ›
-              </button>
-            )}
           </div>
 
-          {/* Tag panel */}
+          {/* Prev/Next — positioned relative to the full column height (incl. toolbar)
+              so they align with ImageLightbox's buttons which span the full viewport */}
+          {onPrev && (
+            <button
+              onClick={(e) => { e.stopPropagation(); onPrev() }}
+              className="absolute left-2 top-1/2 -translate-y-1/2 z-10 w-12 h-12 rounded-full flex items-center justify-center text-lg transition-opacity hover:opacity-100 opacity-70"
+              style={{ backgroundColor: 'rgba(0,0,0,0.4)', color: '#fff' }}
+              aria-label="Previous"
+            >
+              ‹
+            </button>
+          )}
+          {onNext && (
+            <button
+              onClick={(e) => { e.stopPropagation(); onNext() }}
+              className="absolute right-2 top-1/2 -translate-y-1/2 z-10 w-12 h-12 rounded-full flex items-center justify-center text-lg transition-opacity hover:opacity-100 opacity-70"
+              style={{ backgroundColor: 'rgba(0,0,0,0.4)', color: '#fff' }}
+              aria-label="Next"
+            >
+              ›
+            </button>
+          )}
+        </div>
+
+        {/* ── Tag panel ── bottom half on mobile, right sidebar on desktop */}
+        {showTags && (
           <div
-            className="w-80 h-full max-h-full flex-shrink-0 rounded-xl overflow-y-auto p-4 flex flex-col gap-4"
+            className="flex-shrink-0 flex flex-col overflow-hidden w-full max-h-[50vh] md:w-80 md:max-h-none md:h-full"
             style={{ backgroundColor: 'var(--surface)', border: '1px solid var(--border)' }}
             onClick={(e) => e.stopPropagation()}
           >
-            {/* Panel header — hide panel + AI tagger + close */}
-            <div className="flex items-center justify-between flex-shrink-0">
+            {/* Panel header — ‹ hide  |  ✕ close */}
+            <div className="flex items-center justify-between p-4 flex-shrink-0">
               <button
                 onClick={() => setShowTags(false)}
                 className={smallBtn}
@@ -182,23 +174,29 @@ export default function VideoPlayerModal({ url, name, onClose, onPrev, onNext, i
                 ›
               </button>
               <div className="flex items-center gap-1.5">
+                <button
+                  onClick={onClose}
+                  className={smallBtn}
+                  style={{ backgroundColor: 'var(--border)', color: 'var(--text-secondary)', fontSize: '0.85rem' }}
+                  onMouseEnter={(e) => (e.currentTarget as HTMLElement).style.backgroundColor = 'var(--text-secondary)'}
+                  onMouseLeave={(e) => (e.currentTarget as HTMLElement).style.backgroundColor = 'var(--border)'}
+                  aria-label="Close"
+                  title="Close"
+                >
+                  ✕
+                </button>
+              </div>
+            </div>
+
+            {/* Tags */}
+            <div className="overflow-y-auto flex-1 min-h-0 px-4 pb-4" style={{ borderTop: '1px solid var(--border)' }}>
+              <div className="flex items-center justify-between mt-4 mb-3">
+                <p className="text-xs font-semibold uppercase tracking-wider" style={{ color: 'var(--text-secondary)' }}>
+                  Tags
+                </p>
                 {onAiTag && (
                   <button
-                    onClick={async (e) => {
-                      e.stopPropagation()
-                      setAiTagging(true)
-                      setAiTagError(null)
-                      try {
-                        await onAiTag()
-                        setTagRefreshKey((k) => k + 1)
-                        onTagsChanged?.()
-                      } catch (err) {
-                        setAiTagError(err instanceof Error ? err.message : 'AI tagging failed')
-                        setTimeout(() => setAiTagError(null), 4000)
-                      } finally {
-                        setAiTagging(false)
-                      }
-                    }}
+                    onClick={handleAiTag}
                     disabled={aiTagging}
                     className={`${smallBtn} disabled:opacity-50`}
                     style={{
@@ -218,64 +216,13 @@ export default function VideoPlayerModal({ url, name, onClose, onPrev, onNext, i
                     {aiTagging ? <span className="animate-spin" style={{ display: 'inline-block' }}>⟳</span> : '✨'}
                   </button>
                 )}
-                <button
-                  onClick={onClose}
-                  className={smallBtn}
-                  style={{ backgroundColor: 'var(--border)', color: 'var(--text-secondary)', fontSize: '0.85rem' }}
-                  onMouseEnter={(e) => (e.currentTarget as HTMLElement).style.backgroundColor = 'var(--text-secondary)'}
-                  onMouseLeave={(e) => (e.currentTarget as HTMLElement).style.backgroundColor = 'var(--border)'}
-                  aria-label="Close"
-                  title="Close"
-                >
-                  ✕
-                </button>
               </div>
-            </div>
-
-            {/* Tags section */}
-            <div style={{ borderTop: '1px solid var(--border)', paddingTop: '1rem' }}>
-              <p className="text-xs font-semibold uppercase tracking-wider mb-3" style={{ color: 'var(--text-secondary)' }}>
-                Tags
-              </p>
+              {aiTagError && <p className="text-xs mb-2" style={{ color: '#f87171' }}>{aiTagError}</p>}
               <TagSelector itemKey={itemKey!} onTagsChanged={onTagsChanged} refreshKey={tagRefreshKey} />
             </div>
           </div>
-        </div>
-      ) : (
-        <div className="w-full flex-1 min-h-0 flex items-center justify-center overflow-hidden max-h-full relative">
-          <video
-            key={url}
-            src={url}
-            controls
-            autoPlay={autoPlay}
-            muted={muted}
-            loop={loop}
-            className="w-full h-full max-w-4xl object-contain rounded-lg"
-            style={{ backgroundColor: '#000' }}
-            onClick={(e) => e.stopPropagation()}
-          />
-          {onPrev && (
-            <button
-              onClick={(e) => { e.stopPropagation(); onPrev() }}
-              className="absolute left-2 top-1/2 -translate-y-1/2 w-12 h-12 rounded-full flex items-center justify-center text-lg transition-opacity hover:opacity-100 opacity-70"
-              style={{ backgroundColor: 'rgba(0,0,0,0.4)', color: '#fff' }}
-              aria-label="Previous"
-            >
-              ‹
-            </button>
-          )}
-          {onNext && (
-            <button
-              onClick={(e) => { e.stopPropagation(); onNext() }}
-              className="absolute right-2 top-1/2 -translate-y-1/2 w-12 h-12 rounded-full flex items-center justify-center text-lg transition-opacity hover:opacity-100 opacity-70"
-              style={{ backgroundColor: 'rgba(0,0,0,0.4)', color: '#fff' }}
-              aria-label="Next"
-            >
-              ›
-            </button>
-          )}
-        </div>
-      )}
+        )}
+      </div>
     </div>
   )
 }

src/lib/ai-jobs.ts

@@ -269,7 +269,7 @@ async function processNextJob(): Promise<boolean> {
         await generateItemDescription(row.item_key)
         break
       case 'extract':
-        await extractItemText(row.item_key, jobPayload?.ocrLanguages)
+        await extractItemText(row.item_key, jobPayload?.ocrLanguages, jobPayload?.ocrMode)
         break
       case 'translate':
         await translateItemText(row.item_key, sourceLanguage || undefined)

src/lib/ai-tagger.ts

@@ -538,7 +538,7 @@ async function extractWithTesseract(
  * Translation is not performed automatically — call translateItemText() separately.
  * Returns { extractedText, translatedText } where translatedText is always null.
  */
-export async function extractItemText(itemKey: string, ocrLanguagesOverride?: string): Promise<{ extractedText: string; translatedText: string | null }> {
+export async function extractItemText(itemKey: string, ocrLanguagesOverride?: string, ocrModeOverride?: string): Promise<{ extractedText: string; translatedText: string | null }> {
   const libraryId = itemKey.split(':')[0]
   const config = getEffectiveAiConfig(libraryId)
 
@@ -567,7 +567,8 @@ export async function extractItemText(itemKey: string, ocrLanguagesOverride?: st
     throw Object.assign(new Error('Text extraction is only available for images'), { code: 'NO_IMAGE' })
   }
 
-  const { ocrMode, ocrLanguages: configOcrLanguages, ocrConfidenceThreshold } = config
+  const { ocrMode: configOcrMode, ocrLanguages: configOcrLanguages, ocrConfidenceThreshold } = config
+  const ocrMode = ocrModeOverride ?? configOcrMode
   const ocrLanguages = ocrLanguagesOverride?.trim() || configOcrLanguages
 
   // ── Tesseract path ────────────────────────────────────────────────────────
@@ -655,6 +656,14 @@ export function updateExtractedText(itemKey: string, text: string): void {
   db.prepare('UPDATE media_items SET extracted_text = ? WHERE item_key = ?').run(text, itemKey)
 }
 
+/**
+ * Update the ai_description of an item.
+ */
+export function updateAiDescription(itemKey: string, description: string): void {
+  const db = getDb()
+  db.prepare('UPDATE media_items SET ai_description = ? WHERE item_key = ?').run(description, itemKey)
+}
+
 /**
  * Translate text to a target language using the chat API.
  * Returns null if the text is already in the target language.