gpatti/MediaLore
Archived
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

feat: per-extraction OCR language override

Browse Source 
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>
This commit is contained in:
Garret Patti
2026-04-13 21:55:07 -04:00
parent 96cfb8aae7
commit db2e446ef4
8 changed files with 206 additions and 70 deletions
Download Patch File Download Diff File Expand all files Collapse all files

11
src/app/api/ai-settings/ocr/route.ts Normal file
View File

@@ -0,0 +1,11 @@
import { NextRequest, NextResponse } from 'next/server'
import { requireAuth } from '@/lib/auth'
import { getAiConfig } from '@/lib/app-settings'
export async function GET(request: NextRequest) {
const auth = await requireAuth(request)
if (auth instanceof NextResponse) return auth
const { ocrMode, ocrLanguages } = getAiConfig()
return NextResponse.json({ ocrMode, ocrLanguages })
}

12
src/app/api/ai-tagging/extract-text/route.ts
View File

@@ -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 }
let body: { itemKey?: string; ocrLanguages?: string }
try {
body = await request.json()
} catch {
return NextResponse.json({ error: 'Invalid JSON body' }, { status: 400 })
}
const { itemKey } = body
const { itemKey, ocrLanguages } = body
if (!itemKey || typeof itemKey !== 'string') {
return NextResponse.json({ error: 'itemKey is required' }, { status: 400 })
}
@@ -19,6 +19,12 @@ export async function POST(request: NextRequest) {
const auth = await requireLibraryAccess(request, libraryId)
if (auth instanceof NextResponse) return auth
const jobId = enqueueJob(itemKey, 'extract', libraryId)
const jobId = enqueueJob(
itemKey,
'extract',
libraryId,
undefined,
ocrLanguages ? { ocrLanguages } : undefined,
)
return NextResponse.json({ jobId }, { status: 202 })
}

2
src/components/DoomScrollView.tsx
View File

@@ -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()}
>

37
src/components/mixed/ImageLightbox.tsx
View File

@@ -39,6 +39,11 @@ export default function ImageLightbox({ url, name, onClose, onPrev, onNext, item
const [descPending, setDescPending] = useState(false)
const [descError, setDescError] = useState<string | null>(null)
// OCR settings
const [ocrMode, setOcrMode] = useState<string | null>(null)
const [defaultOcrLanguages, setDefaultOcrLanguages] = useState('eng')
const [ocrLanguageInput, setOcrLanguageInput] = useState('')
// Text overlay state
const [showTextOverlay, setShowTextOverlay] = useState(false)
const [showOriginal, setShowOriginal] = useState(false)
@@ -68,6 +73,13 @@ export default function ImageLightbox({ url, name, onClose, onPrev, onNext, item
useEffect(() => {
fetchAiFields()
fetch('/api/ai-settings/ocr')
.then((r) => r.json())
.then((d: { ocrMode: string; ocrLanguages: string }) => {
setOcrMode(d.ocrMode)
setDefaultOcrLanguages(d.ocrLanguages)
})
.catch(() => {})
return () => {
if (pollRef.current) clearInterval(pollRef.current)
}
@@ -439,6 +451,7 @@ export default function ImageLightbox({ url, name, onClose, onPrev, onNext, item
Text Extraction
</p>
<div className="flex items-center gap-2 flex-wrap">
<button
onClick={async () => {
setExtracting(true)
@@ -448,7 +461,10 @@ export default function ImageLightbox({ url, name, onClose, onPrev, onNext, item
const res = await fetch('/api/ai-tagging/extract-text', {
method: 'POST',
headers: { 'Content-Type': 'application/json' },
body: JSON.stringify({ itemKey }),
body: JSON.stringify({
itemKey,
...(ocrLanguageInput.trim() && { ocrLanguages: ocrLanguageInput.trim() }),
}),
})
if (res.status === 202) {
setExtractPending(true)
@@ -471,7 +487,7 @@ export default function ImageLightbox({ url, name, onClose, onPrev, onNext, item
}
}}
disabled={extracting || extractPending}
className="text-xs px-2 py-1 rounded-lg transition-colors disabled:opacity-50 self-start"
className="text-xs px-2 py-1 rounded-lg transition-colors disabled:opacity-50 self-start flex-shrink-0"
style={{
backgroundColor: extractPending ? 'var(--accent)' : 'var(--border)',
color: extractPending ? '#fff' : 'var(--text-secondary)',
@@ -491,6 +507,23 @@ export default function ImageLightbox({ url, name, onClose, onPrev, onNext, item
>
{extracting ? '⟳ Extracting…' : extractPending ? '⟳ Queued…' : extractedText ? '🔍 Re-extract Text' : '🔍 Extract Text'}
</button>
{ocrMode && ocrMode !== 'llm' && (
<input
type="text"
value={ocrLanguageInput}
onChange={(e) => setOcrLanguageInput(e.target.value)}
placeholder={defaultOcrLanguages}
className="text-xs px-2 py-0.5 rounded-full outline-none"
style={{
backgroundColor: 'var(--background)',
border: '1px solid var(--border)',
color: 'var(--text-primary)',
width: 120,
}}
title="Tesseract language(s) for this extraction (e.g. jpn+jpn_vert). Leave blank to use the configured default."
/>
)}
</div>
{extractError && (
<p className="text-xs" style={{ color: '#f87171' }}>{extractError}</p>

81
src/components/mixed/MixedView.tsx
View File

@@ -83,6 +83,9 @@ export default function MixedView({ libraryId, initialPath }: Props) {
setDoomScrollLoading(false)
}, [currentPath])
const [ocrMode, setOcrMode] = useState<string | null>(null)
const [defaultOcrLanguages, setDefaultOcrLanguages] = useState('eng')
const fetchAssignments = useCallback(() => {
fetch(`/api/tags/library-assignments?libraryId=${encodeURIComponent(libraryId)}`)
.then((r) => r.json())
@@ -92,6 +95,16 @@ export default function MixedView({ libraryId, initialPath }: Props) {
useEffect(() => { fetchAssignments() }, [fetchAssignments])
useEffect(() => {
fetch('/api/ai-settings/ocr')
.then((r) => r.json())
.then((d: { ocrMode: string; ocrLanguages: string }) => {
setOcrMode(d.ocrMode)
setDefaultOcrLanguages(d.ocrLanguages)
})
.catch(() => {})
}, [])
const filtersActive = search !== '' || selectedTagIds.size > 0
const fetchRecursive = useCallback(() => {
@@ -387,6 +400,8 @@ export default function MixedView({ libraryId, initialPath }: Props) {
entry={entry}
onOpen={handleEntry}
onTag={handleTagEntry}
ocrMode={ocrMode}
defaultOcrLanguages={defaultOcrLanguages}
onAiTag={async (e) => {
const itemKey = itemKeyFor(e)
const res = await fetch('/api/ai-tagging', {
@@ -401,7 +416,7 @@ export default function MixedView({ libraryId, initialPath }: Props) {
fetchAssignments()
setFilterRefreshKey((k) => k + 1)
}}
onExtractText={async (e) => {
onExtractText={async (e, ocrLanguages) => {
if (e.type === 'directory') {
// Bulk extract for directory
const dirRel = filtersActive ? e.name : (currentPath ? `${currentPath}/${e.name}` : e.name)
@@ -420,7 +435,7 @@ export default function MixedView({ libraryId, initialPath }: Props) {
const res = await fetch('/api/ai-tagging/extract-text', {
method: 'POST',
headers: { 'Content-Type': 'application/json' },
body: JSON.stringify({ itemKey }),
body: JSON.stringify({ itemKey, ...(ocrLanguages && { ocrLanguages }) }),
})
if (!res.ok) {
const data = await res.json().catch(() => ({}))
@@ -594,7 +609,7 @@ export default function MixedView({ libraryId, initialPath }: Props) {
)
}
function EntryTile({ entry, onOpen, onTag, onDelete, onRename, onAiTag, onExtractText, onDescribe, onTranslate }: { entry: FileEntry; onOpen: (e: FileEntry) => void; onTag: (e: FileEntry) => void; onDelete?: (e: FileEntry) => void; onRename?: (e: FileEntry, newName: string) => Promise<boolean>; onAiTag?: (e: FileEntry) => Promise<void>; onExtractText?: (e: FileEntry) => Promise<void>; onDescribe?: (e: FileEntry) => Promise<void>; onTranslate?: (e: FileEntry) => Promise<void> }) {
function EntryTile({ entry, onOpen, onTag, onDelete, onRename, onAiTag, onExtractText, onDescribe, onTranslate, ocrMode, defaultOcrLanguages }: { entry: FileEntry; onOpen: (e: FileEntry) => void; onTag: (e: FileEntry) => void; onDelete?: (e: FileEntry) => void; onRename?: (e: FileEntry, newName: string) => Promise<boolean>; onAiTag?: (e: FileEntry) => Promise<void>; onExtractText?: (e: FileEntry, ocrLanguages?: string) => Promise<void>; onDescribe?: (e: FileEntry) => Promise<void>; onTranslate?: (e: FileEntry) => Promise<void>; ocrMode?: string | null; defaultOcrLanguages?: string }) {
type ImgState = 'loading' | 'loaded' | 'error'
const [imgState, setImgState] = useState<ImgState>(
entry.thumbnailUrl ? 'loading' : 'error'
@@ -615,6 +630,8 @@ function EntryTile({ entry, onOpen, onTag, onDelete, onRename, onAiTag, onExtrac
const [describeError, setDescribeError] = useState<string | null>(null)
const [translating, setTranslating] = useState(false)
const [translateError, setTranslateError] = useState<string | null>(null)
const [showOcrPrompt, setShowOcrPrompt] = useState(false)
const [ocrLanguageInput, setOcrLanguageInput] = useState('')
useEffect(() => {
if (!menuOpen) return
@@ -804,16 +821,21 @@ function EntryTile({ entry, onOpen, onTag, onDelete, onRename, onAiTag, onExtrac
📝 Describe Folder
</button>
)}
{onExtractText && entry.mediaType === 'image' && (
{onExtractText && entry.mediaType === 'image' && !showOcrPrompt && (
<button
onClick={(e) => {
e.stopPropagation()
if (ocrMode && ocrMode !== 'llm') {
setOcrLanguageInput('')
setShowOcrPrompt(true)
} else {
setMenuOpen(false)
setTextExtracting(true)
setTextExtractError(null)
onExtractText(entry)
.catch((err) => setTextExtractError(err instanceof Error ? err.message : 'Text extraction failed'))
.finally(() => setTextExtracting(false))
}
}}
disabled={textExtracting}
className="flex items-center gap-2 w-full px-4 py-2 text-sm text-left transition-colors disabled:opacity-50"
@@ -824,6 +846,57 @@ function EntryTile({ entry, onOpen, onTag, onDelete, onRename, onAiTag, onExtrac
🔍 Extract Text
</button>
)}
{onExtractText && entry.mediaType === 'image' && showOcrPrompt && (
<div className="px-4 py-2 flex flex-col gap-2" onClick={(e) => e.stopPropagation()}>
<p className="text-xs" style={{ color: 'var(--text-secondary)' }}>OCR language</p>
<input
autoFocus
type="text"
value={ocrLanguageInput}
onChange={(e) => setOcrLanguageInput(e.target.value)}
onKeyDown={(e) => {
if (e.key === 'Escape') { setShowOcrPrompt(false) }
if (e.key === 'Enter') {
setShowOcrPrompt(false)
setMenuOpen(false)
setTextExtracting(true)
setTextExtractError(null)
onExtractText(entry, ocrLanguageInput.trim() || undefined)
.catch((err) => setTextExtractError(err instanceof Error ? err.message : 'Text extraction failed'))
.finally(() => setTextExtracting(false))
}
}}
placeholder={defaultOcrLanguages ?? 'eng'}
className="text-xs px-2 py-1 rounded-lg outline-none w-full"
style={{ backgroundColor: 'var(--background)', border: '1px solid var(--border)', color: 'var(--text-primary)' }}
title="Tesseract language(s) for this extraction (e.g. jpn+jpn_vert). Leave blank to use the configured default."
/>
<div className="flex gap-2">
<button
onClick={() => {
setShowOcrPrompt(false)
setMenuOpen(false)
setTextExtracting(true)
setTextExtractError(null)
onExtractText(entry, ocrLanguageInput.trim() || undefined)
.catch((err) => setTextExtractError(err instanceof Error ? err.message : 'Text extraction failed'))
.finally(() => setTextExtracting(false))
}}
className="text-xs px-2 py-1 rounded-lg"
style={{ backgroundColor: 'var(--accent)', color: '#fff' }}
>
Extract
</button>
<button
onClick={() => setShowOcrPrompt(false)}
className="text-xs px-2 py-1"
style={{ color: 'var(--text-secondary)' }}
>
Cancel
</button>
</div>
</div>
)}
{onExtractText && entry.type === 'directory' && (
<button
onClick={(e) => {

12
src/lib/ai-jobs.ts
View File

@@ -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)

5
src/lib/ai-tagger.ts
View File

@@ -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): Promise<{ extractedText: string; translatedText: string | null }> {
export async function extractItemText(itemKey: string, ocrLanguagesOverride?: 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): Promise<{ extractedText:
throw Object.assign(new Error('Text extraction is only available for images'), { code: 'NO_IMAGE' })
}
const { ocrMode, ocrLanguages, ocrConfidenceThreshold } = config
const { ocrMode, ocrLanguages: configOcrLanguages, ocrConfidenceThreshold } = config
const ocrLanguages = ocrLanguagesOverride?.trim() || configOcrLanguages
// ── Tesseract path ────────────────────────────────────────────────────────
if (ocrMode === 'tesseract' || ocrMode === 'hybrid') {

8
src/lib/db.ts
View File

@@ -338,4 +338,12 @@ function migrateAiJobs(db: Database.Database): void {
CREATE INDEX IF NOT EXISTS ai_jobs_status ON ai_jobs(status);
CREATE INDEX IF NOT EXISTS ai_jobs_created_at ON ai_jobs(created_at);
`)
// Add payload column if not present
const aiJobsRow = db
.prepare("SELECT sql FROM sqlite_master WHERE type='table' AND name='ai_jobs'")
.get() as { sql: string } | undefined
if (aiJobsRow && !aiJobsRow.sql.includes('payload')) {
db.exec('ALTER TABLE ai_jobs ADD COLUMN payload TEXT')
}
}