On-Device-OCR via ML Kit in Image-Preview-Flow integrieren (v0.6.0) #425
Labels
No labels
app/archiv
app/einkaufslisten
app/imap-client
app/wissensbasis
arch-answered
arch-question
area/api
area/auth
area/infra
area/mobile
area/shared
area/ui
area/web
portfolio-status
prio/high
prio/low
prio/medium
roadmap/public
size/l
size/m
size/s
size/xl
size/xs
status/blocked
status/needs-info
type/bug
type/chore
type/docs
type/feature
type/idea
type/refactor
No milestone
No project
No assignees
1 participant
Notifications
Due date
No due date set.
Dependencies
No dependencies set.
Reference
admin-mrrm/mrrmlabapp#425
Loading…
Add table
Add a link
Reference in a new issue
No description provided.
Delete branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Kontext
#423/PR #424 hat ML Kit Text Recognition v2 als On-Device-OCR-Stack validiert (Spike-Screen
mrrmlab:///ocr-spike, handschriftliches „Oregano“ → CER ≈ 0.28). Jetzt produktiv in den User-Flow integrieren — Ziel: v0.6.0-Release.Why
Status quo
apps/mobile/app/lists/[listId]/image-preview.tsx→parseImage.mutate({ file, model })→ Server-OCR (EasyOCR/de)OCR_MODELSinpackages/feature-shopping-list/src/components/shopping-list-screen.tsxenthält bisher nureasyocr/deScope v0.6.0 (minimal pragmatisch)
Mobile:
OCR_MODELSummlkit/deerweitern (Label „Deutsch (on-device)“, als Default setzen)image-preview.tsx: beimodel === mlkit/deon-device-Pfad nehmen:ocrService.recognize(uri)→ Plaintext/lists/parse-text-Endpoint → Items zurück → Review-ScreenBackend:
4. Neuer Endpoint
POST /lists/parse-text(NestJS): Body{ text: string }, Response{ items: ShoppingItem[] }. Reuse der existierenden Item-Extraktions-Logik (parseItems), nur ohne den OCR-Schritt davor.5. API-Client:
parseShoppingListText(text)neben dem bestehendenparseShoppingListImage6. Hook in
feature-shopping-list:useParseShoppingListTextanalog zuuseParseShoppingListImageAcceptance Criteria
ocr-spike.tsxentfernt (Route + File)Out-of-Scope (Folge-Stories)
llama.rn): heavier scope, eigener SpikeRelease
Nach Merge: v0.6.0 cutten. Changelog: „On-Device-OCR — schneller und privater scannen, funktioniert offline.“
Verwandt