fix: Paddle Direct keeps preprocessing (orient/deskew/dewarp/crop)
Some checks failed
CI / nodejs-lint (push) Has been cancelled
CI / go-lint (push) Has been cancelled
CI / python-lint (push) Has been cancelled
CI / test-go-school (push) Has been cancelled
CI / test-go-edu-search (push) Has been cancelled
CI / test-python-klausur (push) Has been cancelled
CI / test-python-agent-core (push) Has been cancelled
CI / test-nodejs-website (push) Has been cancelled

Uses the cropped/dewarped image instead of the original so the overlay
shows the correctly oriented page. 5 steps instead of 2.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
This commit is contained in:
Benjamin Admin
2026-03-12 16:56:18 +01:00
parent 90c1efd9b0
commit c743a38eaf
3 changed files with 30 additions and 15 deletions

View File

@@ -71,10 +71,10 @@ export default function OcrOverlayPage() {
setSteps(
PADDLE_DIRECT_STEPS.map((s, i) => ({
...s,
status: i < 1 ? 'completed' : i === 1 ? 'active' : 'pending',
status: i < 4 ? 'completed' : i === 4 ? 'active' : 'pending',
})),
)
setCurrentStep(1)
setCurrentStep(4)
} else {
setMode('pipeline')
// Map DB step to overlay UI step
@@ -231,6 +231,12 @@ export default function OcrOverlayPage() {
case 0:
return <StepOrientation sessionId={sessionId} onNext={handleOrientationComplete} />
case 1:
return <StepDeskew sessionId={sessionId} onNext={handleNext} />
case 2:
return <StepDewarp sessionId={sessionId} onNext={handleNext} />
case 3:
return <StepCrop sessionId={sessionId} onNext={handleNext} />
case 4:
return <PaddleDirectStep sessionId={sessionId} onNext={handleNext} />
default:
return null
@@ -472,7 +478,7 @@ export default function OcrOverlayPage() {
: 'text-gray-500 dark:text-gray-400 hover:text-gray-700 dark:hover:text-gray-300'
}`}
>
Paddle Direct (2 Schritte)
Paddle Direct (5 Schritte)
</button>
</div>

View File

@@ -49,11 +49,14 @@ export const OVERLAY_UI_TO_DB: Record<number, number> = {
}
/**
* 2-step pipeline for Paddle Direct mode.
* Upload → PaddleOCR+Overlay (skips deskew/dewarp/crop/rows)
* 5-step pipeline for Paddle Direct mode.
* Same preprocessing (orient/deskew/dewarp/crop), then PaddleOCR replaces rows+words+overlay.
*/
export const PADDLE_DIRECT_STEPS: PipelineStep[] = [
{ id: 'orientation', name: 'Upload', icon: '📤', status: 'pending' },
{ id: 'orientation', name: 'Orientierung', icon: '🔄', status: 'pending' },
{ id: 'deskew', name: 'Begradigung', icon: '📐', status: 'pending' },
{ id: 'dewarp', name: 'Entzerrung', icon: '🔧', status: 'pending' },
{ id: 'crop', name: 'Zuschneiden', icon: '✂️', status: 'pending' },
{ id: 'paddle-direct', name: 'PaddleOCR + Overlay', icon: '⚡', status: 'pending' },
]

View File

@@ -2511,17 +2511,23 @@ async def _word_stream_generator(
@router.post("/sessions/{session_id}/paddle-direct")
async def paddle_direct(session_id: str):
"""Run PaddleOCR on the original image and build a word grid directly.
"""Run PaddleOCR on the preprocessed image and build a word grid directly.
Skips deskew/dewarp/crop/rows — just Upload → PaddleOCR → Overlay.
The original image is stored as cropped_png so OverlayReconstruction
Expects orientation/deskew/dewarp/crop to be done already.
Uses the cropped image (falls back to dewarped, then original).
The used image is stored as cropped_png so OverlayReconstruction
can display it as the background.
"""
original_png = await get_session_image(session_id, "original")
if not original_png:
raise HTTPException(status_code=404, detail="No original image found for this session")
# Try preprocessed images first (crop > dewarp > original)
img_png = await get_session_image(session_id, "cropped")
if not img_png:
img_png = await get_session_image(session_id, "dewarped")
if not img_png:
img_png = await get_session_image(session_id, "original")
if not img_png:
raise HTTPException(status_code=404, detail="No image found for this session")
img_arr = np.frombuffer(original_png, dtype=np.uint8)
img_arr = np.frombuffer(img_png, dtype=np.uint8)
img_bgr = cv2.imdecode(img_arr, cv2.IMREAD_COLOR)
if img_bgr is None:
raise HTTPException(status_code=400, detail="Failed to decode original image")
@@ -2562,11 +2568,11 @@ async def paddle_direct(session_id: str):
},
}
# Store original image as cropped_png so OverlayReconstruction shows it
# Store preprocessed image as cropped_png so OverlayReconstruction shows it
await update_session_db(
session_id,
word_result=word_result,
cropped_png=original_png,
cropped_png=img_png,
current_step=8,
)