fix: mock

backend/app/api/v1/images.py

@@ -65,7 +65,13 @@ async def upload_images(
 
     # trigger OCR tasks in-process background (non-blocking for API response)
     from app.workers.ocr_tasks import process_images_ocr_batch_async
-    pending_ids = [str(img.id) for img in results if img.ocr_status.value == "pending"]
+    pending_imgs = [img for img in results if img.ocr_status.value == "pending"]
+    for img in pending_imgs:
+        img.ocr_status = OcrStatus.processing
+    if pending_imgs:
+        await db.flush()
+        await db.commit()
+    pending_ids = [str(img.id) for img in pending_imgs]
     if pending_ids:
         asyncio.create_task(
             process_images_ocr_batch_async(
@@ -171,20 +177,25 @@ async def start_case_ocr(
     image_ids = payload.image_ids if payload else []
     if image_ids:
         images = await repo.list_by_ids_in_case(case_id, image_ids)
+        # Never submit images that are already processing: this prevents
+        # duplicate OCR tasks when users trigger OCR from multiple pages.
+        images = [img for img in images if img.ocr_status != OcrStatus.processing]
         # For explicit re-run, mark selected images as processing immediately
         # so frontend can reflect state transition without full page refresh.
         for img in images:
             img.ocr_status = OcrStatus.processing
-        await db.flush()
-        await db.commit()
+        if images:
+            await db.flush()
+            await db.commit()
     else:
         images = await repo.list_for_ocr(case_id, include_done=include_done)
         # Mark queued images as processing immediately, including when OCR is
         # triggered from workspace page, so UI can show progress right away.
         for img in images:
             img.ocr_status = OcrStatus.processing
-        await db.flush()
-        await db.commit()
+        if images:
+            await db.flush()
+            await db.commit()
 
     from app.workers.ocr_tasks import process_images_ocr_batch_async
 

backend/app/core/config.py

@@ -2,9 +2,12 @@ from pathlib import Path
 from pydantic_settings import BaseSettings, SettingsConfigDict
 
 
+ENV_FILE_PATH = Path(__file__).resolve().parents[2] / ".env"
+
+
 class Settings(BaseSettings):
     model_config = SettingsConfigDict(
-        env_file=".env",
+        env_file=ENV_FILE_PATH,
         env_file_encoding="utf-8",
         extra="ignore",
     )
@@ -18,6 +21,7 @@ class Settings(BaseSettings):
     OCR_API_KEY: str = ""
     OCR_API_URL: str = ""
     OCR_MODEL: str = ""
+    OCR_ALLOW_MOCK_FALLBACK: bool = False
     OCR_PARALLELISM: int = 4
     LLM_API_KEY: str = ""
     LLM_API_URL: str = ""

backend/app/repositories/image_repo.py

@@ -51,8 +51,12 @@ class ImageRepository(BaseRepository[EvidenceImage]):
 
     async def list_for_ocr(self, case_id: UUID, include_done: bool = False) -> list[EvidenceImage]:
         query = select(EvidenceImage).where(EvidenceImage.case_id == case_id)
-        if not include_done:
-            query = query.where(EvidenceImage.ocr_status != OcrStatus.done)
+        # Always exclude currently-processing images to avoid duplicate OCR
+        # submission from different trigger paths (upload/workspace/screenshots).
+        if include_done:
+            query = query.where(EvidenceImage.ocr_status != OcrStatus.processing)
+        else:
+            query = query.where(EvidenceImage.ocr_status.in_([OcrStatus.pending, OcrStatus.failed]))
         result = await self.session.execute(query.order_by(EvidenceImage.uploaded_at.desc()))
         return list(result.scalars().all())
 

backend/app/services/ocr_service.py

@@ -43,6 +43,17 @@ def _ocr_available() -> bool:
     return bool(url and key and model)
 
 
+def _missing_ocr_fields() -> list[str]:
+    missing: list[str] = []
+    if not (settings.OCR_API_URL or settings.LLM_API_URL):
+        missing.append("OCR_API_URL(or LLM_API_URL)")
+    if not (settings.OCR_API_KEY or settings.LLM_API_KEY):
+        missing.append("OCR_API_KEY(or LLM_API_KEY)")
+    if not (settings.OCR_MODEL or settings.LLM_MODEL):
+        missing.append("OCR_MODEL(or LLM_MODEL)")
+    return missing
+
+
 def _llm_available() -> bool:
     url, key, model = _llm_config()
     return bool(url and key and model)
@@ -54,6 +65,10 @@ async def classify_page(image_path: str) -> tuple[SourceApp, PageType]:
     """Identify the source app and page type of a screenshot."""
     if _ocr_available():
         return await _classify_via_api(image_path)
+    if not settings.OCR_ALLOW_MOCK_FALLBACK:
+        missing = ", ".join(_missing_ocr_fields()) or "unknown"
+        raise RuntimeError(f"OCR configuration missing: {missing}")
+    logger.warning("OCR unavailable, falling back to mock classification for image: %s", image_path)
     return _classify_mock(image_path)
 
 
@@ -63,6 +78,10 @@ async def extract_transaction_fields(
     """Extract structured transaction fields from a screenshot."""
     if _ocr_available():
         return await _extract_via_api(image_path, source_app, page_type)
+    if not settings.OCR_ALLOW_MOCK_FALLBACK:
+        missing = ", ".join(_missing_ocr_fields()) or "unknown"
+        raise RuntimeError(f"OCR configuration missing: {missing}")
+    logger.warning("OCR unavailable, falling back to mock extraction for image: %s", image_path)
     mock_data = _extract_mock(image_path, source_app, page_type)
     return mock_data, json.dumps(mock_data, ensure_ascii=False)
 

backend/app/workers/ocr_tasks.py

@@ -28,6 +28,9 @@ async def process_images_ocr_batch_async(image_ids: list[str], max_concurrency:
     """Process many images with bounded OCR concurrency."""
     if not image_ids:
         return
+    # De-duplicate in-memory to prevent repeated processing of same image id
+    # in a single batch submission.
+    image_ids = list(dict.fromkeys(image_ids))
     concurrency = max(1, max_concurrency)
     semaphore = asyncio.Semaphore(concurrency)