fix

gateway/app/services/ocr.py

@@ -25,8 +25,13 @@ def recognize_image_bytes(image_bytes: bytes) -> Dict[str, Any]:
     img = Image.open(io.BytesIO(image_bytes))
     width, height = img.size
 
-    # Run RapidOCR
-    result, _ = _ocr_engine(image_bytes)
+    # Run RapidOCR with parameters tuned for numbers, single digits, isolated symbols (- / .), and table cells
+    result, _ = _ocr_engine(
+        image_bytes,
+        box_thresh=0.08,
+        text_score=0.08,
+        unclip_ratio=2.5,
+    )
 
     lines = []
     if result:
@@ -81,7 +86,7 @@ def recognize_image_bytes(image_bytes: bytes) -> Dict[str, Any]:
     }
 
 
-def process_pdf_page_ocr(doc: Any, page_index: int, dpi: int = 150) -> Dict[str, Any]:
+def process_pdf_page_ocr(doc: Any, page_index: int, dpi: int = 200) -> Dict[str, Any]:
     if not is_ocr_available():
         raise RuntimeError("RapidOCR engine is not available")
 

gateway/tests/test_number_ocr.py

@@ -0,0 +1,30 @@
+import io
+from PIL import Image, ImageDraw, ImageFont
+from rapidocr_onnxruntime import RapidOCR
+
+# Create synthetic table image with single numbers 1, 2, 3, 4, 5, 6, 7, 8
+img = Image.new("RGB", (300, 400), color=(255, 255, 255))
+draw = ImageDraw.Draw(img)
+
+draw.text((20, 20), "#", fill=(0, 0, 0))
+draw.text((80, 20), "Item Description", fill=(0, 0, 0))
+
+numbers = ["1", "2", "3", "4", "5", "6", "7", "8"]
+y = 50
+for num in numbers:
+    draw.text((20, y), num, fill=(0, 0, 0))
+    draw.text((80, y), f"Item {num}", fill=(0, 0, 0))
+    y += 35
+
+buf = io.BytesIO()
+img.save(buf, format="PNG")
+png_bytes = buf.getvalue()
+
+from app.services.ocr import recognize_image_bytes
+
+print("--- Testing Gateway recognize_image_bytes with tuned number parameters ---")
+res = recognize_image_bytes(png_bytes)
+for line in res["lines"]:
+    safe_txt = repr(line["text"]).encode('ascii', 'backslashreplace').decode('ascii')
+    print("  Detected:", safe_txt, "| box:", line["box"], "| score:", line["confidence"])
+