Update ocr_engine.py

ocr_engine.py

@@ -6,80 +6,57 @@ import re
 reader = easyocr.Reader(['en'], gpu=False)
 
 def enhance_image(img):
-    # Convert to grayscale
     gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
 
-    # Apply sharpening kernel
-    kernel = np.array([[0, -1, 0], [-1, 5,-1], [0, -1, 0]])
-    sharp = cv2.filter2D(gray, -1, kernel)
+    # CLAHE (adaptive histogram equalization for better contrast)
+    clahe = cv2.createCLAHE(clipLimit=2.0, tileGridSize=(8, 8))
+    contrast = clahe.apply(gray)
 
-    # Contrast Limited Adaptive Histogram Equalization (CLAHE)
-    clahe = cv2.createCLAHE(clipLimit=2.0, tileGridSize=(8,8))
-    contrast = clahe.apply(sharp)
+    # Sharpen image
+    kernel = np.array([[0, -1, 0], [-1, 5, -1], [0, -1, 0]])
+    sharpened = cv2.filter2D(contrast, -1, kernel)
 
-    # Denoising
-    denoised = cv2.fastNlMeansDenoising(contrast, h=30)
+    # Resize if very small
+    h, w = sharpened.shape
+    if max(h, w) < 500:
+        sharpened = cv2.resize(sharpened, None, fx=2, fy=2, interpolation=cv2.INTER_CUBIC)
 
-    # Adaptive threshold for very dim images
-    thresh = cv2.adaptiveThreshold(denoised, 255, cv2.ADAPTIVE_THRESH_GAUSSIAN_C, 
-                                   cv2.THRESH_BINARY, 11, 2)
-
-    return thresh
+    return sharpened
 
 def extract_weight_from_image(pil_img):
     try:
         img = np.array(pil_img)
+        img = cv2.cvtColor(img, cv2.COLOR_RGB2BGR)  # convert from PIL to OpenCV BGR
 
-        # Resize if too large or too small
-        max_dim = 1000
-        height, width = img.shape[:2]
-        if max(height, width) > max_dim:
-            scale = max_dim / max(height, width)
-            img = cv2.resize(img, None, fx=scale, fy=scale, interpolation=cv2.INTER_AREA)
-        elif max(height, width) < 400:
-            scale = 2.5  # Upscale very small images
-            img = cv2.resize(img, None, fx=scale, fy=scale, interpolation=cv2.INTER_CUBIC)
-
-        # Enhance image for OCR
-        preprocessed = enhance_image(img)
+        # Preprocess image
+        processed = enhance_image(img)
 
-        results = reader.readtext(preprocessed)
+        results = reader.readtext(processed)
 
         best_weight = None
         best_conf = 0.0
 
-        for item in results:
-            if len(item) != 2 or not isinstance(item[1], tuple):
-                continue
-            text, conf = item[1]
-            cleaned = text.lower().strip()
+        for (bbox, text, conf) in results:
+            original_text = text
+            text = text.lower().strip()
 
-            cleaned = cleaned.replace(",", ".")
-            cleaned = cleaned.replace("o", "0").replace("O", "0")
-            cleaned = cleaned.replace("s", "5").replace("S", "5")
-            cleaned = cleaned.replace("g", "9").replace("G", "6")
-            cleaned = cleaned.replace("kg", "").replace("kgs", "")
-            cleaned = re.sub(r"[^\d\.]", "", cleaned)
+            # Fix common OCR errors
+            text = text.replace(",", ".")
+            text = text.replace("o", "0").replace("O", "0")
+            text = text.replace("s", "5").replace("S", "5")
+            text = text.replace("g", "9").replace("G", "6")
+            text = text.replace("kgs", "").replace("kg", "")
+            text = re.sub(r"[^\d\.]", "", text)
 
-            if re.fullmatch(r"\d{2,4}(\.\d{1,3})?", cleaned):
+            if re.fullmatch(r"\d{1,4}(\.\d{1,3})?", text):
                 if conf > best_conf:
-                    best_weight = cleaned
-                    best_conf = conf
-
-        if not best_weight:
-            for item in results:
-                if len(item) != 2 or not isinstance(item[1], tuple):
-                    continue
-                text, conf = item[1]
-                fallback = re.sub(r"[^\d\.]", "", text)
-                if fallback and fallback.replace(".", "").isdigit():
-                    best_weight = fallback
+                    best_weight = text
                     best_conf = conf
-                    break
 
         if not best_weight:
             return "Not detected", 0.0
 
+        # Format output
         if "." in best_weight:
             int_part, dec_part = best_weight.split(".")
             int_part = int_part.lstrip("0") or "0"