Update utils/image_processor.py

utils/image_processor.py

@@ -2,34 +2,69 @@ import os
 import cv2
 import numpy as np
 from PIL import Image
-import torch
-from transformers import TrOCRProcessor, VisionEncoderDecoderModel
+import easyocr
 
-# ✅ Load TrOCR processor and model once
-try:
-    processor = TrOCRProcessor.from_pretrained("microsoft/trocr-base-printed")
-    trocr_model = VisionEncoderDecoderModel.from_pretrained("microsoft/trocr-base-printed")
+# ✅ Safe directory for OCR model storage
+EASYOCR_DIR = "/tmp/.easyocr"
+os.makedirs(EASYOCR_DIR, exist_ok=True)
+
+# ✅ Set environment variable before import (optional but safe)
+os.environ["EASYOCR_HOME"] = EASYOCR_DIR
 
-    print("✅ TrOCR model loaded.")
+# ✅ Initialize EasyOCR
+try:
+    reader = easyocr.Reader(['en'], model_storage_directory=EASYOCR_DIR)
+    ocr_available = True
+    print("✅ EasyOCR initialized.")
 except Exception as e:
-    print(f"❌ Failed to load TrOCR model: {e}")
-    processor = None
-    trocr_model = None
+    print(f"❌ EasyOCR initialization failed: {str(e)}")
+    ocr_available = False
+
+def preprocess_image(image):
+    """
+    Preprocess image to improve OCR accuracy
+    """
+    if len(image.shape) == 3:
+        gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
+    else:
+        gray = image
+    denoised = cv2.fastNlMeansDenoising(gray, h=10)
+    processed = cv2.adaptiveThreshold(
+        denoised, 255, cv2.ADAPTIVE_THRESH_GAUSSIAN_C,
+        cv2.THRESH_BINARY, 11, 2
+    )
+    return processed
 
 def extract_text_from_image(image_path):
+    """
+    Extract text from image using EasyOCR
+    """
     print(f"📂 Reading image from: {image_path}")
     try:
-        if trocr_model is None or processor is None:
-            raise RuntimeError("TrOCR model or processor not loaded.")
+        if not ocr_available:
+            raise ValueError("EasyOCR not available")
+
+        image = cv2.imread(image_path)
+        if image is None:
+            raise ValueError(f"Could not read image at {image_path}")
+        
+        processed = preprocess_image(image)
+        temp_path = os.path.join(os.path.dirname(image_path), f"temp_{os.path.basename(image_path)}")
+        cv2.imwrite(temp_path, processed)
+
+        results = reader.readtext(temp_path)
+        os.remove(temp_path)
+
+        text = ' '.join([res[1] for res in results]).strip()
 
-        image = Image.open(image_path).convert("RGB")
-        pixel_values = processor(images=image, return_tensors="pt").pixel_values
+        # Fallback to original if empty
+        if not text:
+            results = reader.readtext(image_path)
+            text = ' '.join([res[1] for res in results]).strip()
 
-        generated_ids = trocr_model.generate(pixel_values)
-        generated_text = processor.batch_decode(generated_ids, skip_special_tokens=True)[0]
+        print("📝 Extracted Text:", text)
+        return text or "Text extraction failed. Please enter text manually."
 
-        return generated_text.strip() if generated_text else "Text extraction failed."
-    
     except Exception as e:
-        print(f"❌ OCR failed on {image_path}: {e}")
-        return "Text extraction failed."
+        print(f"❌ OCR failed: {e}")
+        return "Text extraction failed. Please enter text manually."