Upload 2 files

Dockerfile

@@ -20,11 +20,8 @@ COPY requirements.txt .
 RUN pip install --no-cache-dir -r requirements.txt
 
 # 🛑 CRITICAL FIX FOR EASYOCR PERMISSION: 🛑
-# Create the necessary writable directories as root during the build process.
-# This prevents the application from crashing at runtime when the non-root user tries to create it.
-RUN mkdir -p /app/easyocr_models/user_network
-RUN chmod -R 777 /app/easyocr_models
-# NOTE: This ensures both model storage and user network directories are writable.
+# Create the necessary writable directories including user_network
+RUN mkdir -p /app/easyocr_models/user_network && chmod -R 777 /app/easyocr_models
 
 # 3b. Copy the rest of your application code and models
 COPY . .

ocr_utils.py

@@ -3,18 +3,20 @@ import os
 
 # Initialize the OCR reader once to save time
 # 'en' is for English language
-# In ocr_utils.py
-# Set model_storage_directory to a folder inside your writable app directory (/app)
-# ⚠️ Use absolute path to prevent EasyOCR from defaulting to /.EasyOCR
-MODELS_DIR = "/app/easyocr_models"  # <-- Absolute path
-USER_NETWORK_DIR = "/app/easyocr_models/user_network"  # <-- Added to fix PermissionError
-os.makedirs(MODELS_DIR, exist_ok=True) # Ensure the directory exists
-os.makedirs(USER_NETWORK_DIR, exist_ok=True) # Ensure the user network directory exists
+# Set both model storage and user network directories inside /app (writable)
+
+MODELS_DIR = "/app/easyocr_models"
+USER_NETWORK_DIR = "/app/easyocr_models/user_network"
+
+# Ensure directories exist
+os.makedirs(MODELS_DIR, exist_ok=True)
+os.makedirs(USER_NETWORK_DIR, exist_ok=True)
 
 # DEBUG: Verify the directories
 print(f"[DEBUG] EasyOCR models directory: {MODELS_DIR}")
 print(f"[DEBUG] EasyOCR user network directory: {USER_NETWORK_DIR}")
 
+# Initialize EasyOCR reader with both directories specified
 reader = easyocr.Reader(
     ['en'],
     model_storage_directory=MODELS_DIR,
@@ -24,26 +26,11 @@ reader = easyocr.Reader(
 def extract_keywords_from_report(file_path):
     """
     Performs OCR on the uploaded file and extracts relevant text.
-    
-    Args:
-        file_path (str): The path to the uploaded image or PDF.
-        
-    Returns:
-        str: A single string containing all recognized text.
     """
     try:
-        # EasyOCR works well with image files (JPG, PNG)
-        # For PDFs, you'll need an external tool to convert PDF to image pages first, 
-        # but for simplicity, we'll focus on image inputs for now.
-        
-        # Read the text from the image
         results = reader.readtext(file_path, detail=0)
-        
-        # Join all lines into a single string for keyword searching
         full_text = " ".join(results).lower()
-        
         return full_text
-        
     except Exception as e:
         print(f"OCR Error: {e}")
         return ""
@@ -51,12 +38,6 @@ def extract_keywords_from_report(file_path):
 def score_text_for_risk(text):
     """
     Scores the extracted text and lists the keywords found.
-    
-    Args:
-        text (str): The full text extracted via OCR.
-        
-    Returns:
-        tuple: (float risk_score, list keywords_found)
     """
     high_risk_keywords = [
         "nodule", "abnormal cell", "squamous", "carcinoma", "malignant", 
@@ -69,25 +50,16 @@ def score_text_for_risk(text):
     for keyword in high_risk_keywords:
         if keyword in text:
             score += 0.1 
-            keywords_found.append(keyword.title()) # Capitalize for display
+            keywords_found.append(keyword.title())
             
-    # Cap the score at 1.0
     return min(score, 1.0), keywords_found
 
-# Example usage (Optional, can be removed once integrated into app.py)
+# Example test
 if __name__ == '__main__':
-    # You would test this with a sample image report
-    # print(extract_keywords_from_report('path/to/sample_report.jpg'))
-    pass
-
-if __name__ == '__main__':
-    # ⚠️ Replace 'test_report.png' with the actual file name
     test_file_path = 'test_report.png' 
-
     if os.path.exists(test_file_path):
         extracted_text = extract_keywords_from_report(test_file_path)
         risk_score = score_text_for_risk(extracted_text)
-
         print("--- OCR Test Results ---")
         print(f"Extracted Text: {extracted_text}")
         print(f"Calculated Risk Score: {risk_score}")