Update app.py

app.py

@@ -2,19 +2,20 @@ import gradio as gr
 import google.generativeai as genai
 from PIL import Image
 import os
+import re
 
 # Direct API key placement
 API_KEY = 'AIzaSyBMZrhMXRpQKp7M-JcN2Qk73afeta5Mv5Y'
 
-def extract_malayalam_text(image):
+def extract_mixed_text(image):
     """
-    Extract text specifically in Malayalam from the image
+    Extract Malayalam letters, English letters, and numbers from the image
     
     Args:
         image (PIL.Image): Uploaded image
     
     Returns:
-        str: Extracted Malayalam text
+        str: Extracted text
     """
     # Validate input
     if image is None:
@@ -27,48 +28,60 @@ def extract_malayalam_text(image):
         # Use Gemini 2.0 Flash model
         model = genai.GenerativeModel('gemini-2.0-flash')
         
-        # Highly specific prompt for Malayalam text extraction
+        # Comprehensive prompt for text extraction
         response = model.generate_content(
             [
-                "CRITICAL INSTRUCTIONS FOR MALAYALAM TEXT EXTRACTION: "
-                "1. This image contains text in Malayalam script ONLY. "
-                "2. Extract ONLY Malayalam characters and script. "
-                "3. Use pure Malayalam Unicode characters. "
-                "4. If the text is in another language, state 'No Malayalam text found'. "
-                "5. Preserve exact Malayalam script, including any special characters. "
-                "6. DO NOT translate or convert to any other script. "
-                "7. If there are multiple lines, preserve line breaks. "
-                "8. Focus EXCLUSIVELY on Malayalam script.", 
+                "CRITICAL INSTRUCTIONS FOR TEXT EXTRACTION: "
+                "1. Extract ALL text from this image. "
+                "2. Include: "
+                "   - Malayalam letters "
+                "   - English letters "
+                "   - Numbers "
+                "3. Preserve original formatting and line breaks. "
+                "4. Maintain the exact order of characters. "
+                "5. Do not modify or translate the text.", 
                 image
             ],
             generation_config=genai.types.GenerationConfig(
-                temperature=0.1,  # Ultra-low temperature for precise extraction
-                max_output_tokens=500  # Increased to capture more text
+                temperature=0.1,  # Precise extraction
+                max_output_tokens=500  # Capture full text
             )
         )
         
         # Extract text from response
         extracted_text = response.text
         
-        # Validate Malayalam characters
-        malayalam_chars = [char for char in extracted_text if '\u0D00' <= char <= '\u0D7F']
+        # Filter to keep only Malayalam letters, English letters, and numbers
+        filtered_text = ''.join(
+            char for char in extracted_text 
+            if (
+                # Malayalam Unicode range
+                ('\u0D00' <= char <= '\u0D7F') or 
+                # English letters (uppercase and lowercase)
+                char.isalpha() or 
+                # Numbers
+                char.isdigit() or 
+                # Preserve spaces and basic punctuation
+                char.isspace() or char in '.,/-'
+            )
+        )
         
-        # Check if extracted text contains Malayalam characters
-        if not malayalam_chars:
-            return "No Malayalam text detected. Please verify the image contains Malayalam script."
+        # Validate text extraction
+        if not filtered_text.strip():
+            return "No valid text detected. Please check the image quality."
         
-        return extracted_text
+        return filtered_text
     
     except Exception as e:
         return f"An error occurred: {str(e)}"
 
 # Create Gradio Interface
 demo = gr.Interface(
-    fn=extract_malayalam_text,
-    inputs=gr.Image(type="pil", label="Upload Malayalam Text Image"),
-    outputs=gr.Textbox(label="Extracted Malayalam Text", lines=10),
-    title="Malayalam Text Extractor",
-    description="Upload an image containing Malayalam text for precise extraction."
+    fn=extract_mixed_text,
+    inputs=gr.Image(type="pil", label="Upload Image with Text"),
+    outputs=gr.Textbox(label="Extracted Text", lines=10),
+    title="Text Extractor",
+    description="Upload an image to extract Malayalam letters, English letters, and numbers."
 )
 
 # Launch the app