Update app.py

app.py

@@ -3,59 +3,95 @@ import pytesseract
 from pytesseract import Output
 from PIL import Image
 
-# Ensure Tesseract is found (if needed for specific environments, otherwise leave default)
-# pytesseract.pytesseract.tesseract_cmd = r'/usr/bin/tesseract'
-
 def validate_chat_origin(image):
     if image is None:
         return {"valid": False, "error": "No image uploaded"}
 
     try:
+        # Get data with bounding boxes
         data = pytesseract.image_to_data(image, output_type=Output.DICT)
     except Exception as e:
         return {"valid": False, "error": str(e)}
 
     width, height = image.size
-    midpoint = width / 2
     
-    valid_lines = []
-    ignore_words = ["end-to-end", "encrypted", "business", "account", "block", "today", "joined", "march"]
+    # --- CONFIGURATION ---
+    # 1. Ignore the top 12% of the screen (Status Bar + Header Name)
+    header_margin = height * 0.12
+    
+    # 2. Strict Stranger Zone: Incoming messages MUST start in the first 15% of width
+    #    (e.g., if screen is 1000px wide, message must start within 0-150px)
+    stranger_margin_limit = width * 0.15
 
+    # 3. Ignore words (Case insensitive)
+    ignore_words = [
+        "end-to-end", "encrypted", "business", "account", "block", 
+        "add", "contacts", "today", "joined", "march", "april", "may", "june",
+        "report", "spam", "copied", "messages", "calls", "group"
+    ]
+
+    valid_lines = []
     n_boxes = len(data['text'])
+    
     for i in range(n_boxes):
         text = data['text'][i].strip()
-        if not text or int(data['conf'][i]) < 50:
+        conf = int(data['conf'][i])
+        
+        # Basic filter: Empty text or low confidence
+        if not text or conf < 40:
+            continue
+            
+        # Filter A: Ignore System Header (Top of screen)
+        # This prevents "13:25" or "68%" in status bar from triggering "Right Aligned"
+        if data['top'][i] < header_margin:
             continue
+            
+        # Filter B: Ignore specific system words
         if any(ignored in text.lower() for ignored in ignore_words):
             continue
         
+        # Filter C: purely numeric small chunks often misread (like timestamps isolated)
+        # If it's just digits and very short, skip (unless it's a long phone number)
+        if text.isdigit() and len(text) < 6:
+            continue
+
         valid_lines.append({
             "text": text,
             "top": data['top'][i],
-            "left": data['left'][i]
+            "left": data['left'][i],
+            "width": data['width'][i]
         })
 
+    # Sort by vertical position (Top to Bottom)
     valid_lines.sort(key=lambda x: x['top'])
 
     if not valid_lines:
-        return {"valid": False, "reason": "No text detected"}
+        return {"valid": False, "reason": "No readable chat text found."}
 
-    first_message = valid_lines[0]
+    # --- DECISION LOGIC ---
+    # Pick the very first detected text block
+    first_msg = valid_lines[0]
+    
+    # Check: Does it start on the far left?
+    # Incoming messages align Left. Outgoing align Right. Centered text aligns Middle.
+    # Only "Left" is valid.
+    
+    is_stranger = first_msg['left'] < stranger_margin_limit
     
-    # Check alignment: Left = Stranger (Valid), Right = Me (Invalid)
-    is_left_aligned = first_message['left'] < midpoint
+    debug_msg = f"Detected: '{first_msg['text']}' at X={first_msg['left']} (Limit: {stranger_margin_limit})"
 
-    if is_left_aligned:
-        return {"valid": True, "reason": "First message is from stranger (Left aligned)"}
+    if is_stranger:
+        return {"valid": True, "reason": "First message is from stranger.", "debug": debug_msg}
     else:
-        return {"valid": False, "reason": "First message is from you (Right aligned)"}
+        # If it's not on the left, it's either ME (Right) or SYSTEM (Center)
+        return {"valid": False, "reason": "First message is NOT from stranger (Right/Center aligned).", "debug": debug_msg}
 
-# Define the Interface with JSON output
+# Interface
 iface = gr.Interface(
     fn=validate_chat_origin,
     inputs=gr.Image(type="pil"),
-    outputs=gr.JSON(),  # <--- This ensures the API returns JSON
-    title="WhatsApp Validator API"
+    outputs=gr.JSON(),
+    title="WhatsApp Validator V2"
 )
 
 if __name__ == "__main__":