update the compute changes of the handler.py

handler.py

@@ -44,26 +44,27 @@ class EndpointHandler:
             return decoded
             
     def compute_changes(self, original, enhanced):
-        # Your existing compute_changes logic
         changes = []
-        matcher = SequenceMatcher(None, original.split(), enhanced.split())
+        matcher = SequenceMatcher(None, original, enhanced)  # char-level, not token-level
+
         for tag, i1, i2, j1, j2 in matcher.get_opcodes():
             if tag in ("replace", "insert", "delete"):
-                original_phrase = " ".join(original.split()[i1:i2])
-                new_phrase = " ".join(enhanced.split()[j1:j2])
+                original_phrase = original[i1:i2]
+                new_phrase = enhanced[j1:j2]
                 changes.append({
                     "original_phrase": original_phrase,
                     "new_phrase": new_phrase,
                     "char_start": i1,
                     "char_end": i2,
-                    "token_start": i1,
-                    "token_end": i2,
+                    "token_start": None,  # not token-based anymore
+                    "token_end": None,
                     "explanation": f"{tag} change",
-                    "error_type": "",
-                    "tip": ""
+                    "error_type": "whitespace" if original_phrase.isspace() or new_phrase.isspace() else "",
+                    "tip": "Avoid extra spaces between words." if original_phrase.isspace() or new_phrase.isspace() else ""
                 })
         return changes
 
+
     def __call__(self, inputs):
         # This method is the main entry point for the Hugging Face Endpoint.