save

utils/comparator.py

@@ -1,37 +1,52 @@
 import difflib
 import re
+from typing import List, Tuple
 
 
-def highlight_fuzzy_diff(user_text, original_text):
-    original_words = re.findall(r"\w+|[^\w\s]", original_text)
-    user_words = re.findall(r"\w+|[^\w\s]", user_text)
+def preprocess(text: str) -> List[str]:
+    return re.sub(r"[^\w\s]", "", text.lower()).split()
 
-    matcher = difflib.SequenceMatcher(None, user_words, original_words)
+
+def highlight_fuzzy_diff(user_text: str, original_text: str) -> Tuple[str, int]:
+    original_words_raw = re.findall(r"\w+|[^\w\s]", original_text)
+    user_words_raw = re.findall(r"\w+|[^\w\s]", user_text)
+
+    original_words_clean = preprocess(original_text)
+    user_words_clean = preprocess(user_text)
+
+    matcher = difflib.SequenceMatcher(None, user_words_clean, original_words_clean)
     highlighted = []
     correct_count = 0
     total_count = 0
 
+    original_pointer = 0
+
     for op, i1, i2, j1, j2 in matcher.get_opcodes():
-        segment = original_words[j1:j2]
-        if op == 'equal':
-            for word in segment:
-                if re.match(r"\w+", word):  # тільки слова рахуються в score
+        segment_clean = original_words_clean[j1:j2]
+        segment_raw = []
+
+        # Зібрати наступні слова з raw-представлення (з пунктуацією)
+        while len(segment_raw) < len(segment_clean) and original_pointer < len(original_words_raw):
+            if re.match(r"\w+", original_words_raw[original_pointer]):
+                segment_raw.append(original_words_raw[original_pointer])
+            original_pointer += 1
+
+        for word in segment_raw:
+            if re.match(r"\w+", word):
+                total_count += 1
+                if op == 'equal':
                     correct_count += 1
-                    total_count += 1
-                highlighted.append(
-                    f"<span style='background-color:#e6ffe6; color:#006600; padding:4px 10px; "
-                    f"margin:4px; border-radius:999px; font-weight:500; display:inline-block;'>{word}</span>"
-                )
-        else:
-            for word in segment:
-                if re.match(r"\w+", word):
-                    total_count += 1
+                    highlighted.append(
+                        f"<span style='background-color:#e6ffe6; color:#006600; padding:4px 10px; "
+                        f"margin:4px; border-radius:999px; font-weight:500; display:inline-block;'>{word}</span>"
+                    )
+                else:
                     highlighted.append(
                         f"<span style='background-color:#ffe6e6; color:#990000; padding:4px 10px; "
                         f"margin:4px; border-radius:999px; font-weight:500; display:inline-block;'>...</span>"
                     )
-                else:
-                    highlighted.append(word)
+            else:
+                highlighted.append(word)
 
     score_percent = round((correct_count / total_count) * 100) if total_count > 0 else 0
     return " ".join(highlighted), score_percent