save

utils/comparator.py

@@ -1,66 +1,37 @@
+import difflib
 import re
 
-from rapidfuzz import fuzz
 
+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):
-    return re.sub(r"[^\w\s]", "", text.lower()).split()
-
-
-def is_position_acceptable(expected_index, actual_index, max_shift=1):
-    return abs(expected_index - actual_index) <= max_shift
-
-
-def highlight_fuzzy_diff(user_text, original_text, threshold=92, max_position_shift=1):
-    user_words = preprocess(user_text)
-    original_words = preprocess(original_text)
-
-    used_user_indices = set()
+    matcher = difflib.SequenceMatcher(None, user_words, original_words)
     highlighted = []
     correct_count = 0
-
-    original_display_words = re.findall(r"\w+|[^\w\s]", original_text)
-
-    display_index = 0
-
-    for i, orig_word in enumerate(original_words):
-        best_score = 0
-        best_index = -1
-        for j, user_word in enumerate(user_words):
-            if j in used_user_indices:
-                continue
-            if not is_position_acceptable(i, j, max_position_shift):
-                continue
-            score = fuzz.ratio(orig_word, user_word)
-            if score > best_score:
-                best_score = score
-                best_index = j
-
-        while display_index < len(original_display_words):
-            display_candidate = original_display_words[display_index]
-            if re.match(r"\w+", display_candidate):
-                display_word = display_candidate
-                display_index += 1
-                break
-            else:
-                highlighted.append(display_candidate)
-                display_index += 1
-        else:
-            display_word = orig_word
-
-        if best_score >= threshold and best_index != -1:
-            used_user_indices.add(best_index)
-            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;'>{display_word}</span>"
-            )
-            correct_count += 1
+    total_count = 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
+                    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:
-            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>"
-            )
-
-    score_percent = round((correct_count / len(original_words)) * 100) if original_words else 0
+            for word in segment:
+                if re.match(r"\w+", word):
+                    total_count += 1
+                    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)
+
+    score_percent = round((correct_count / total_count) * 100) if total_count > 0 else 0
     return " ".join(highlighted), score_percent
-