fix: log warning on unmask regex compile failure (v2.6.5)

CHANGELOG.md

@@ -4,6 +4,13 @@ All notable changes to this project will be documented in this file.
 
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/).
 
+## [2.6.5] - 2026-06
+
+### Changed
+- The combined-unmask-regex fallback (v2.6.4) now logs a `WARNING` when the
+  alternation regex fails to compile, instead of silently switching to the
+  slow per-mask path
+
 ## [2.6.4] - 2026-06
 
 ### Performance

data_masking.py

@@ -30,7 +30,7 @@
 # Re-exports from masking package for backward compatibility
 # ============================================================================
 
-__version__ = "2.6.4"
+__version__ = "2.6.5"
 
 from masking.constants import (
     __version__, __author__, __contact__, __phone__, __license__, __year__,

masking/constants.py

@@ -26,7 +26,7 @@
 # ============================================================================
 # МЕТАДАНІ
 # ============================================================================
-__version__ = "2.6.4"
+__version__ = "2.6.5"
 __author__ = "Vladyslav V. Prodan"
 __contact__ = "github.com/click0"
 __phone__ = "+38(099)6053340"

tests/test_unmask.py

@@ -532,6 +532,29 @@ def test_empty_mapping_noop(self):
         assert restored == "текст без масок"
         assert stats["restored_count"] == 0
 
+    def test_regex_compile_failure_logs_warning_and_falls_back(self, monkeypatch, caplog):
+        # Якщо об'єднаний regex не компілюється — попередження в лог
+        # і коректний результат через повільний шлях
+        import re as _re
+        import unmasking.engine as eng
+
+        real_compile = _re.compile
+
+        def boom(pattern, *a, **k):
+            if "коваленко" in pattern.lower():
+                raise _re.error("forced failure")
+            return real_compile(pattern, *a, **k)
+
+        monkeypatch.setattr(eng.re, "compile", boom)
+        masking_map = self._map({
+            "surname": {"петренко": {"masked_as": "коваленко", "instances": [1]}}
+        })
+        with caplog.at_level("WARNING"):
+            restored, stats = eng.unmask_other_data("коваленко тут", masking_map)
+        assert restored == "петренко тут"
+        assert stats["restored_count"] == 1
+        assert any("falling back" in r.message for r in caplog.records)
+
 
 if __name__ == "__main__":
     pytest.main([__file__, "-v"])

unmasking/cli.py

@@ -58,7 +58,7 @@
 # ============================================================================
 # МЕТАДАНІ
 # ============================================================================
-__version__ = "2.6.4"
+__version__ = "2.6.5"
 
 
 def main():

unmasking/engine.py

@@ -7,9 +7,12 @@
 Extracted from unmask_data.py during the package refactoring (v2.5.0).
 """
 
+import logging
 import re
 from typing import Any, Dict, List, Tuple
 
+_logger = logging.getLogger(__name__)
+
 from rank_data import (
     RANK_DECLENSIONS,
     RANK_FEMININE_MAP,
@@ -193,8 +196,14 @@ def unmask_other_data(masked_text: str, masking_map: Dict) -> Tuple[str, Dict]:
             r'(?<!\w)(' + '|'.join(re.escape(m) for m in masks_sorted) + r')(?!\w)',
             re.IGNORECASE,
         )
-    except re.error:
-        # Запасний шлях (екзотичні символи в масках)
+    except re.error as e:
+        # Запасний шлях (екзотичні символи в масках). Не тихо: логуємо,
+        # бо повільний шлях значно дорожчий на великих документах.
+        _logger.warning(
+            "Combined unmask regex failed to compile (%s); falling back to "
+            "per-mask scan over %d masks — this is much slower on large files",
+            e, len(instance_map),
+        )
         return _unmask_other_data_slow(restored_text, instance_map)
 
     seen = {}            # lower(маска) -> скільки входжень уже зустріли