fix(transliterate): romanize royin words containing ฤ

pythainlp/transliterate/royin.py

@@ -274,7 +274,10 @@ def _romanize(word: str) -> str:
         return ""
 
     word = _replace_vowels(_normalize(word))
-    consonants = _RE_CONSONANT.findall(word)
+    # Use the same membership test as _replace_consonants(), which treats
+    # every key of _CONSONANTS as a consonant. _RE_CONSONANT only matches
+    # thai_consonants, so it misses ฤ and the two lists fall out of sync.
+    consonants = [c for c in word if c in _CONSONANTS]
 
     # 2-character word, all consonants
     if len(word) == 2 and len(consonants) == 2:

tests/core/test_transliterate.py

@@ -38,6 +38,13 @@
     "ชารินทร์": "charin",
 }
 
+# words containing ฤ (U+0E24 THAI CHARACTER RU), see issue #1444
+RU_TESTS = {
+    "ฤก": "rueok",
+    "ฤดู": "rueadu",
+    "ฤทธิ์": "rueot",
+}
+
 # these are set of two-syllable words,
 # to test if the transliteration/romanization is consistent, say
 # romanize(1+2) = romanize(1) + romanize(2)
@@ -60,6 +67,12 @@ def test_romanize_royin_basic(self):
         for word, expect in BASIC_TESTS.items():
             self.assertEqual(romanize(word, engine="royin"), expect)  # type: ignore[arg-type]
 
+    def test_romanize_royin_ru(self):
+        # ฤ (U+0E24) is a key of _CONSONANTS but is not in thai_consonants,
+        # so it used to desynchronise the consonant list and raise IndexError.
+        for word, expect in RU_TESTS.items():
+            self.assertEqual(romanize(word, engine="royin"), expect)  # type: ignore[arg-type]
+
     def test_romanize_royin_consistency(self):
         for word, part1, part2 in CONSISTENCY_TESTS:
             self.assertEqual(