Add more endmarks incl. \uFF1F (#146)

* Add more endmarks incl. \uFF1F Fix #145 Signed-off-by: Hiroshi Miura <[email protected]> * PEP8/Black Signed-off-by: Hiroshi Miura <[email protected]>
miurahr · Jul 10, 2021 · 921f159 · 921f159
1 parent d67bd5b
commit 921f159
Show file tree

Hide file tree

Showing 2 changed files with 17 additions and 1 deletion.
diff --git a/src/pykakasi/properties.py b/src/pykakasi/properties.py
@@ -60,7 +60,7 @@ class Ch:
     zenkaku_number_nine = 0xFF1A
     zenkaku_A = 0xFF21
     zenkaku_a = 0xFF41
-    endmark = ")]!,.,\u3001\u3002"
+    endmark = ")]!,.,\u3001\u3002\uff1f\uff10\uff1e\uff1c"
     long_symbols = "\u30FC\u2015\u2212\uFF70"  # "ー  ―  −  ｰ "
     # _UNCHECKED_LONG_SYMBOLS: str = "\u002D\u2010\u2011\u2013\u2014" # "-  ‐ ‑ – —"
 

diff --git a/tests/test_pykakasi_structured.py b/tests/test_pykakasi_structured.py
@@ -406,6 +406,22 @@ def test_kakasi_unidic_noun(case, expected):
     assert result[0]["hira"] == expected
 
 
+def test_issue_145():
+    case = "見えますか？"
+    expected = {
+        "orig": "ますか？",
+        "hira": "ますか？",
+        "kana": "マスカ？",
+        "hepburn": "masuka?",
+        "kunrei": "masuka?",
+        "passport": "masuka?",
+    }
+    kakasi = pykakasi.Kakasi()
+    result = kakasi.convert(case)
+    assert result[1]["orig"] == expected["orig"]
+    assert result[1]["hira"] == expected["hira"]
+
+
 @pytest.mark.parametrize(
     "case, expected",
     [