ENH: Windowsでエンジンの多重起動を可能にする (#1514)

Co-authored-by: Hiroshiba <[email protected]>
Co-authored-by: Hiroshiba Kazuyuki <[email protected]>

Author: sabonerune

poetry.lock

@@ -52,7 +52,7 @@ uvicorn = "^0.32.1"
 soundfile = "^0.12.1"
 pyyaml = "^6.0.1"
 pyworld = "^0.3.0"
-pyopenjtalk = { git = "https://github.com/VOICEVOX/pyopenjtalk", rev = "b35fc89fe42948a28e33aed886ea145a51113f88" }
+pyopenjtalk = { git = "https://github.com/VOICEVOX/pyopenjtalk", rev = "0fcb731c94555e8d160d18e7f1a4d005b2e8e852" }
 semver = "^3.0.0"
 platformdirs = "^4.2.0"
 soxr = "^0.5.0"

pyproject.toml

@@ -20,7 +20,7 @@ pydantic-core==2.23.4 ; python_version >= "3.11" and python_version < "3.12"
 pydantic==2.9.2 ; python_version >= "3.11" and python_version < "3.12"
 pyinstaller-hooks-contrib==2024.10 ; python_version >= "3.11" and python_version < "3.12"
 pyinstaller==5.13.2 ; python_version >= "3.11" and python_version < "3.12"
-pyopenjtalk @ git+https://github.com/VOICEVOX/pyopenjtalk@b35fc89fe42948a28e33aed886ea145a51113f88 ; python_version >= "3.11" and python_version < "3.12"
+pyopenjtalk @ git+https://github.com/VOICEVOX/pyopenjtalk@0fcb731c94555e8d160d18e7f1a4d005b2e8e852 ; python_version >= "3.11" and python_version < "3.12"
 python-multipart==0.0.17 ; python_version >= "3.11" and python_version < "3.12"
 pywin32-ctypes==0.2.3 ; python_version >= "3.11" and python_version < "3.12" and sys_platform == "win32"
 pyworld==0.3.4 ; python_version >= "3.11" and python_version < "3.12"

requirements-build.txt

@@ -74,7 +74,7 @@ pydantic-core==2.23.4 ; python_version >= "3.11" and python_version < "3.12"
 pydantic==2.9.2 ; python_version >= "3.11" and python_version < "3.12"
 pyflakes==3.2.0 ; python_version >= "3.11" and python_version < "3.12"
 pygments==2.18.0 ; python_version >= "3.11" and python_version < "3.12"
-pyopenjtalk @ git+https://github.com/VOICEVOX/pyopenjtalk@b35fc89fe42948a28e33aed886ea145a51113f88 ; python_version >= "3.11" and python_version < "3.12"
+pyopenjtalk @ git+https://github.com/VOICEVOX/pyopenjtalk@0fcb731c94555e8d160d18e7f1a4d005b2e8e852 ; python_version >= "3.11" and python_version < "3.12"
 pyproject-hooks==1.2.0 ; python_version >= "3.11" and python_version < "3.12"
 pysen==0.11.0 ; python_version >= "3.11" and python_version < "3.12"
 pytest==8.3.3 ; python_version >= "3.11" and python_version < "3.12"

requirements-dev.txt

@@ -14,7 +14,7 @@ platformdirs==4.3.6 ; python_version >= "3.11" and python_version < "3.12"
 pycparser==2.22 ; python_version >= "3.11" and python_version < "3.12"
 pydantic-core==2.23.4 ; python_version >= "3.11" and python_version < "3.12"
 pydantic==2.9.2 ; python_version >= "3.11" and python_version < "3.12"
-pyopenjtalk @ git+https://github.com/VOICEVOX/pyopenjtalk@b35fc89fe42948a28e33aed886ea145a51113f88 ; python_version >= "3.11" and python_version < "3.12"
+pyopenjtalk @ git+https://github.com/VOICEVOX/pyopenjtalk@0fcb731c94555e8d160d18e7f1a4d005b2e8e852 ; python_version >= "3.11" and python_version < "3.12"
 python-multipart==0.0.17 ; python_version >= "3.11" and python_version < "3.12"
 pyworld==0.3.4 ; python_version >= "3.11" and python_version < "3.12"
 pyyaml==6.0.2 ; python_version >= "3.11" and python_version < "3.12"

requirements.txt

@@ -44,7 +44,6 @@ def app_params(tmp_path: Path) -> dict[str, Any]:
     user_dict = UserDictionary(
         default_dict_path=_copy_under_dir(DEFAULT_DICT_PATH, tmp_path),
         user_dict_path=_generate_user_dict(tmp_path),
-        compiled_dict_path=tmp_path / "user.dic",
     )
 
     engine_manifest = load_manifest(engine_manifest_path())

test/e2e/conftest.py

@@ -98,8 +98,7 @@ def test_create_word() -> None:
 def test_apply_word_without_json(tmp_path: Path) -> None:
 
     user_dict = UserDictionary(
-        user_dict_path=tmp_path / "test_apply_word_without_json.json",
-        compiled_dict_path=tmp_path / "test_apply_word_without_json.dic",
+        user_dict_path=tmp_path / "test_apply_word_without_json.json"
     )
     user_dict.apply_word(
         WordProperty(surface="test", pronunciation="テスト", accent_type=1)
@@ -119,10 +118,7 @@ def test_apply_word_with_json(tmp_path: Path) -> None:
     user_dict_path.write_text(
         json.dumps(valid_dict_dict_json, ensure_ascii=False), encoding="utf-8"
     )
-    user_dict = UserDictionary(
-        user_dict_path=user_dict_path,
-        compiled_dict_path=tmp_path / "test_apply_word_with_json.dic",
-    )
+    user_dict = UserDictionary(user_dict_path=user_dict_path)
     user_dict.apply_word(
         WordProperty(surface="test2", pronunciation="テストツー", accent_type=3)
     )
@@ -141,10 +137,7 @@ def test_rewrite_word_invalid_id(tmp_path: Path) -> None:
     user_dict_path.write_text(
         json.dumps(valid_dict_dict_json, ensure_ascii=False), encoding="utf-8"
     )
-    user_dict = UserDictionary(
-        user_dict_path=user_dict_path,
-        compiled_dict_path=(tmp_path / "test_rewrite_word_invalid_id.dic"),
-    )
+    user_dict = UserDictionary(user_dict_path=user_dict_path)
     with pytest.raises(UserDictInputError):
         user_dict.rewrite_word(
             "c2be4dc5-d07d-4767-8be1-04a1bb3f05a9",
@@ -157,10 +150,7 @@ def test_rewrite_word_valid_id(tmp_path: Path) -> None:
     user_dict_path.write_text(
         json.dumps(valid_dict_dict_json, ensure_ascii=False), encoding="utf-8"
     )
-    user_dict = UserDictionary(
-        user_dict_path=user_dict_path,
-        compiled_dict_path=tmp_path / "test_rewrite_word_valid_id.dic",
-    )
+    user_dict = UserDictionary(user_dict_path=user_dict_path)
     user_dict.rewrite_word(
         "aab7dda2-0d97-43c8-8cb7-3f440dab9b4e",
         WordProperty(surface="test2", pronunciation="テストツー", accent_type=2),
@@ -178,10 +168,7 @@ def test_delete_word_invalid_id(tmp_path: Path) -> None:
     user_dict_path.write_text(
         json.dumps(valid_dict_dict_json, ensure_ascii=False), encoding="utf-8"
     )
-    user_dict = UserDictionary(
-        user_dict_path=user_dict_path,
-        compiled_dict_path=tmp_path / "test_delete_word_invalid_id.dic",
-    )
+    user_dict = UserDictionary(user_dict_path=user_dict_path)
     with pytest.raises(UserDictInputError):
         user_dict.delete_word(word_uuid="c2be4dc5-d07d-4767-8be1-04a1bb3f05a9")
 
@@ -191,10 +178,7 @@ def test_delete_word_valid_id(tmp_path: Path) -> None:
     user_dict_path.write_text(
         json.dumps(valid_dict_dict_json, ensure_ascii=False), encoding="utf-8"
     )
-    user_dict = UserDictionary(
-        user_dict_path=user_dict_path,
-        compiled_dict_path=tmp_path / "test_delete_word_valid_id.dic",
-    )
+    user_dict = UserDictionary(user_dict_path=user_dict_path)
     user_dict.delete_word(word_uuid="aab7dda2-0d97-43c8-8cb7-3f440dab9b4e")
     assert len(user_dict.read_dict()) == 0
 
@@ -218,13 +202,10 @@ def test_priority() -> None:
 
 def test_import_dict(tmp_path: Path) -> None:
     user_dict_path = tmp_path / "test_import_dict.json"
-    compiled_dict_path = tmp_path / "test_import_dict.dic"
     user_dict_path.write_text(
         json.dumps(valid_dict_dict_json, ensure_ascii=False), encoding="utf-8"
     )
-    user_dict = UserDictionary(
-        user_dict_path=user_dict_path, compiled_dict_path=compiled_dict_path
-    )
+    user_dict = UserDictionary(user_dict_path=user_dict_path)
     user_dict.import_user_dict(
         {"b1affe2a-d5f0-4050-926c-f28e0c1d9a98": import_word}, override=False
     )
@@ -236,13 +217,10 @@ def test_import_dict(tmp_path: Path) -> None:
 
 def test_import_dict_no_override(tmp_path: Path) -> None:
     user_dict_path = tmp_path / "test_import_dict_no_override.json"
-    compiled_dict_path = tmp_path / "test_import_dict_no_override.dic"
     user_dict_path.write_text(
         json.dumps(valid_dict_dict_json, ensure_ascii=False), encoding="utf-8"
     )
-    user_dict = UserDictionary(
-        user_dict_path=user_dict_path, compiled_dict_path=compiled_dict_path
-    )
+    user_dict = UserDictionary(user_dict_path=user_dict_path)
     user_dict.import_user_dict(
         {"aab7dda2-0d97-43c8-8cb7-3f440dab9b4e": import_word}, override=False
     )
@@ -253,13 +231,10 @@ def test_import_dict_no_override(tmp_path: Path) -> None:
 
 def test_import_dict_override(tmp_path: Path) -> None:
     user_dict_path = tmp_path / "test_import_dict_override.json"
-    compiled_dict_path = tmp_path / "test_import_dict_override.dic"
     user_dict_path.write_text(
         json.dumps(valid_dict_dict_json, ensure_ascii=False), encoding="utf-8"
     )
-    user_dict = UserDictionary(
-        user_dict_path=user_dict_path, compiled_dict_path=compiled_dict_path
-    )
+    user_dict = UserDictionary(user_dict_path=user_dict_path)
     user_dict.import_user_dict(
         {"aab7dda2-0d97-43c8-8cb7-3f440dab9b4e": import_word}, override=True
     )
@@ -268,15 +243,12 @@ def test_import_dict_override(tmp_path: Path) -> None:
 
 def test_import_invalid_word(tmp_path: Path) -> None:
     user_dict_path = tmp_path / "test_import_invalid_dict.json"
-    compiled_dict_path = tmp_path / "test_import_invalid_dict.dic"
     invalid_accent_associative_rule_word = deepcopy(import_word)
     invalid_accent_associative_rule_word.accent_associative_rule = "invalid"
     user_dict_path.write_text(
         json.dumps(valid_dict_dict_json, ensure_ascii=False), encoding="utf-8"
     )
-    user_dict = UserDictionary(
-        user_dict_path=user_dict_path, compiled_dict_path=compiled_dict_path
-    )
+    user_dict = UserDictionary(user_dict_path=user_dict_path)
     with pytest.raises(AssertionError):
         user_dict.import_user_dict(
             {
@@ -299,10 +271,7 @@ def test_import_invalid_word(tmp_path: Path) -> None:
 
 def test_update_dict(tmp_path: Path) -> None:
     user_dict_path = tmp_path / "test_update_dict.json"
-    compiled_dict_path = tmp_path / "test_update_dict.dic"
-    user_dict = UserDictionary(
-        user_dict_path=user_dict_path, compiled_dict_path=compiled_dict_path
-    )
+    user_dict = UserDictionary(user_dict_path=user_dict_path)
     user_dict.update_dict()
     test_text = "テスト用の文字列"
     success_pronunciation = "デフォルトノジショデハゼッタイニセイセイサレナイヨミ"

test/unit/user_dict/test_user_dict.py

@@ -50,7 +50,6 @@ def func(*args: Any, **kw: Any) -> Any:
 # デフォルトのファイルパス
 DEFAULT_DICT_PATH: Final = resource_dir / "default.csv"  # VOICEVOXデフォルト辞書
 _USER_DICT_PATH: Final = save_dir / "user_dict.json"  # ユーザー辞書
-_COMPILED_DICT_PATH: Final = save_dir / "user.dic"  # コンパイル済み辞書
 
 
 # 同時書き込みの制御
@@ -61,14 +60,69 @@ def func(*args: Any, **kw: Any) -> Any:
 _save_format_dict_adapter = TypeAdapter(dict[str, SaveFormatUserDictWord])
 
 
+def _delete_file_on_close(file_path: Path) -> None:
+    """
+    ファイルのハンドルが全て閉じたときにファイルを削除する。OpenJTalk用のカスタム辞書用。
+
+    WindowsではCreateFileW関数で`FILE_FLAG_DELETE_ON_CLOSE`を付けてすぐに閉じることで、
+    `FILE_SHARE_DELETE`を付けて開かれているファイルのハンドルが全て閉じた時に削除されるようにする。
+
+    Windows以外では即座にファイルを削除する。
+    """
+    if sys.platform == "win32":
+        import ctypes
+        from ctypes.wintypes import DWORD, HANDLE, LPCWSTR
+
+        _CreateFileW = ctypes.windll.kernel32.CreateFileW
+        _CreateFileW.argtypes = [
+            LPCWSTR,
+            DWORD,
+            DWORD,
+            ctypes.c_void_p,
+            DWORD,
+            DWORD,
+            HANDLE,
+        ]
+        _CreateFileW.restype = HANDLE
+        _CloseHandle = ctypes.windll.kernel32.CloseHandle
+        _CloseHandle.argtypes = [HANDLE]
+
+        _FILE_SHARE_DELETE = 0x00000004
+        _FILE_SHARE_READ = 0x00000001
+        _OPEN_EXISTING = 3
+        _FILE_FLAG_DELETE_ON_CLOSE = 0x04000000
+        _INVALID_HANDLE_VALUE = HANDLE(-1).value
+
+        h_file = _CreateFileW(
+            str(file_path),
+            0,
+            _FILE_SHARE_DELETE | _FILE_SHARE_READ,
+            None,
+            _OPEN_EXISTING,
+            _FILE_FLAG_DELETE_ON_CLOSE,
+            None,
+        )
+        if h_file == _INVALID_HANDLE_VALUE:
+            raise RuntimeError(
+                f"Failed to CreateFileW for {file_path}"
+            ) from ctypes.WinError()
+
+        result = _CloseHandle(h_file)
+        if result == 0:
+            raise RuntimeError(
+                f"Failed to CloseHandle for {file_path}"
+            ) from ctypes.WinError()
+    else:
+        file_path.unlink()
+
+
 class UserDictionary:
     """ユーザー辞書"""
 
     def __init__(
         self,
         default_dict_path: Path = DEFAULT_DICT_PATH,
         user_dict_path: Path = _USER_DICT_PATH,
-        compiled_dict_path: Path = _COMPILED_DICT_PATH,
     ) -> None:
         """
         Parameters
@@ -77,12 +131,9 @@ def __init__(
             デフォルト辞書ファイルのパス
         user_dict_path : Path
             ユーザー辞書ファイルのパス
-        compiled_dict_path : Path
-            コンパイル済み辞書ファイルのパス
         """
         self._default_dict_path = default_dict_path
         self._user_dict_path = user_dict_path
-        self._compiled_dict_path = compiled_dict_path
         self.update_dict()
 
     @mutex_wrapper(mutex_user_dict)
@@ -99,14 +150,14 @@ def _write_to_json(self, user_dict: dict[str, UserDictWord]) -> None:
     def update_dict(self) -> None:
         """辞書を更新する。"""
         default_dict_path = self._default_dict_path
-        compiled_dict_path = self._compiled_dict_path
+        user_dict_path = self._user_dict_path
 
         random_string = uuid4()
-        tmp_csv_path = compiled_dict_path.with_suffix(
-            f".dict_csv-{random_string}.tmp"
+        tmp_csv_path = user_dict_path.with_name(
+            f"user.dict_csv-{random_string}.tmp"
         )  # csv形式辞書データの一時保存ファイル
-        tmp_compiled_path = compiled_dict_path.with_suffix(
-            f".dict_compiled-{random_string}.tmp"
+        tmp_compiled_path = user_dict_path.with_name(
+            f"user.dict_compiled-{random_string}.tmp"
         )  # コンパイル済み辞書データの一時保存ファイル
 
         try:
@@ -157,11 +208,10 @@ def update_dict(self) -> None:
             if not tmp_compiled_path.is_file():
                 raise RuntimeError("辞書のコンパイル時にエラーが発生しました。")
 
-            # コンパイル済み辞書の置き換え・読み込み
-            pyopenjtalk.unset_user_dict()
-            tmp_compiled_path.replace(compiled_dict_path)
-            if compiled_dict_path.is_file():
-                pyopenjtalk.set_user_dict(str(compiled_dict_path.resolve(strict=True)))
+            # コンパイル済み辞書の読み込み
+            pyopenjtalk.set_user_dict(
+                str(tmp_compiled_path.resolve(strict=True))
+            )  # NOTE: resolveによりコンパイル実行時でも相対パスを正しく認識できる
 
         except Exception as e:
             print("Error: Failed to update dictionary.", file=sys.stderr)
@@ -172,7 +222,7 @@ def update_dict(self) -> None:
             if tmp_csv_path.exists():
                 tmp_csv_path.unlink()
             if tmp_compiled_path.exists():
-                tmp_compiled_path.unlink()
+                _delete_file_on_close(tmp_compiled_path)
 
     @mutex_wrapper(mutex_user_dict)
     def read_dict(self) -> dict[str, UserDictWord]: