APIのstyle_idの型をStyleId型に (#966)

* run.pyのとこStyleIdに

* API引数をStyleId化

* 間違えて追加してしまっていた

* 漏れ

* StyleIdの場所変更

* pysen

* 自動import箇所が意図とあってなさそうだった

* ワークアラウンドなことをコメント、FIXMEをコメント

run.py

@@ -35,6 +35,7 @@
 from voicevox_engine.engine_manifest import EngineManifestLoader
 from voicevox_engine.engine_manifest.EngineManifest import EngineManifest
 from voicevox_engine.library_manager import LibraryManager
+from voicevox_engine.metas.Metas import StyleId
 from voicevox_engine.metas.MetasStore import MetasStore, construct_lookup
 from voicevox_engine.model import (
     AccentPhrase,
@@ -47,7 +48,6 @@
     ParseKanaError,
     Speaker,
     SpeakerInfo,
-    StyleId,
     StyleIdNotFoundError,
     SupportedDevicesInfo,
     UserDictWord,
@@ -93,17 +93,17 @@
 
 
 def get_style_id_from_deprecated(
-    style_id: int | None, speaker_id: int | None
+    style_id: StyleId | None, speaker_id: StyleId | None
 ) -> StyleId:
     """
     style_idとspeaker_id両方ともNoneかNoneでないかをチェックし、
     どちらか片方しかNoneが存在しなければstyle_idを返す
     """
     if speaker_id is not None and style_id is None:
         warnings.warn("speakerは非推奨です。style_idを利用してください。", stacklevel=1)
-        return StyleId(speaker_id)
+        return speaker_id
     elif style_id is not None and speaker_id is None:
-        return StyleId(style_id)
+        return style_id
     raise HTTPException(
         status_code=400, detail="speakerとstyle_idが両方とも存在しないか、両方とも存在しています。"
     )
@@ -282,8 +282,8 @@ def get_core(core_version: Optional[str]) -> CoreAdapter:
     )
     def audio_query(
         text: str,
-        style_id: int | None = Query(default=None),  # noqa: B008
-        speaker: int | None = Query(default=None, deprecated=True),  # noqa: B008
+        style_id: StyleId | None = Query(default=None),  # noqa: B008
+        speaker: StyleId | None = Query(default=None, deprecated=True),  # noqa: B008
         core_version: str | None = None,
     ) -> AudioQuery:
         """
@@ -333,9 +333,7 @@ def audio_query_from_preset(
         else:
             raise HTTPException(status_code=422, detail="該当するプリセットIDが見つかりません")
 
-        accent_phrases = engine.create_accent_phrases(
-            text, StyleId(selected_preset.style_id)
-        )
+        accent_phrases = engine.create_accent_phrases(text, selected_preset.style_id)
         return AudioQuery(
             accent_phrases=accent_phrases,
             speedScale=selected_preset.speedScale,
@@ -363,8 +361,8 @@ def audio_query_from_preset(
     )
     def accent_phrases(
         text: str,
-        style_id: int | None = Query(default=None),  # noqa: B008
-        speaker: int | None = Query(default=None, deprecated=True),  # noqa: B008
+        style_id: StyleId | None = Query(default=None),  # noqa: B008
+        speaker: StyleId | None = Query(default=None, deprecated=True),  # noqa: B008
         is_kana: bool = False,
         core_version: str | None = None,
     ) -> list[AccentPhrase]:
@@ -401,8 +399,8 @@ def accent_phrases(
     )
     def mora_data(
         accent_phrases: list[AccentPhrase],
-        style_id: int | None = Query(default=None),  # noqa: B008
-        speaker: int | None = Query(default=None, deprecated=True),  # noqa: B008
+        style_id: StyleId | None = Query(default=None),  # noqa: B008
+        speaker: StyleId | None = Query(default=None, deprecated=True),  # noqa: B008
         core_version: str | None = None,
     ) -> list[AccentPhrase]:
         style_id = get_style_id_from_deprecated(style_id=style_id, speaker_id=speaker)
@@ -417,8 +415,8 @@ def mora_data(
     )
     def mora_length(
         accent_phrases: list[AccentPhrase],
-        style_id: int | None = Query(default=None),  # noqa: B008
-        speaker: int | None = Query(default=None, deprecated=True),  # noqa: B008
+        style_id: StyleId | None = Query(default=None),  # noqa: B008
+        speaker: StyleId | None = Query(default=None, deprecated=True),  # noqa: B008
         core_version: str | None = None,
     ) -> list[AccentPhrase]:
         style_id = get_style_id_from_deprecated(style_id=style_id, speaker_id=speaker)
@@ -433,8 +431,8 @@ def mora_length(
     )
     def mora_pitch(
         accent_phrases: list[AccentPhrase],
-        style_id: int | None = Query(default=None),  # noqa: B008
-        speaker: int | None = Query(default=None, deprecated=True),  # noqa: B008
+        style_id: StyleId | None = Query(default=None),  # noqa: B008
+        speaker: StyleId | None = Query(default=None, deprecated=True),  # noqa: B008
         core_version: str | None = None,
     ) -> list[AccentPhrase]:
         style_id = get_style_id_from_deprecated(style_id=style_id, speaker_id=speaker)
@@ -456,8 +454,8 @@ def mora_pitch(
     )
     def synthesis(
         query: AudioQuery,
-        style_id: int | None = Query(default=None),  # noqa: B008
-        speaker: int | None = Query(default=None, deprecated=True),  # noqa: B008
+        style_id: StyleId | None = Query(default=None),  # noqa: B008
+        speaker: StyleId | None = Query(default=None, deprecated=True),  # noqa: B008
         enable_interrogative_upspeak: bool = Query(  # noqa: B008
             default=True,
             description="疑問系のテキストが与えられたら語尾を自動調整する",
@@ -497,8 +495,8 @@ def synthesis(
     def cancellable_synthesis(
         query: AudioQuery,
         request: Request,
-        style_id: int | None = Query(default=None),  # noqa: B008
-        speaker: int | None = Query(default=None, deprecated=True),  # noqa: B008
+        style_id: StyleId | None = Query(default=None),  # noqa: B008
+        speaker: StyleId | None = Query(default=None, deprecated=True),  # noqa: B008
         core_version: str | None = None,
     ) -> FileResponse:
         style_id = get_style_id_from_deprecated(style_id=style_id, speaker_id=speaker)
@@ -536,8 +534,8 @@ def cancellable_synthesis(
     )
     def multi_synthesis(
         queries: list[AudioQuery],
-        style_id: int | None = Query(default=None),  # noqa: B008
-        speaker: int | None = Query(default=None, deprecated=True),  # noqa: B008
+        style_id: StyleId | None = Query(default=None),  # noqa: B008
+        speaker: StyleId | None = Query(default=None, deprecated=True),  # noqa: B008
         core_version: str | None = None,
     ) -> FileResponse:
         style_id = get_style_id_from_deprecated(style_id=style_id, speaker_id=speaker)
@@ -576,7 +574,7 @@ def multi_synthesis(
         summary="指定した話者に対してエンジン内の話者がモーフィングが可能か判定する",
     )
     def morphable_targets(
-        base_speakers: list[int],
+        base_speakers: list[int],  # FIXME: StyleId型にする
         core_version: str | None = None,
     ) -> list[dict[str, MorphableTargetInfo]]:
         """
@@ -617,7 +615,7 @@ def morphable_targets(
     )
     def _synthesis_morphing(
         query: AudioQuery,
-        base_speaker: int,
+        base_speaker: int,  # FIXME: StyleId型にする
         target_speaker: int,
         morph_rate: float = Query(..., ge=0.0, le=1.0),  # noqa: B008
         core_version: str | None = None,
@@ -1001,7 +999,7 @@ def uninstall_library(library_uuid: str) -> Response:
 
     @app.post("/initialize_style_id", status_code=204, tags=["その他"])
     def initialize_style_id(
-        style_id: int,
+        style_id: StyleId,
         skip_reinit: bool = Query(  # noqa: B008
             False, description="既に初期化済みのスタイルの再初期化をスキップするかどうか"
         ),
@@ -1012,23 +1010,23 @@ def initialize_style_id(
         実行しなくても他のAPIは使用できますが、初回実行時に時間がかかることがあります。
         """
         core = get_core(core_version)
-        core.initialize_style_id_synthesis(StyleId(style_id), skip_reinit=skip_reinit)
+        core.initialize_style_id_synthesis(style_id, skip_reinit=skip_reinit)
         return Response(status_code=204)
 
     @app.get("/is_initialized_style_id", response_model=bool, tags=["その他"])
     def is_initialized_style_id(
-        style_id: int,
+        style_id: StyleId,
         core_version: str | None = None,
     ) -> bool:
         """
         指定されたstyle_idのスタイルが初期化されているかどうかを返します。
         """
         core = get_core(core_version)
-        return core.is_initialized_style_id_synthesis(StyleId(style_id))
+        return core.is_initialized_style_id_synthesis(style_id)
 
     @app.post("/initialize_speaker", status_code=204, tags=["その他"], deprecated=True)
     def initialize_speaker(
-        speaker: int,
+        speaker: StyleId,
         skip_reinit: bool = Query(  # noqa: B008
             False, description="既に初期化済みの話者の再初期化をスキップするかどうか"
         ),
@@ -1044,14 +1042,14 @@ def initialize_speaker(
             stacklevel=1,
         )
         return initialize_style_id(
-            StyleId(speaker), skip_reinit=skip_reinit, core_version=core_version
+            speaker, skip_reinit=skip_reinit, core_version=core_version
         )
 
     @app.get(
         "/is_initialized_speaker", response_model=bool, tags=["その他"], deprecated=True
     )
     def is_initialized_speaker(
-        speaker: int,
+        speaker: StyleId,
         core_version: str | None = None,
     ) -> bool:
         """
@@ -1062,7 +1060,7 @@ def is_initialized_speaker(
             "使用しているAPI(/is_initialize_speaker)は非推奨です。/is_initialized_style_idを利用してください。",
             stacklevel=1,
         )
-        return is_initialized_style_id(StyleId(speaker), core_version=core_version)
+        return is_initialized_style_id(speaker, core_version=core_version)
 
     @app.get("/user_dict", response_model=dict[str, UserDictWord], tags=["ユーザー辞書"])
     def get_user_dict_words() -> dict[str, UserDictWord]:

test/test_mock_tts_engine.py

@@ -1,7 +1,8 @@
 from unittest import TestCase
 
 from voicevox_engine.dev.tts_engine import MockTTSEngine
-from voicevox_engine.model import AccentPhrase, AudioQuery, Mora, StyleId
+from voicevox_engine.metas.Metas import StyleId
+from voicevox_engine.model import AccentPhrase, AudioQuery, Mora
 from voicevox_engine.tts_pipeline.kana_converter import create_kana
 
 

test/test_tts_engine.py

@@ -5,7 +5,8 @@
 
 import numpy
 
-from voicevox_engine.model import AccentPhrase, AudioQuery, Mora, StyleId
+from voicevox_engine.metas.Metas import StyleId
+from voicevox_engine.model import AccentPhrase, AudioQuery, Mora
 from voicevox_engine.tts_pipeline import TTSEngine
 from voicevox_engine.tts_pipeline.acoustic_feature_extractor import Phoneme
 from voicevox_engine.tts_pipeline.tts_engine import (

test/test_tts_engine_base.py

@@ -1,7 +1,8 @@
 from unittest import TestCase
 
 from voicevox_engine.dev.core.mock import MockCoreWrapper
-from voicevox_engine.model import AccentPhrase, Mora, StyleId
+from voicevox_engine.metas.Metas import StyleId
+from voicevox_engine.model import AccentPhrase, Mora
 from voicevox_engine.tts_pipeline import TTSEngine
 from voicevox_engine.tts_pipeline.tts_engine import (
     apply_interrogative_upspeak,  # FIXME: この関数を使うテストをTTSEngine用のテストに移動する

voicevox_engine/cancellable_engine.py

@@ -17,7 +17,8 @@
 from fastapi import HTTPException, Request
 
 from .core_initializer import initialize_cores
-from .model import AudioQuery, StyleId
+from .metas.Metas import StyleId
+from .model import AudioQuery
 from .tts_pipeline import make_tts_engines_from_cores
 from .utility import get_latest_core_version
 

voicevox_engine/core_adapter.py

@@ -4,7 +4,7 @@
 from numpy import ndarray
 
 from .core_wrapper import CoreWrapper, OldCoreError
-from .model import StyleId
+from .metas.Metas import StyleId
 
 
 class CoreAdapter:

voicevox_engine/dev/tts_engine/mock.py

@@ -6,7 +6,8 @@
 from pyopenjtalk import tts
 from soxr import resample
 
-from ...model import AudioQuery, StyleId
+from ...metas.Metas import StyleId
+from ...model import AudioQuery
 from ...tts_pipeline import TTSEngine
 from ...tts_pipeline.tts_engine import to_flatten_moras
 from ..core.mock import MockCoreWrapper

voicevox_engine/metas/Metas.py

@@ -1,16 +1,20 @@
 from enum import Enum
-from typing import List, Optional
+from typing import List, NewType, Optional
 
 from pydantic import BaseModel, Field
 
+# NOTE: 循環importを防ぐためにとりあえずここに書いている
+# FIXME: 他のmodelに依存せず、全modelから参照できる場所に配置する
+StyleId = NewType("StyleId", int)
+
 
 class SpeakerStyle(BaseModel):
     """
     話者のスタイル情報
     """
 
     name: str = Field(title="スタイル名")
-    id: int = Field(title="スタイルID")
+    id: StyleId = Field(title="スタイルID")
 
 
 class SpeakerSupportPermittedSynthesisMorphing(str, Enum):
@@ -67,7 +71,7 @@ class StyleInfo(BaseModel):
     スタイルの追加情報
     """
 
-    id: int = Field(title="スタイルID")
+    id: StyleId = Field(title="スタイルID")
     icon: str = Field(title="当該スタイルのアイコンをbase64エンコードしたもの")
     portrait: Optional[str] = Field(title="当該スタイルのportrait.pngをbase64エンコードしたもの")
     voice_samples: List[str] = Field(title="voice_sampleのwavファイルをbase64エンコードしたもの")

voicevox_engine/model.py

@@ -1,6 +1,6 @@
 from enum import Enum
 from re import findall, fullmatch
-from typing import Any, Dict, List, NewType, Optional
+from typing import Any, Dict, List, Optional
 
 from pydantic import BaseModel, Field, StrictStr, validator
 
@@ -45,9 +45,6 @@ def __hash__(self):
         return hash(tuple(sorted(items)))
 
 
-StyleId = NewType("StyleId", int)
-
-
 class AudioQuery(BaseModel):
     """
     音声合成用のクエリ

voicevox_engine/morphing.py

@@ -8,9 +8,14 @@
 from soxr import resample
 
 from .core_adapter import CoreAdapter
-from .metas.Metas import Speaker, SpeakerStyle, SpeakerSupportPermittedSynthesisMorphing
+from .metas.Metas import (
+    Speaker,
+    SpeakerStyle,
+    SpeakerSupportPermittedSynthesisMorphing,
+    StyleId,
+)
 from .metas.MetasStore import construct_lookup
-from .model import AudioQuery, MorphableTargetInfo, StyleId, StyleIdNotFoundError
+from .model import AudioQuery, MorphableTargetInfo, StyleIdNotFoundError
 from .tts_pipeline import TTSEngine
 
 
@@ -52,7 +57,7 @@ def create_morphing_parameter(
 def get_morphable_targets(
     speakers: List[Speaker],
     base_speakers: List[int],
-) -> List[Dict[int, MorphableTargetInfo]]:
+) -> List[Dict[StyleId, MorphableTargetInfo]]:
     """
     speakers: 全話者の情報
     base_speakers: モーフィング可能か判定したいベースの話者リスト（スタイルID）

voicevox_engine/preset/Preset.py

@@ -1,5 +1,7 @@
 from pydantic import BaseModel, Field
 
+from voicevox_engine.metas.Metas import StyleId
+
 
 class Preset(BaseModel):
     """
@@ -9,7 +11,7 @@ class Preset(BaseModel):
     id: int = Field(title="プリセットID")
     name: str = Field(title="プリセット名")
     speaker_uuid: str = Field(title="話者のUUID")
-    style_id: int = Field(title="スタイルID")
+    style_id: StyleId = Field(title="スタイルID")
     speedScale: float = Field(title="全体の話速")
     pitchScale: float = Field(title="全体の音高")
     intonationScale: float = Field(title="全体の抑揚")

voicevox_engine/tts_pipeline/tts_engine.py

@@ -7,7 +7,8 @@
 
 from ..core_adapter import CoreAdapter
 from ..core_wrapper import CoreWrapper
-from ..model import AccentPhrase, AudioQuery, Mora, StyleId
+from ..metas.Metas import StyleId
+from ..model import AccentPhrase, AudioQuery, Mora
 from .acoustic_feature_extractor import Phoneme
 from .mora_list import openjtalk_mora2text
 from .text_analyzer import text_to_accent_phrases