ENH: pydanticをv2へ (VOICEVOX#1184)

* ENH: pydanticをv2へ

* FIX: json indent=4

* FIX: 変数名

* FIX: コメント追加

* FIX: スキーマー修正

* Update voicevox_engine/app/application.py

---------

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

pyproject.toml

@@ -56,7 +56,7 @@ pyopenjtalk = { git = "https://github.com/VOICEVOX/pyopenjtalk", rev = "b35fc89f
 semver = "^3.0.0"
 platformdirs = "^4.2.0"
 soxr = "^0.3.6"
-pydantic = "^1.10.15"
+pydantic = "^2.7.3"
 starlette = "^0.37.0"
 
 [tool.poetry.group.dev.dependencies]

requirements-build.txt

@@ -1,4 +1,5 @@
 altgraph==0.17.4 ; python_version >= "3.11" and python_version < "3.12"
+annotated-types==0.6.0 ; python_version >= "3.11" and python_version < "3.12"
 anyio==4.3.0 ; python_version >= "3.11" and python_version < "3.12"
 cffi==1.16.0 ; python_version >= "3.11" and python_version < "3.12"
 click==8.1.7 ; python_version >= "3.11" and python_version < "3.12"
@@ -15,7 +16,8 @@ packaging==24.0 ; python_version >= "3.11" and python_version < "3.12"
 pefile==2023.2.7 ; python_version >= "3.11" and python_version < "3.12" and sys_platform == "win32"
 platformdirs==4.2.2 ; python_version >= "3.11" and python_version < "3.12"
 pycparser==2.22 ; python_version >= "3.11" and python_version < "3.12"
-pydantic==1.10.15 ; python_version >= "3.11" and python_version < "3.12"
+pydantic-core==2.18.4 ; python_version >= "3.11" and python_version < "3.12"
+pydantic==2.7.3 ; python_version >= "3.11" and python_version < "3.12"
 pyinstaller-hooks-contrib==2024.6 ; 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"

requirements-dev.txt

@@ -1,3 +1,4 @@
+annotated-types==0.6.0 ; python_version >= "3.11" and python_version < "3.12"
 anyio==4.3.0 ; python_version >= "3.11" and python_version < "3.12"
 attrs==23.2.0 ; python_version >= "3.11" and python_version < "3.12"
 authlib==1.3.0 ; python_version >= "3.11" and python_version < "3.12"
@@ -68,7 +69,8 @@ prettytable==3.10.0 ; python_version >= "3.11" and python_version < "3.12"
 ptyprocess==0.7.0 ; python_version >= "3.11" and python_version < "3.12"
 pycodestyle==2.11.1 ; python_version >= "3.11" and python_version < "3.12"
 pycparser==2.22 ; python_version >= "3.11" and python_version < "3.12"
-pydantic==1.10.15 ; python_version >= "3.11" and python_version < "3.12"
+pydantic-core==2.18.4 ; python_version >= "3.11" and python_version < "3.12"
+pydantic==2.7.3 ; 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"

requirements.txt

@@ -1,3 +1,4 @@
+annotated-types==0.6.0 ; python_version >= "3.11" and python_version < "3.12"
 anyio==4.3.0 ; python_version >= "3.11" and python_version < "3.12"
 cffi==1.16.0 ; python_version >= "3.11" and python_version < "3.12"
 click==8.1.7 ; python_version >= "3.11" and python_version < "3.12"
@@ -11,7 +12,8 @@ markupsafe==2.1.5 ; python_version >= "3.11" and python_version < "3.12"
 numpy==1.26.4 ; python_version >= "3.11" and python_version < "3.12"
 platformdirs==4.2.2 ; python_version >= "3.11" and python_version < "3.12"
 pycparser==2.22 ; python_version >= "3.11" and python_version < "3.12"
-pydantic==1.10.15 ; python_version >= "3.11" and python_version < "3.12"
+pydantic-core==2.18.4 ; python_version >= "3.11" and python_version < "3.12"
+pydantic==2.7.3 ; 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"
 python-multipart==0.0.9 ; python_version >= "3.11" and python_version < "3.12"
 pyworld==0.3.4 ; python_version >= "3.11" and python_version < "3.12"

test/e2e/test_speakers.py

@@ -6,11 +6,13 @@
 from test.utility import hash_long_string
 
 from fastapi.testclient import TestClient
-from pydantic import parse_obj_as
+from pydantic import TypeAdapter
 from syrupy.assertion import SnapshotAssertion
 
 from voicevox_engine.metas.Metas import Speaker
 
+_speaker_list_adapter = TypeAdapter(list[Speaker])
+
 
 def test_話者一覧が取得できる(
     client: TestClient, snapshot_json: SnapshotAssertion
@@ -23,7 +25,7 @@ def test_話者一覧が取得できる(
 def test_話者の情報を取得できる(
     client: TestClient, snapshot_json: SnapshotAssertion
 ) -> None:
-    speakers = parse_obj_as(list[Speaker], client.get("/speakers").json())
+    speakers = _speaker_list_adapter.validate_python(client.get("/speakers").json())
     for speaker in speakers:
         response = client.get(
             "/speaker_info", params={"speaker_uuid": speaker.speaker_uuid}
@@ -44,7 +46,7 @@ def test_歌手一覧が取得できる(
 def test_歌手の情報を取得できる(
     client: TestClient, snapshot_json: SnapshotAssertion
 ) -> None:
-    singers = parse_obj_as(list[Speaker], client.get("/singers").json())
+    singers = _speaker_list_adapter.validate_python(client.get("/singers").json())
     for singer in singers:
         response = client.get(
             "/singer_info", params={"speaker_uuid": singer.speaker_uuid}

test/unit/setting/test_setting.py

@@ -15,7 +15,7 @@ def test_setting_handler_load_not_exist_file() -> None:
     # Expects
     true_setting = {"allow_origin": None, "cors_policy_mode": CorsPolicyMode.localapps}
     # Outputs
-    setting = settings.dict()
+    setting = settings.model_dump()
     # Test
     assert true_setting == setting
 
@@ -29,7 +29,7 @@ def test_setting_handler_load_exist_file_1() -> None:
     # Expects
     true_setting = {"allow_origin": None, "cors_policy_mode": CorsPolicyMode.localapps}
     # Outputs
-    setting = settings.dict()
+    setting = settings.model_dump()
     # Test
     assert true_setting == setting
 
@@ -43,7 +43,7 @@ def test_setting_handler_load_exist_file_2() -> None:
     # Expects
     true_setting = {"allow_origin": None, "cors_policy_mode": "all"}
     # Outputs
-    setting = settings.dict()
+    setting = settings.model_dump()
     # Test
     assert true_setting == setting
 
@@ -60,7 +60,7 @@ def test_setting_handler_load_exist_file_3() -> None:
         "cors_policy_mode": CorsPolicyMode.localapps,
     }
     # Outputs
-    setting = settings.dict()
+    setting = settings.model_dump()
     # Test
     assert true_setting == setting
 
@@ -76,7 +76,7 @@ def test_setting_handler_save(tmp_path: Path) -> None:
     # Outputs
     setting_loader.save(new_setting)
     # NOTE: `.load()` の正常動作を前提とする
-    setting = setting_loader.load().dict()
+    setting = setting_loader.load().model_dump()
     # Test
     assert true_setting == setting
 

test/utility.py

@@ -1,11 +1,10 @@
 import hashlib
 import io
-import json
 from typing import Any
 
 import numpy as np
 import soundfile as sf
-from pydantic.json import pydantic_encoder
+from fastapi.encoders import jsonable_encoder
 
 
 def round_floats(value: Any, round_value: int) -> Any:
@@ -22,7 +21,7 @@ def round_floats(value: Any, round_value: int) -> Any:
 
 def pydantic_to_native_type(value: Any) -> Any:
     """pydanticの型をnativeな型に変換する"""
-    return json.loads(json.dumps(value, default=pydantic_encoder))
+    return jsonable_encoder(value)
 
 
 def hash_long_string(value: Any) -> Any:

voicevox_engine/app/application.py

@@ -53,6 +53,7 @@ def generate_app(
         title=engine_manifest.name,
         description=f"{engine_manifest.brand_name} の音声合成エンジンです。",
         version=__version__,
+        separate_input_output_schemas=False,  # Pydantic V1 のときのスキーマに合わせるため
     )
     app = configure_middlewares(app, cors_policy_mode, allow_origin)
     app = configure_global_exception_handlers(app)

voicevox_engine/app/openapi_schema.py

@@ -20,13 +20,17 @@ def custom_openapi() -> Any:
         openapi_schema = get_openapi(
             title=app.title,
             version=app.version,
+            openapi_version=app.openapi_version,
+            summary=app.summary,
             description=app.description,
-            routes=app.routes,
-            tags=app.openapi_tags,
-            servers=app.servers,
             terms_of_service=app.terms_of_service,
             contact=app.contact,
             license_info=app.license_info,
+            routes=app.routes,
+            webhooks=app.webhooks.routes,
+            tags=app.openapi_tags,
+            servers=app.servers,
+            separate_input_output_schemas=app.separate_input_output_schemas,
         )
         if manage_library:
             additional_models: list[type[BaseModel]] = [
@@ -35,10 +39,12 @@ def custom_openapi() -> Any:
             ]
             for model in additional_models:
                 # ref_templateを指定しない場合、definitionsを参照してしまうので、手動で指定する
-                schema = model.schema(ref_template="#/components/schemas/{model}")
+                schema = model.model_json_schema(
+                    ref_template="#/components/schemas/{model}"
+                )
                 # definitionsは既存のモデルを重複して定義するため、不要なので削除
-                if "definitions" in schema:
-                    del schema["definitions"]
+                if "$defs" in schema:
+                    del schema["$defs"]
                 openapi_schema["components"]["schemas"][schema["title"]] = schema
         app.openapi_schema = openapi_schema
         return openapi_schema

voicevox_engine/app/routers/engine_info.py

@@ -4,6 +4,7 @@
 
 from fastapi import APIRouter, HTTPException
 from pydantic import BaseModel, Field
+from pydantic.json_schema import SkipJsonSchema
 
 from voicevox_engine import __version__
 from voicevox_engine.core.core_adapter import DeviceSupport
@@ -47,7 +48,9 @@ async def core_versions() -> list[str]:
         return core_manager.versions()
 
     @router.get("/supported_devices")
-    def supported_devices(core_version: str | None = None) -> SupportedDevicesInfo:
+    def supported_devices(
+        core_version: str | SkipJsonSchema[None] = None,
+    ) -> SupportedDevicesInfo:
         """対応デバイスの一覧を取得します。"""
         supported_devices = core_manager.get_core(core_version).supported_devices
         if supported_devices is None:

voicevox_engine/app/routers/morphing.py

@@ -6,6 +6,7 @@
 
 import soundfile
 from fastapi import APIRouter, HTTPException, Query
+from pydantic.json_schema import SkipJsonSchema
 from starlette.background import BackgroundTask
 from starlette.responses import FileResponse
 
@@ -45,7 +46,7 @@ def generate_morphing_router(
         summary="指定したスタイル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_style_ids: list[StyleId], core_version: str | None = None
+        base_style_ids: list[StyleId], core_version: str | SkipJsonSchema[None] = None
     ) -> list[dict[str, MorphableTargetInfo]]:
         """
         指定されたベーススタイルに対してエンジン内の各話者がモーフィング機能を利用可能か返します。
@@ -83,7 +84,7 @@ def _synthesis_morphing(
         base_style_id: Annotated[StyleId, Query(alias="base_speaker")],
         target_style_id: Annotated[StyleId, Query(alias="target_speaker")],
         morph_rate: Annotated[float, Query(ge=0.0, le=1.0)],
-        core_version: str | None = None,
+        core_version: str | SkipJsonSchema[None] = None,
     ) -> FileResponse:
         """
         指定された2種類のスタイルで音声を合成、指定した割合でモーフィングした音声を得ます。

voicevox_engine/app/routers/setting.py

@@ -4,6 +4,7 @@
 
 from fastapi import APIRouter, Depends, Form, Request, Response
 from fastapi.templating import Jinja2Templates
+from pydantic.json_schema import SkipJsonSchema
 
 from voicevox_engine.engine_manifest import BrandName
 from voicevox_engine.setting.model import CorsPolicyMode
@@ -53,7 +54,7 @@ def setting_get(request: Request) -> Response:
     )
     def setting_post(
         cors_policy_mode: Annotated[CorsPolicyMode, Form()],
-        allow_origin: Annotated[str | None, Form()] = None,
+        allow_origin: Annotated[str | SkipJsonSchema[None], Form()] = None,
     ) -> None:
         """
         設定を更新します。

voicevox_engine/app/routers/speaker.py

@@ -5,7 +5,8 @@
 from typing import Literal
 
 from fastapi import APIRouter, HTTPException
-from pydantic import parse_obj_as
+from pydantic import TypeAdapter
+from pydantic.json_schema import SkipJsonSchema
 
 from voicevox_engine.core.core_initializer import CoreManager
 from voicevox_engine.metas.Metas import Speaker, SpeakerInfo
@@ -25,14 +26,16 @@ def generate_speaker_router(
     router = APIRouter(tags=["その他"])
 
     @router.get("/speakers")
-    def speakers(core_version: str | None = None) -> list[Speaker]:
+    def speakers(core_version: str | SkipJsonSchema[None] = None) -> list[Speaker]:
         """話者情報の一覧を取得します。"""
         core = core_manager.get_core(core_version)
         speakers = metas_store.load_combined_metas(core.speakers)
         return filter_speakers_and_styles(speakers, "speaker")
 
     @router.get("/speaker_info")
-    def speaker_info(speaker_uuid: str, core_version: str | None = None) -> SpeakerInfo:
+    def speaker_info(
+        speaker_uuid: str, core_version: str | SkipJsonSchema[None] = None
+    ) -> SpeakerInfo:
         """
         指定されたspeaker_uuidの話者に関する情報をjson形式で返します。
         画像や音声はbase64エンコードされたものが返されます。
@@ -43,6 +46,8 @@ def speaker_info(speaker_uuid: str, core_version: str | None = None) -> SpeakerI
             core_version=core_version,
         )
 
+    _speaker_list_adapter = TypeAdapter(list[Speaker])
+
     # FIXME: この関数をどこかに切り出す
     def _speaker_info(
         speaker_uuid: str,
@@ -73,8 +78,8 @@ def _speaker_info(
         #           ...
 
         # 該当話者を検索する
-        speakers = parse_obj_as(
-            list[Speaker], core_manager.get_core(core_version).speakers
+        speakers = _speaker_list_adapter.validate_python(
+            core_manager.get_core(core_version).speakers, from_attributes=True
         )
         speakers = filter_speakers_and_styles(speakers, speaker_or_singer)
         speaker = next(
@@ -135,14 +140,16 @@ def _speaker_info(
         return spk_info
 
     @router.get("/singers")
-    def singers(core_version: str | None = None) -> list[Speaker]:
+    def singers(core_version: str | SkipJsonSchema[None] = None) -> list[Speaker]:
         """歌手情報の一覧を取得します"""
         core = core_manager.get_core(core_version)
         singers = metas_store.load_combined_metas(core.speakers)
         return filter_speakers_and_styles(singers, "singer")
 
     @router.get("/singer_info")
-    def singer_info(speaker_uuid: str, core_version: str | None = None) -> SpeakerInfo:
+    def singer_info(
+        speaker_uuid: str, core_version: str | SkipJsonSchema[None] = None
+    ) -> SpeakerInfo:
         """
         指定されたspeaker_uuidの歌手に関する情報をjson形式で返します。
         画像や音声はbase64エンコードされたものが返されます。