test: cancellable_synthesis, connect_waves API のスナップショットテスト (#1538)

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

Author: nanae772

test/e2e/single_api/tts_pipeline/__snapshots__/test_cancellable_synthesis.ambr

@@ -0,0 +1,4 @@
+# serializer version: 1
+# name: test_post_cancellable_synthesis_200
+  'MD5:f7d42ce5787856549abc3d2d7561c06f'
+# ---

test/e2e/single_api/tts_pipeline/__snapshots__/test_connect_waves.ambr

@@ -0,0 +1,4 @@
+# serializer version: 1
+# name: test_post_connect_waves_200
+  'MD5:09eeac5fac75ce47f5002c48fef3044d'
+# ---

test/e2e/single_api/tts_pipeline/test_audio/sample1.wav

@@ -0,0 +1,65 @@
+"""
+/cancellable_synthesis API のテスト
+"""
+
+from test.e2e.single_api.utils import gen_mora
+from test.utility import hash_wave_floats_from_wav_bytes
+
+import pytest
+from fastapi.testclient import TestClient
+from syrupy.assertion import SnapshotAssertion
+
+from voicevox_engine.app.application import generate_app
+from voicevox_engine.cancellable_engine import CancellableEngine
+
+
+@pytest.fixture()
+def cancellable_client(app_params: dict) -> TestClient:
+    app_params["cancellable_engine"] = CancellableEngine(
+        init_processes=1,
+        use_gpu=False,
+        enable_mock=True,
+    )
+    cancellable_app = generate_app(**app_params)
+    return TestClient(cancellable_app)
+
+
+def test_post_cancellable_synthesis_200(
+    cancellable_client: TestClient, snapshot: SnapshotAssertion
+) -> None:
+    query = {
+        "accent_phrases": [
+            {
+                "moras": [
+                    gen_mora("テ", "t", 2.3, "e", 0.8, 3.3),
+                    gen_mora("ス", "s", 2.1, "U", 0.3, 0.0),
+                    gen_mora("ト", "t", 2.3, "o", 1.8, 4.1),
+                ],
+                "accent": 1,
+                "pause_mora": None,
+                "is_interrogative": False,
+            }
+        ],
+        "speedScale": 1.0,
+        "pitchScale": 1.0,
+        "intonationScale": 1.0,
+        "volumeScale": 1.0,
+        "prePhonemeLength": 0.1,
+        "postPhonemeLength": 0.1,
+        "pauseLength": None,
+        "pauseLengthScale": 1.0,
+        "outputSamplingRate": 24000,
+        "outputStereo": False,
+        "kana": "テ'_スト",
+    }
+    response = cancellable_client.post(
+        "/cancellable_synthesis", params={"speaker": 0}, json=query
+    )
+    assert response.status_code == 200
+
+    # 音声波形が一致する
+    assert response.headers["content-type"] == "audio/wav"
+    assert snapshot == hash_wave_floats_from_wav_bytes(response.read())
+
+
+# TODO: キャンセルするテストを追加する

test/e2e/single_api/tts_pipeline/test_audio/sample2.wav

@@ -0,0 +1,27 @@
+"""
+/connect_waves API のテスト
+"""
+
+import base64
+from pathlib import Path
+from test.utility import hash_wave_floats_from_wav_bytes
+
+from fastapi.testclient import TestClient
+from syrupy.assertion import SnapshotAssertion
+
+
+def test_post_connect_waves_200(
+    client: TestClient, snapshot: SnapshotAssertion
+) -> None:
+    test_audio_dir = Path(__file__).parent / "test_audio"
+    wavs = [
+        base64.b64encode((test_audio_dir / "sample1.wav").read_bytes()).decode(),
+        base64.b64encode((test_audio_dir / "sample2.wav").read_bytes()).decode(),
+    ]
+
+    response = client.post("/connect_waves", json=wavs)
+    assert response.status_code == 200
+
+    # 音声波形が一致する
+    assert response.headers["content-type"] == "audio/wav"
+    assert snapshot == hash_wave_floats_from_wav_bytes(response.read())

test/e2e/single_api/tts_pipeline/test_cancellable_synthesis.py

@@ -52,11 +52,11 @@ def __init__(
         self,
         init_processes: int,
         use_gpu: bool,
-        voicelib_dirs: list[Path] | None,
-        voicevox_dir: Path | None,
-        runtime_dirs: list[Path] | None,
-        cpu_num_threads: int | None,
-        enable_mock: bool,
+        voicelib_dirs: list[Path] | None = None,
+        voicevox_dir: Path | None = None,
+        runtime_dirs: list[Path] | None = None,
+        cpu_num_threads: int | None = None,
+        enable_mock: bool = True,
     ) -> None:
         """
         変数の初期化を行う