feat: support fish audio TTS (#7982)

api/tests/integration_tests/model_runtime/__mock/fishaudio.py

@@ -0,0 +1,82 @@
+import os
+from collections.abc import Callable
+from typing import Literal
+
+import httpx
+import pytest
+from _pytest.monkeypatch import MonkeyPatch
+
+
+def mock_get(*args, **kwargs):
+    if kwargs.get("headers", {}).get("Authorization") != "Bearer test":
+        raise httpx.HTTPStatusError(
+            "Invalid API key",
+            request=httpx.Request("GET", ""),
+            response=httpx.Response(401),
+        )
+
+    return httpx.Response(
+        200,
+        json={
+            "items": [
+                {"title": "Model 1", "_id": "model1"},
+                {"title": "Model 2", "_id": "model2"},
+            ]
+        },
+        request=httpx.Request("GET", ""),
+    )
+
+
+def mock_stream(*args, **kwargs):
+    class MockStreamResponse:
+        def __init__(self):
+            self.status_code = 200
+
+        def __enter__(self):
+            return self
+
+        def __exit__(self, exc_type, exc_val, exc_tb):
+            pass
+
+        def iter_bytes(self):
+            yield b"Mocked audio data"
+
+    return MockStreamResponse()
+
+
+def mock_fishaudio(
+    monkeypatch: MonkeyPatch,
+    methods: list[Literal["list-models", "tts"]],
+) -> Callable[[], None]:
+    """
+    mock fishaudio module
+
+    :param monkeypatch: pytest monkeypatch fixture
+    :return: unpatch function
+    """
+
+    def unpatch() -> None:
+        monkeypatch.undo()
+
+    if "list-models" in methods:
+        monkeypatch.setattr(httpx, "get", mock_get)
+
+    if "tts" in methods:
+        monkeypatch.setattr(httpx, "stream", mock_stream)
+
+    return unpatch
+
+
+MOCK = os.getenv("MOCK_SWITCH", "false").lower() == "true"
+
+
+@pytest.fixture
+def setup_fishaudio_mock(request, monkeypatch):
+    methods = request.param if hasattr(request, "param") else []
+    if MOCK:
+        unpatch = mock_fishaudio(monkeypatch, methods=methods)
+
+    yield
+
+    if MOCK:
+        unpatch()

api/tests/integration_tests/model_runtime/fishaudio/test_provider.py

@@ -0,0 +1,33 @@
+import os
+
+import httpx
+import pytest
+
+from core.model_runtime.errors.validate import CredentialsValidateFailedError
+from core.model_runtime.model_providers.fishaudio.fishaudio import FishAudioProvider
+from tests.integration_tests.model_runtime.__mock.fishaudio import setup_fishaudio_mock
+
+
+@pytest.mark.parametrize("setup_fishaudio_mock", [["list-models"]], indirect=True)
+def test_validate_provider_credentials(setup_fishaudio_mock):
+    print("-----", httpx.get)
+    provider = FishAudioProvider()
+
+    with pytest.raises(CredentialsValidateFailedError):
+        provider.validate_provider_credentials(
+            credentials={
+                "api_key": "bad_api_key",
+                "api_base": os.environ.get("FISH_AUDIO_API_BASE", "https://api.fish.audio"),
+                "use_public_models": "false",
+                "latency": "normal",
+            }
+        )
+
+    provider.validate_provider_credentials(
+        credentials={
+            "api_key": os.environ.get("FISH_AUDIO_API_KEY", "test"),
+            "api_base": os.environ.get("FISH_AUDIO_API_BASE", "https://api.fish.audio"),
+            "use_public_models": "false",
+            "latency": "normal",
+        }
+    )

api/tests/integration_tests/model_runtime/fishaudio/test_tts.py

@@ -0,0 +1,32 @@
+import os
+
+import pytest
+
+from core.model_runtime.model_providers.fishaudio.tts.tts import (
+    FishAudioText2SpeechModel,
+)
+from tests.integration_tests.model_runtime.__mock.fishaudio import setup_fishaudio_mock
+
+
+@pytest.mark.parametrize("setup_fishaudio_mock", [["tts"]], indirect=True)
+def test_invoke_model(setup_fishaudio_mock):
+    model = FishAudioText2SpeechModel()
+
+    result = model.invoke(
+        model="tts-default",
+        tenant_id="test",
+        credentials={
+            "api_key": os.environ.get("FISH_AUDIO_API_KEY", "test"),
+            "api_base": os.environ.get("FISH_AUDIO_API_BASE", "https://api.fish.audio"),
+            "use_public_models": "false",
+            "latency": "normal",
+        },
+        content_text="Hello, world!",
+        voice="03397b4c4be74759b72533b663fbd001",
+    )
+
+    content = b""
+    for chunk in result:
+        content += chunk
+
+    assert content != b""