feat: support speech2text and tts for gpustack

api/core/model_runtime/model_providers/gpustack/gpustack.yaml

@@ -9,6 +9,8 @@ supported_model_types:
   - llm
   - text-embedding
   - rerank
+  - speech2text
+  - tts
 configurate_methods:
   - customizable-model
 model_credential_schema:
@@ -118,3 +120,19 @@ model_credential_schema:
           label:
             en_US: Not Support
             zh_Hans: 不支持
+    - variable: voices
+      show_on:
+        - variable: __model_type
+          value: tts
+      label:
+        en_US: Available Voices (comma-separated)
+        zh_Hans: 可用声音（用英文逗号分隔）
+      type: text-input
+      required: false
+      default: "Chinese Female"
+      placeholder:
+        en_US: "Chinese Female, Chinese Male, Japanese Male, Cantonese Female, English Female, English Male, Korean Female"
+        zh_Hans: "Chinese Female, Chinese Male, Japanese Male, Cantonese Female, English Female, English Male, Korean Female"
+      help:
+        en_US: "List voice names separated by commas. First voice will be used as default."
+        zh_Hans: "用英文逗号分隔的声音列表。第一个声音将作为默认值。"

api/core/model_runtime/model_providers/gpustack/speech2text/speech2text.py

@@ -0,0 +1,43 @@
+from typing import IO, Optional
+
+from core.model_runtime.model_providers.openai_api_compatible.speech2text.speech2text import OAICompatSpeech2TextModel
+
+
+class GPUStackSpeech2TextModel(OAICompatSpeech2TextModel):
+    """
+    Model class for GPUStack Speech to text model.
+    """
+
+    def _invoke(self, model: str, credentials: dict, file: IO[bytes], user: Optional[str] = None) -> str:
+        """
+        Invoke speech2text model
+        :param model: model name
+        :param credentials: model credentials
+        :param file: audio file
+        :param user: unique user id
+        :return: text for given audio file
+        """
+        compatible_credentials = self._get_compatible_credentials(credentials)
+        return super()._invoke(model, compatible_credentials, file)
+
+    def validate_credentials(self, model: str, credentials: dict) -> None:
+        """
+        Validate model credentials
+
+        :param model: model name
+        :param credentials: model credentials
+        """
+        compatible_credentials = self._get_compatible_credentials(credentials)
+        super().validate_credentials(model, compatible_credentials)
+
+    def _get_compatible_credentials(self, credentials: dict) -> dict:
+        """
+        Get compatible credentials
+
+        :param credentials: model credentials
+        :return: compatible credentials
+        """
+        compatible_credentials = credentials.copy()
+        base_url = credentials["endpoint_url"].rstrip("/").removesuffix("/v1-openai")
+        compatible_credentials["endpoint_url"] = f"{base_url}/v1-openai"
+        return compatible_credentials

api/core/model_runtime/model_providers/gpustack/tts/tts.py

@@ -0,0 +1,57 @@
+from typing import Any, Optional
+
+from core.model_runtime.model_providers.openai_api_compatible.tts.tts import OAICompatText2SpeechModel
+
+
+class GPUStackText2SpeechModel(OAICompatText2SpeechModel):
+    """
+    Model class for GPUStack Text to Speech model.
+    """
+
+    def _invoke(
+        self, model: str, tenant_id: str, credentials: dict, content_text: str, voice: str, user: Optional[str] = None
+    ) -> Any:
+        """
+        Invoke text2speech model
+
+        :param model: model name
+        :param tenant_id: user tenant id
+        :param credentials: model credentials
+        :param content_text: text content to be translated
+        :param voice: model timbre
+        :param user: unique user id
+        :return: text translated to audio file
+        """
+        compatible_credentials = self._get_compatible_credentials(credentials)
+        return super()._invoke(
+            model=model,
+            tenant_id=tenant_id,
+            credentials=compatible_credentials,
+            content_text=content_text,
+            voice=voice,
+            user=user,
+        )
+
+    def validate_credentials(self, model: str, credentials: dict, user: Optional[str] = None) -> None:
+        """
+        Validate model credentials
+
+        :param model: model name
+        :param credentials: model credentials
+        :param user: unique user id
+        """
+        compatible_credentials = self._get_compatible_credentials(credentials)
+        super().validate_credentials(model, compatible_credentials)
+
+    def _get_compatible_credentials(self, credentials: dict) -> dict:
+        """
+        Get compatible credentials
+
+        :param credentials: model credentials
+        :return: compatible credentials
+        """
+        compatible_credentials = credentials.copy()
+        base_url = credentials["endpoint_url"].rstrip("/").removesuffix("/v1-openai")
+        compatible_credentials["endpoint_url"] = f"{base_url}/v1-openai"
+
+        return compatible_credentials