Support audio model in Azure provider

g4f/Provider/needs_auth/Azure.py

@@ -4,7 +4,7 @@ import os
 import json
 
 from ...typing import Messages, AsyncResult
-from ...errors import MissingAuthError
+from ...errors import MissingAuthError, ModelNotFoundError
 from ..template import OpenaiTemplate
 
 class Azure(OpenaiTemplate):
@@ -15,9 +15,20 @@ class Azure(OpenaiTemplate):
     active_by_default = True
     login_url = "https://discord.gg/qXA4Wf4Fsm"
     routes: dict[str, str] = {}
+    audio_models = ["gpt-4o-mini-audio-preview"]
+    model_extra_body = {
+        "gpt-4o-mini-audio-preview": {
+            "audio": {
+                "voice": "alloy",
+                "format": "mp3"
+            },
+            "modalities": ["text", "audio"],
+            "stream": False
+        }
+    }
 
     @classmethod
-    def get_models(cls, **kwargs) -> list[str]:
+    def get_models(cls, api_key: str = None, **kwargs) -> list[str]:
         routes = os.environ.get("AZURE_ROUTES")
         if routes:
             try:
@@ -27,7 +38,7 @@ class Azure(OpenaiTemplate):
             cls.routes = routes
         if cls.routes:
             return list(cls.routes.keys())
-        return super().get_models(**kwargs)
+        return super().get_models(api_key=api_key, **kwargs)
 
     @classmethod
     async def create_async_generator(
@@ -40,6 +51,9 @@ class Azure(OpenaiTemplate):
     ) -> AsyncResult:
         if not model:
             model = os.environ.get("AZURE_DEFAULT_MODEL", cls.default_model)
+        if model in cls.model_extra_body:
+            for key, value in cls.model_extra_body[model].items():
+                kwargs.setdefault(key, value)
         if not api_key:
             raise ValueError(f"API key is required for Azure provider. Ask for API key in the {cls.login_url} Discord server.")
         if not api_endpoint:
@@ -47,7 +61,7 @@ class Azure(OpenaiTemplate):
                 cls.get_models()
             api_endpoint = cls.routes.get(model)
             if cls.routes and not api_endpoint:
-                raise ValueError(f"No API endpoint found for model: {model}")
+                raise ModelNotFoundError(f"No API endpoint found for model: {model}")
         if not api_endpoint:
             api_endpoint = os.environ.get("AZURE_API_ENDPOINT")
         try:

g4f/Provider/template/OpenaiTemplate.py

@@ -7,6 +7,7 @@ from ..base_provider import AsyncGeneratorProvider, ProviderModelMixin, RaiseErr
 from ...typing import Union, AsyncResult, Messages, MediaListType
 from ...requests import StreamSession, raise_for_status
 from ...image import use_aspect_ratio
+from ...image.copy_images import save_response_media
 from ...providers.response import FinishReason, ToolCalls, Usage, ImageResponse, ProviderInfo
 from ...tools.media import render_messages
 from ...errors import MissingAuthError, ResponseError
@@ -62,7 +63,7 @@ class OpenaiTemplate(AsyncGeneratorProvider, ProviderModelMixin, RaiseErrorMixin
         max_tokens: int = None,
         top_p: float = None,
         stop: Union[str, list[str]] = None,
-        stream: bool = False,
+        stream: bool = None,
         prompt: str = None,
         headers: dict = None,
         impersonate: str = None,
@@ -115,7 +116,7 @@ class OpenaiTemplate(AsyncGeneratorProvider, ProviderModelMixin, RaiseErrorMixin
                 max_tokens=max_tokens,
                 top_p=top_p,
                 stop=stop,
-                stream=stream,
+                stream="audio" not in extra_parameters if stream is None else stream,
                 **extra_parameters,
                 **extra_body
             )
@@ -136,10 +137,18 @@ class OpenaiTemplate(AsyncGeneratorProvider, ProviderModelMixin, RaiseErrorMixin
                         yield Usage(**data["usage"])
                     if "choices" in data:
                         choice = next(iter(data["choices"]), None)
-                        if choice and "content" in choice["message"] and choice["message"]["content"]:
+                        message = choice.get("message", {})
-                            yield choice["message"]["content"].strip()
+                        if choice and "content" in message and message["content"]:
-                        if "tool_calls" in choice["message"]:
+                            yield message["content"].strip()
-                            yield ToolCalls(choice["message"]["tool_calls"])
+                        if "tool_calls" in message:
+                            yield ToolCalls(message["tool_calls"])
+                        audio = message.get("audio", {})
+                        if "data" in audio:
+                            async for chunk in save_response_media(audio["data"], prompt, [model, extra_body.get("audio", {}).get("voice")]):
+                                yield chunk
+                        if "transcript" in audio:
+                            yield "\n\n"
+                            yield audio["transcript"]
                         if choice and "finish_reason" in choice and choice["finish_reason"] is not None:
                             yield FinishReason(choice["finish_reason"])
                             return

g4f/client/service.py

@@ -23,7 +23,7 @@ def convert_to_provider(provider: str) -> ProviderType:
 
 def get_model_and_provider(model    : Union[Model, str], 
                            provider : Union[ProviderType, str, None], 
-                           stream   : bool,
+                           stream   : bool = False,
                            ignore_working: bool = False,
                            ignore_stream: bool = False,
                            logging: bool = True,

g4f/gui/server/api.py

@@ -149,7 +149,6 @@ class Api:
             "model": model,
             "provider": provider,
             "messages": messages,
-            "stream": True,
             "ignore_stream": True,
             **kwargs
         }
@@ -166,8 +165,6 @@ class Api:
         try:
             model, provider_handler = get_model_and_provider(
                 kwargs.get("model"), provider,
-                stream=True,
-                ignore_stream=True,
                 has_images="media" in kwargs,
             )
             if "user" in kwargs:

g4f/image/copy_images.py

@@ -112,7 +112,7 @@ def get_filename(tags: list[str], alt: str, extension: str, image: str) -> str:
     return "".join((
         f"{int(time.time())}_",
         f"{secure_filename(tags + alt)}_" if alt else secure_filename(tags),
-        hashlib.sha256(image.encode()).hexdigest()[:16],
+        hashlib.sha256(str(time.time()).encode() if image is None else image.encode()).hexdigest()[:16],
         extension
     ))
 

g4f/providers/base_provider.py

@@ -292,7 +292,7 @@ class AsyncGeneratorProvider(AbstractProvider):
         cls,
         model: str,
         messages: Messages,
-        stream: bool = True,
+        stream: bool = None,
         timeout: int = None,
         **kwargs
     ) -> CreateResult:
@@ -312,7 +312,7 @@ class AsyncGeneratorProvider(AbstractProvider):
         """
         return to_sync_generator(
             cls.create_async_generator(model, messages, stream=stream, **kwargs),
-            stream=stream,
+            stream=stream is not False,
             timeout=timeout
         )
 
@@ -321,7 +321,6 @@ class AsyncGeneratorProvider(AbstractProvider):
     async def create_async_generator(
         model: str,
         messages: Messages,
-        stream: bool = True,
         **kwargs
     ) -> AsyncResult:
         """

projects/windows/docker-compose.yml

@@ -17,3 +17,4 @@ services:
     restart: on-failure
     volumes:
       - /var/win:/storage
+      - ./:/data