SD2-1327-ll-ms-make-it-possible-to-generate-audio-files

edenai_apis/apis/amazon/amazon_llm_api.py

@@ -1,4 +1,4 @@
-from typing import Dict, List, Type, Union, Optional
+from typing import Dict, List, Type, Union, Optional, Literal
 
 import httpx
 from openai import BaseModel
@@ -25,6 +25,8 @@ def llm__chat(
         presence_penalty: Optional[float] = None,
         frequency_penalty: Optional[float] = None,
         logit_bias: Optional[dict] = None,
+        modalities: Optional[List[Literal["text", "audio"]]] = None,
+        audio: Optional[Dict] = None,
         # openai v1.0+ new params
         response_format: Optional[
             Union[dict, Type[BaseModel]]
@@ -81,6 +83,8 @@ def llm__chat(
             model_list=model_list,
             drop_invalid_params=drop_invalid_params,
             user=user,
+            modalities=modalities,
+            audio=audio,
             **kwargs,
         )
         return response

edenai_apis/apis/anthropic/anthropic_api.py

@@ -147,6 +147,8 @@ def llm__chat(
         presence_penalty: Optional[float] = None,
         frequency_penalty: Optional[float] = None,
         logit_bias: Optional[dict] = None,
+        modalities: Optional[List[Literal["text", "audio"]]] = None,
+        audio: Optional[Dict] = None,
         # openai v1.0+ new params
         response_format: Optional[
             Union[dict, Type[BaseModel]]
@@ -203,6 +205,8 @@ def llm__chat(
             model_list=model_list,
             drop_invalid_params=drop_invalid_params,
             user=user,
+            modalities=modalities,
+            audio=audio,
             **kwargs,
         )
         return response

edenai_apis/apis/cohere/cohere_api.py

@@ -346,6 +346,8 @@ def llm__chat(
         presence_penalty: Optional[float] = None,
         frequency_penalty: Optional[float] = None,
         logit_bias: Optional[dict] = None,
+        modalities: Optional[List[Literal["text", "audio"]]] = None,
+        audio: Optional[Dict] = None,
         # openai v1.0+ new params
         response_format: Optional[
             Union[dict, Type[BaseModel]]
@@ -402,6 +404,8 @@ def llm__chat(
             model_list=model_list,
             drop_invalid_params=drop_invalid_params,
             user=user,
+            modalities=modalities,
+            audio=audio,
             **kwargs,
         )
         return response

edenai_apis/apis/deepseek/deepseek_api.py

@@ -73,6 +73,8 @@ def llm__chat(
         presence_penalty: Optional[float] = None,
         frequency_penalty: Optional[float] = None,
         logit_bias: Optional[dict] = None,
+        modalities: Optional[List[Literal["text", "audio"]]] = None,
+        audio: Optional[Dict] = None,
         # openai v1.0+ new params
         response_format: Optional[
             Union[dict, Type[BaseModel]]
@@ -129,6 +131,8 @@ def llm__chat(
             model_list=model_list,
             drop_invalid_params=drop_invalid_params,
             user=user,
+            modalities=modalities,
+            audio=audio,
             **kwargs,
         )
         return response

edenai_apis/apis/google/google_llm_api.py

@@ -1,4 +1,4 @@
-from typing import List, Type, Union, Optional
+from typing import List, Type, Union, Optional, Literal, Dict
 
 import httpx
 from openai import BaseModel
@@ -25,6 +25,8 @@ def llm__chat(
         presence_penalty: Optional[float] = None,
         frequency_penalty: Optional[float] = None,
         logit_bias: Optional[dict] = None,
+        modalities: Optional[List[Literal["text", "audio"]]] = None,
+        audio: Optional[Dict] = None,
         # openai v1.0+ new params
         response_format: Optional[
             Union[dict, Type[BaseModel]]
@@ -81,6 +83,8 @@ def llm__chat(
             model_list=model_list,
             drop_invalid_params=drop_invalid_params,
             user=user,
+            modalities=modalities,
+            audio=audio,
             **kwargs,
         )
         return response

edenai_apis/apis/groq/groq_api.py

@@ -74,6 +74,8 @@ def llm__chat(
         presence_penalty: Optional[float] = None,
         frequency_penalty: Optional[float] = None,
         logit_bias: Optional[dict] = None,
+        modalities: Optional[List[Literal["text", "audio"]]] = None,
+        audio: Optional[Dict] = None,
         # openai v1.0+ new params
         response_format: Optional[
             Union[dict, Type[BaseModel]]
@@ -130,6 +132,8 @@ def llm__chat(
             model_list=model_list,
             drop_invalid_params=drop_invalid_params,
             user=user,
+            modalities=modalities,
+            audio=audio,
             **kwargs,
         )
         return response

edenai_apis/apis/meta/meta_api.py

@@ -140,6 +140,8 @@ def llm__chat(
         presence_penalty: Optional[float] = None,
         frequency_penalty: Optional[float] = None,
         logit_bias: Optional[dict] = None,
+        modalities: Optional[List[Literal["text", "audio"]]] = None,
+        audio: Optional[Dict] = None,
         # openai v1.0+ new params
         response_format: Optional[
             Union[dict, Type[BaseModel]]
@@ -196,6 +198,8 @@ def llm__chat(
             model_list=model_list,
             drop_invalid_params=drop_invalid_params,
             user=user,
+            modalities=modalities,
+            audio=audio,
             **kwargs,
         )
         return response

edenai_apis/apis/microsoft/microsoft_llm_api.py

@@ -1,4 +1,4 @@
-from typing import List, Type, Union, Optional
+from typing import List, Type, Union, Optional, Literal, Dict
 
 import httpx
 from openai import BaseModel
@@ -24,6 +24,8 @@ def llm__chat(
         presence_penalty: Optional[float] = None,
         frequency_penalty: Optional[float] = None,
         logit_bias: Optional[dict] = None,
+        modalities: Optional[List[Literal["text", "audio"]]] = None,
+        audio: Optional[Dict] = None,
         # openai v1.0+ new params
         response_format: Optional[
             Union[dict, Type[BaseModel]]
@@ -80,6 +82,8 @@ def llm__chat(
             model_list=model_list,
             drop_invalid_params=drop_invalid_params,
             user=user,
+            modalities=modalities,
+            audio=audio,
             **kwargs,
         )
         return response

edenai_apis/apis/mistral/mistral_api.py

@@ -181,6 +181,8 @@ def llm__chat(
         presence_penalty: Optional[float] = None,
         frequency_penalty: Optional[float] = None,
         logit_bias: Optional[dict] = None,
+        modalities: Optional[List[Literal["text", "audio"]]] = None,
+        audio: Optional[Dict] = None,
         # openai v1.0+ new params
         response_format: Optional[
             Union[dict, Type[BaseModel]]
@@ -237,6 +239,8 @@ def llm__chat(
             model_list=model_list,
             drop_invalid_params=drop_invalid_params,
             user=user,
+            modalities=modalities,
+            audio=audio,
             **kwargs,
         )
         return response

edenai_apis/apis/openai/openai_llm_api.py

@@ -1,7 +1,6 @@
-from typing import Dict, List, Type, Union, Optional
+from typing import Dict, List, Type, Union, Optional, Literal
 import httpx
-from openai import BaseModel, OpenAI
-from openai import NOT_GIVEN
+from openai import BaseModel
 from edenai_apis.features.llm.llm_interface import LlmInterface
 from edenai_apis.features.llm.chat.chat_dataclass import ChatDataClass
 
@@ -25,6 +24,8 @@ def llm__chat(
         presence_penalty: Optional[float] = None,
         frequency_penalty: Optional[float] = None,
         logit_bias: Optional[dict] = None,
+        modalities: Optional[List[Literal["text", "audio"]]] = None,
+        audio: Optional[Dict] = None,
         # openai v1.0+ new params
         response_format: Optional[
             Union[dict, Type[BaseModel]]
@@ -80,6 +81,8 @@ def llm__chat(
             model_list=model_list,
             drop_invalid_params=drop_invalid_params,
             user=user,
+            modalities=modalities,
+            audio=audio,
             **kwargs,
         )
         return response

edenai_apis/apis/replicate/replicate_api.py

@@ -258,6 +258,8 @@ def llm__chat(
         presence_penalty: Optional[float] = None,
         frequency_penalty: Optional[float] = None,
         logit_bias: Optional[dict] = None,
+        modalities: Optional[List[Literal["text", "audio"]]] = None,
+        audio: Optional[Dict] = None,
         # openai v1.0+ new params
         response_format: Optional[
             Union[dict, Type[BaseModel]]
@@ -314,6 +316,8 @@ def llm__chat(
             model_list=model_list,
             drop_invalid_params=drop_invalid_params,
             user=user,
+            modalities=modalities,
+            audio=audio,
             **kwargs,
         )
         return response

edenai_apis/apis/together_ai/together_ai_api.py

@@ -77,6 +77,8 @@ def llm__chat(
         presence_penalty: Optional[float] = None,
         frequency_penalty: Optional[float] = None,
         logit_bias: Optional[dict] = None,
+        modalities: Optional[List[Literal["text", "audio"]]] = None,
+        audio: Optional[Dict] = None,
         # openai v1.0+ new params
         response_format: Optional[
             Union[dict, Type[BaseModel]]
@@ -133,6 +135,8 @@ def llm__chat(
             model_list=model_list,
             drop_invalid_params=drop_invalid_params,
             user=user,
+            modalities=modalities,
+            audio=audio,
             **kwargs,
         )
         return response

edenai_apis/apis/xai/xai_llm_api.py

@@ -1,4 +1,4 @@
-from typing import List, Type, Union, Optional
+from typing import List, Type, Union, Optional, Literal, Dict
 
 import httpx
 from openai import BaseModel
@@ -25,6 +25,8 @@ def llm__chat(
         presence_penalty: Optional[float] = None,
         frequency_penalty: Optional[float] = None,
         logit_bias: Optional[dict] = None,
+        modalities: Optional[List[Literal["text", "audio"]]] = None,
+        audio: Optional[Dict] = None,
         # openai v1.0+ new params
         response_format: Optional[
             Union[dict, Type[BaseModel]]
@@ -81,6 +83,8 @@ def llm__chat(
             model_list=model_list,
             drop_invalid_params=drop_invalid_params,
             user=user,
+            modalities=modalities,
+            audio=audio,
             **kwargs,
         )
         return response

edenai_apis/features/llm/llm_interface.py

@@ -1,5 +1,5 @@
 from abc import abstractmethod
-from typing import Optional, List, Dict, Type, Union
+from typing import Optional, List, Dict, Type, Union, Literal
 
 from openai import BaseModel
 import httpx
@@ -27,6 +27,8 @@ def llm__chat(
         presence_penalty: Optional[float] = None,
         frequency_penalty: Optional[float] = None,
         logit_bias: Optional[dict] = None,
+        modalities: Optional[List[Literal["text", "audio"]]] = None,
+        audio: Optional[Dict] = None,
         # openai v1.0+ new params
         response_format: Optional[
             Union[dict, Type[BaseModel]]

edenai_apis/llmengine/llm_engine.py

@@ -8,10 +8,8 @@
 from typing import Dict, List, Literal, Optional, Type, Union
 
 import httpx
-from loaders.data_loader import ProviderDataEnum
-from loaders.loaders import load_provider
 from pydantic import BaseModel
-
+from litellm.types.llms.openai import ChatCompletionAudioParam
 from edenai_apis.features.image import (
     ExplicitContentDataClass,
     GeneratedImageDataClass,
@@ -113,7 +111,9 @@ def _execute_completion(self, params: Dict, response_class: Type, **kwargs):
                 "An error occurred while parsing the response."
             ) from exc
         except ValueError as exc_v:
-            raise ProviderException("Provider returned an empty or mal-formatted response") from exc_v
+            raise ProviderException(
+                "Provider returned an empty or mal-formatted response"
+            ) from exc_v
         standardized_response = response_class(**result)
         return ResponseType[response_class](
             original_response=response.to_dict(),
@@ -743,6 +743,8 @@ def completion(
         timeout: Optional[Union[float, str, httpx.Timeout]] = None,
         temperature: Optional[float] = None,
         top_p: Optional[float] = None,
+        modalities: Optional[List[Literal["text", "audio"]]] = None,
+        audio: Optional[ChatCompletionAudioParam] = None,
         n: Optional[int] = None,
         stream: Optional[bool] = None,
         stream_options: Optional[dict] = None,
@@ -779,6 +781,10 @@ def completion(
         kwargs.pop("moderate_content", None)
         try:
             completion_params = {"messages": messages, "model": model}
+            if modalities is not None:
+                completion_params["modalities"] = modalities
+            if audio is not None:
+                completion_params["audio"] = audio
             if timeout is not None:
                 completion_params["timeout"] = timeout
             if temperature is not None: