diff --git a/litellm/litellm_core_utils/get_llm_provider_logic.py b/litellm/litellm_core_utils/get_llm_provider_logic.py index 9af9dc8ddbb..566f129c5c2 100644 --- a/litellm/litellm_core_utils/get_llm_provider_logic.py +++ b/litellm/litellm_core_utils/get_llm_provider_logic.py @@ -358,7 +358,7 @@ def get_llm_provider( custom_llm_provider = "hyperbolic" dynamic_api_key = get_secret_str("HYPERBOLIC_API_KEY") elif endpoint == "https://router.requesty.ai/v1": - custom_llm_provider = "requesty" # rebind-ok: provider resolution rebinds the parameter like every branch above + custom_llm_provider = "requesty" # rebind-ok: provider resolution dynamic_api_key = get_secret_str("REQUESTY_API_KEY") elif endpoint == "https://ai-gateway.vercel.sh/v1": custom_llm_provider = "vercel_ai_gateway" @@ -844,8 +844,8 @@ def _get_openai_compatible_provider_info( ) = litellm.HyperbolicChatConfig()._get_openai_compatible_provider_info(api_base, api_key) elif custom_llm_provider == "requesty": ( - api_base, # rebind-ok: provider resolution rebinds the parameter like every branch above - dynamic_api_key, # rebind-ok: provider resolution rebinds the parameter like every branch above + api_base, # rebind-ok: provider resolution + dynamic_api_key, # rebind-ok: provider resolution ) = litellm.RequestyConfig()._get_openai_compatible_provider_info(api_base, api_key) elif custom_llm_provider == "vercel_ai_gateway": ( diff --git a/litellm/llms/requesty/chat/transformation.py b/litellm/llms/requesty/chat/transformation.py index 04610aa1e68..ed629917c7d 100644 --- a/litellm/llms/requesty/chat/transformation.py +++ b/litellm/llms/requesty/chat/transformation.py @@ -47,21 +47,21 @@ class RequestyConfig(OpenrouterConfig): except Exception: return False - def get_supported_openai_params(self, model: str) -> list[str]: # mutable-ok: OpenrouterConfig signature - supported_params: Final[list[str]] = super(OpenrouterConfig, self).get_supported_openai_params(model=model) # mutable-ok: OpenrouterConfig.get_supported_openai_params returns a list + def get_supported_openai_params(self, model: str) -> list[str]: # mutable-ok: base signature + supported_params: Final = super(OpenrouterConfig, self).get_supported_openai_params(model=model) if not self._supports_reasoning(model): return supported_params - return list(dict.fromkeys((*supported_params, *REQUESTY_REASONING_PARAMS))) # mutable-ok: OpenrouterConfig.get_supported_openai_params returns a list + return list(dict.fromkeys((*supported_params, *REQUESTY_REASONING_PARAMS))) # mutable-ok: base returns a list def map_openai_params( self, - non_default_params: dict[str, object], # mutable-ok: OpenrouterConfig signature - optional_params: dict, # mutable-ok: OpenrouterConfig signature + non_default_params: dict[str, object], # mutable-ok: base signature + optional_params: dict, # mutable-ok: base signature model: str, drop_params: bool, - ) -> dict: # mutable-ok: OpenrouterConfig signature + ) -> dict: # mutable-ok: base signature mapped_params: Final = ( - {**non_default_params, "reasoning_effort": "xhigh"} # mutable-ok: optional params are plain JSON dicts the base transform mutates + {**non_default_params, "reasoning_effort": "xhigh"} # mutable-ok: JSON params dict if non_default_params.get("reasoning_effort") == "max" else non_default_params ) @@ -70,18 +70,18 @@ class RequestyConfig(OpenrouterConfig): def transform_request( self, model: str, - messages: list[AllMessageValues], # mutable-ok: OpenrouterConfig signature - optional_params: dict, # mutable-ok: OpenrouterConfig signature - litellm_params: dict, # mutable-ok: OpenrouterConfig signature - headers: dict, # mutable-ok: OpenrouterConfig signature - ) -> dict: # mutable-ok: OpenrouterConfig signature + messages: list[AllMessageValues], # mutable-ok: base signature + optional_params: dict, # mutable-ok: base signature + litellm_params: dict, # mutable-ok: base signature + headers: dict, # mutable-ok: base signature + ) -> dict: # mutable-ok: base signature transformed_messages: Final = ( self._move_cache_control_to_content(messages) if self._supports_cache_control_in_content(model) else messages ) - extra_body: Final = optional_params.pop("extra_body", {}) # mutable-ok: extra_body is a JSON dict from the caller + extra_body: Final = optional_params.pop("extra_body", {}) # mutable-ok: caller JSON dict response: Final = super(OpenrouterConfig, self).transform_request( model, transformed_messages, optional_params, litellm_params, headers ) @@ -90,9 +90,17 @@ class RequestyConfig(OpenrouterConfig): # `messages`), otherwise a caller could route to an unauthorized model after # model-authorization and request-inspection checks have run. protected_fields: Final = frozenset({"model", "messages"}) - return {**response, **{key: value for key, value in extra_body.items() if key not in protected_fields}} # mutable-ok: request body is a plain JSON dict + return { # mutable-ok: JSON body + **response, + **{key: value for key, value in extra_body.items() if key not in protected_fields}, # mutable-ok: JSON body + } - def get_error_class(self, error_message: str, status_code: int, headers: dict | httpx.Headers) -> BaseLLMException: # mutable-ok: OpenrouterConfig signature + def get_error_class( + self, + error_message: str, + status_code: int, + headers: dict | httpx.Headers, # mutable-ok: base signature + ) -> BaseLLMException: return RequestyException( message=error_message, status_code=status_code, @@ -112,21 +120,24 @@ class RequestyConfig(OpenrouterConfig): ) +def _delta_with_reasoning(delta: dict) -> dict: # mutable-ok: wire JSON delta + return {**delta, "reasoning_content": delta.get("reasoning")} # mutable-ok: wire JSON delta + + class RequestyChatCompletionStreamingHandler(BaseModelResponseIterator): - def chunk_parser(self, chunk: dict) -> ModelResponseStream: # mutable-ok: OpenrouterConfig signature + def chunk_parser(self, chunk: dict) -> ModelResponseStream: # mutable-ok: base signature try: if "error" in chunk: error_chunk: Final = chunk["error"] + error_metadata: Final = error_chunk.get("metadata", {}) # mutable-ok: JSON metadata raise RequestyException( - message="Message: {}, Metadata: {}".format(error_chunk["message"], error_chunk.get("metadata", {})), # mutable-ok: error metadata is a plain JSON dict + message="Message: {}, Metadata: {}".format(error_chunk["message"], error_metadata), status_code=error_chunk["code"], - headers=error_chunk.get("metadata", {}).get("headers", {}), # mutable-ok: headers are a plain JSON dict + headers=error_metadata.get("headers", {}), # mutable-ok: JSON headers ) - choices: Final = [ # mutable-ok: StreamingChoices expects a list - StreamingChoices( - **{**choice, "delta": {**choice["delta"], "reasoning_content": choice["delta"].get("reasoning")}} # mutable-ok: wire delta is a plain JSON dict - ) + choices: Final = [ # mutable-ok: choices list + StreamingChoices(**{**choice, "delta": _delta_with_reasoning(choice["delta"])}) # mutable-ok: JSON for choice in chunk["choices"] ] return ModelResponseStream( @@ -141,5 +152,5 @@ class RequestyChatCompletionStreamingHandler(BaseModelResponseIterator): raise RequestyException( message=f"KeyError: {e}, Got unexpected response from Requesty: {chunk}", status_code=400, - headers={"Content-Type": "application/json"}, # mutable-ok: headers are a plain JSON dict + headers={"Content-Type": "application/json"}, # mutable-ok: JSON headers )