refactor(rust_bridge): share call_hook instead of per-route native lambdas

Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>

litellm/chat_completions/dispatch.py

@@ -10,7 +10,7 @@ from litellm.rust_bridge.chat_completions.entrypoints import (
     NATIVE_COMPLETION,
     LiteLLMChatCompletionsRequest,
 )
-from litellm.rust_bridge.dispatch import PublicDispatch
+from litellm.rust_bridge.dispatch import PublicDispatch, call_hook
 from litellm.rust_bridge.public_call import (
     bind,
     optional_bool,
@@ -103,7 +103,7 @@ def completion(
         kwargs,
         python=python,
         binding=NATIVE_COMPLETION,
-        native=lambda hook, request, call_args, call_kwargs: hook(request, call_args, call_kwargs),
+        native=call_hook,
     )
 
 
@@ -114,7 +114,7 @@ async def acompletion(*args: object, **kwargs: object) -> ChatResult:  # kwargs-
         kwargs,
         python=python,
         binding=NATIVE_ACOMPLETION,
-        native=lambda hook, request, call_args, call_kwargs: hook(request, call_args, call_kwargs),
+        native=call_hook,
     )
 
 

litellm/messages/dispatch.py

@@ -5,7 +5,7 @@ from typing import Final, TypeAlias, cast  # noqa: TID251  # native binding sele
 
 from litellm.llms.anthropic.experimental_pass_through.messages import handler as main
 from litellm.rust_bridge.catalog import Context, Delivery, Route
-from litellm.rust_bridge.dispatch import PublicDispatch
+from litellm.rust_bridge.dispatch import PublicDispatch, call_hook
 from litellm.rust_bridge.messages.entrypoints import (
     NATIVE_AMESSAGES,
     NATIVE_MESSAGES,
@@ -102,7 +102,7 @@ def anthropic_messages_handler(
         kwargs,
         python=python,
         binding=NATIVE_MESSAGES,
-        native=lambda hook, request, call_args, call_kwargs: hook(request, call_args, call_kwargs),
+        native=call_hook,
     )
 
 
@@ -113,7 +113,7 @@ async def anthropic_messages(*args: object, **kwargs: object) -> MessagesResult:
         kwargs,
         python=python,
         binding=NATIVE_AMESSAGES,
-        native=lambda hook, request, call_args, call_kwargs: hook(request, call_args, call_kwargs),
+        native=call_hook,
     )
 
 

litellm/ocr/dispatch.py

@@ -7,7 +7,7 @@ from litellm.llms.base_llm.ocr.transformation import OCRResponse
 from litellm.ocr import main
 from litellm.ocr.main import convert_file_document_to_url_document, get_mime_type
 from litellm.rust_bridge.catalog import Context, Route
-from litellm.rust_bridge.dispatch import PublicDispatch
+from litellm.rust_bridge.dispatch import PublicDispatch, call_hook
 from litellm.rust_bridge.ocr.entrypoints import NATIVE_AOCR, NATIVE_OCR, LiteLLMOcrRequest
 
 __all__ = ("aocr", "convert_file_document_to_url_document", "get_mime_type", "ocr")
@@ -77,7 +77,7 @@ def ocr(
         kwargs,
         python=_PYTHON_OCR,
         binding=NATIVE_OCR,
-        native=lambda hook, request, call_args, call_kwargs: hook(request, call_args, call_kwargs),
+        native=call_hook,
     )
 
 
@@ -87,5 +87,5 @@ async def aocr(*args: object, **kwargs: object) -> OCRResponse:  # kwargs-ok: pr
         kwargs,
         python=_PYTHON_AOCR,
         binding=NATIVE_AOCR,
-        native=lambda hook, request, call_args, call_kwargs: hook(request, call_args, call_kwargs),
+        native=call_hook,
     )

litellm/responses/dispatch.py

@@ -6,7 +6,7 @@ from typing import Final, TypeAlias, cast  # noqa: TID251  # native binding sele
 from litellm.responses import main
 from litellm.responses.streaming_iterator import BaseResponsesAPIStreamingIterator
 from litellm.rust_bridge.catalog import Context, Delivery, Route
-from litellm.rust_bridge.dispatch import PublicDispatch
+from litellm.rust_bridge.dispatch import PublicDispatch, call_hook
 from litellm.rust_bridge.public_call import bind, optional_bool, optional_mapping, optional_str, signature
 from litellm.rust_bridge.responses.entrypoints import (
     NATIVE_ARESPONSES,
@@ -95,7 +95,7 @@ def responses(
         kwargs,
         python=python,
         binding=NATIVE_RESPONSES,
-        native=lambda hook, request, call_args, call_kwargs: hook(request, call_args, call_kwargs),
+        native=call_hook,
     )
 
 
@@ -106,7 +106,7 @@ async def aresponses(*args: object, **kwargs: object) -> ResponsesResult:  # kwa
         kwargs,
         python=python,
         binding=NATIVE_ARESPONSES,
-        native=lambda hook, request, call_args, call_kwargs: hook(request, call_args, call_kwargs),
+        native=call_hook,
     )
 
 

litellm/rust_bridge/dispatch.py

@@ -14,6 +14,17 @@ RequestT = TypeVar("RequestT")
 NativeT = TypeVar("NativeT")
 ResultT = TypeVar("ResultT")
 
+NativeHook = Callable[[RequestT, tuple[object, ...], Mapping[str, object]], ResultT]
+
+
+def call_hook(
+    hook: NativeHook[RequestT, ResultT],
+    request: RequestT,
+    args: tuple[object, ...],
+    kwargs: Mapping[str, object],
+) -> ResultT:
+    return hook(request, args, kwargs)
+
 
 @dataclass(frozen=True, slots=True)
 class PublicDispatch(Generic[RequestT]):