Configurable max_tokens/max_completion_tokens key (#399)

## Summary

<!--
Include a short paragraph of the changes introduced in this PR.
If this PR requires additional context or rationale, explain why
the changes are necessary.
-->

Makes the `max_tokens` request key configurable through an environment
variable per endpoint type. Defaults to `max_tokens` for legacy
`completions` and `max_completion_tokens` for `chat/completions`

## Details

<!--
Provide a detailed list of all changes introduced in this pull request.
-->
- Add the `GUIDELLM__OPENAI__MAX_OUTPUT_KEY` config option which is a
dict mapping from route name -> output tokens key. Default is
`{"text_completions": "max_tokens", "chat_completions":
"max_completion_tokens"}`

## Test Plan

<!--
List the steps needed to test this PR.
-->
-

## Related Issues

<!--
Link any relevant issues that this PR addresses.
-->
- Closes #395
- Closes #269
- Related #210

---

- [x] "I certify that all code in this PR is my own, except as noted
below."

## Use of AI

- [ ] Includes AI-assisted code completion
- [ ] Includes code generated by an AI application
- [ ] Includes AI-generated tests (NOTE: AI written tests should have a
docstring that includes `## WRITTEN BY AI ##`)

---------

Signed-off-by: Tyler Michael Smith <[email protected]>
Signed-off-by: Samuel Monson <[email protected]>
Co-authored-by: Tyler Michael Smith <[email protected]>

Author: sjmonson

src/guidellm/backend/openai.py

@@ -31,10 +31,11 @@
 TEXT_COMPLETIONS_PATH = "/v1/completions"
 CHAT_COMPLETIONS_PATH = "/v1/chat/completions"
 
-EndpointType = Literal["chat_completions", "models", "text_completions"]
-CHAT_COMPLETIONS: EndpointType = "chat_completions"
+CompletionEndpointType = Literal["text_completions", "chat_completions"]
+EndpointType = Union[Literal["models"], CompletionEndpointType]
+CHAT_COMPLETIONS: CompletionEndpointType = "chat_completions"
 MODELS: EndpointType = "models"
-TEXT_COMPLETIONS: EndpointType = "text_completions"
+TEXT_COMPLETIONS: CompletionEndpointType = "text_completions"
 
 
 @Backend.register("openai_http")
@@ -447,7 +448,7 @@ def _extra_body(self, endpoint_type: EndpointType) -> dict[str, Any]:
 
     def _completions_payload(
         self,
-        endpoint_type: EndpointType,
+        endpoint_type: CompletionEndpointType,
         orig_kwargs: Optional[dict],
         max_output_tokens: Optional[int],
         **kwargs,
@@ -467,8 +468,10 @@ def _completions_payload(
                 self.__class__.__name__,
                 max_output_tokens or self.max_output_tokens,
             )
-            payload["max_tokens"] = max_output_tokens or self.max_output_tokens
-            payload["max_completion_tokens"] = payload["max_tokens"]
+            max_output_key = settings.openai.max_output_key.get(
+                endpoint_type, "max_tokens"
+            )
+            payload[max_output_key] = max_output_tokens or self.max_output_tokens
 
             if max_output_tokens:
                 # only set stop and ignore_eos if max_output_tokens set at request level

src/guidellm/config.py

@@ -88,6 +88,10 @@ class OpenAISettings(BaseModel):
     base_url: str = "http://localhost:8000"
     max_output_tokens: int = 16384
     verify: bool = True
+    max_output_key: dict[Literal["text_completions", "chat_completions"], str] = {
+        "text_completions": "max_tokens",
+        "chat_completions": "max_completion_tokens",
+    }
 
 
 class ReportGenerationSettings(BaseModel):

tests/unit/conftest.py

@@ -132,7 +132,6 @@ async def _mock_completions_response(request) -> AsyncIterable[str]:
             assert payload["stream_options"] == {"include_usage": True}
             assert payload["prompt"] is not None
             assert len(payload["prompt"]) > 0
-            assert payload["max_completion_tokens"] > 0
             assert payload["max_tokens"] > 0
 
             return httpx.Response(  # type: ignore
@@ -141,7 +140,7 @@ async def _mock_completions_response(request) -> AsyncIterable[str]:
                     type_="text",
                     prompt=payload["prompt"],
                     output_token_count=(
-                        payload["max_completion_tokens"]
+                        payload["max_tokens"]
                         if payload.get("ignore_eos", False)
                         else None
                     ),
@@ -162,7 +161,6 @@ async def _mock_chat_completions_response(request):
             assert payload["messages"] is not None
             assert len(payload["messages"]) > 0
             assert payload["max_completion_tokens"] > 0
-            assert payload["max_tokens"] > 0
 
             return httpx.Response(  # type: ignore
                 200,