Rebase to main

Signed-off-by: Mert Toslali <[email protected]>

Author: toslali-ibm

src/guidellm/schemas/request_stats.py

@@ -48,6 +48,9 @@ class GenerativeRequestStats(StandardBaseDict):
     request_type: GenerativeRequestType | str = Field(
         description="Type of generative request (text_completion or chat_completion)"
     )
+    response_id: str | None = Field(
+        default=None, description="vLLM-assigned unique request ID"
+    )
     request_args: str | None = Field(
         default=None, description="Backend arguments used for this request"
     )

src/guidellm/schemas/response.py

@@ -112,6 +112,7 @@ def compile_stats(
 
         return GenerativeRequestStats(
             request_id=self.request_id,
+            response_id=self.response_id,
             request_type=request.request_type,
             request_args=str(
                 request.arguments.model_dump() if request.arguments else {}