[bugfix] /chat/completion doesn't read extra_body for diffusion model (#2042)

Signed-off-by: Huang, Zeyu <11222265+fhfuih@users.noreply.github.com>
Signed-off-by: Samit <285365963@qq.com>
Co-authored-by: Samit <285365963@qq.com>

.buildkite/test-nightly.yml

@@ -185,7 +185,7 @@ steps:
   - label: ":email: Nightly Collection & Email"
     key: nightly-perf-distribution
     depends_on:
-    - nightly-performance
+      - nightly-performance
     if: build.env("NIGHTLY") == "1"
     commands:
       - pip install openpyxl

vllm_omni/entrypoints/openai/serving_chat.py

@@ -297,7 +297,13 @@ class OmniOpenAIServingChat(OpenAIServingChat, AudioMixin):
                 if not extracted_prompt:
                     return self.create_error_response("No text prompt found in messages")
 
-                extra_body = getattr(request, "extra_body", None) or {}
+                # [NOTE] When sending request via openai client Python library,
+                #   `extra_body` is flattented and merged into the payload's root.
+                #   These extra fields are accessible via `model_extra` property (from Pydantic base class).
+                #   When sending raw request with curl, no flattening happens. Directly read the `extra_body` dict.
+                extra_body = getattr(request, "extra_body", None)
+                if not extra_body:
+                    extra_body = request.model_extra or {}
                 height = extra_body.get("height")
                 width = extra_body.get("width")
                 if "size" in extra_body:
@@ -2036,7 +2042,13 @@ class OmniOpenAIServingChat(OpenAIServingChat, AudioMixin):
 
             # Extract generation parameters from extra_body (preferred)
             # Reference: text_to_image.py and text_to_video.py for supported parameters
-            extra_body = getattr(request, "extra_body", None) or {}
+            # [NOTE] When sending request via openai client Python library,
+            #   `extra_body` is flattented and merged into the payload's root.
+            #   These extra fields are accessible via `model_extra` property (from Pydantic base class).
+            #   When sending raw request with curl, no flattening happens. Directly read the `extra_body` dict.
+            extra_body = getattr(request, "extra_body", None)
+            if not extra_body:
+                extra_body = request.model_extra or {}
 
             # Parse size if provided (supports "1024x1024" format)
             height = extra_body.get("height")