chore: apply Gemini review

Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>

Author: max-wittig

src/vllm_router/utils.py

@@ -77,7 +77,7 @@ class ModelType(enum.Enum):
     @staticmethod
     def get_url(model_type: str):
         match ModelType[model_type]:
-            case ModelType.chat:
+            case ModelType.chat | ModelType.vision:
                 return "/v1/chat/completions"
             case ModelType.completion:
                 return "/v1/completions"
@@ -89,8 +89,6 @@ def get_url(model_type: str):
                 return "/v1/score"
             case ModelType.transcription:
                 return "/v1/audio/transcriptions"
-            case ModelType.vision:
-                return "/v1/chat/completions"
 
     @staticmethod
     def get_test_payload(model_type: str):
@@ -233,7 +231,7 @@ def is_model_healthy(url: str, model: str, model_type: str) -> bool:
             # we will use pre-generated silent wav bytes
             response = requests.post(
                 f"{url}{model_url}",
-                files=ModelType[model_type].get_test_payload(
+                files=ModelType.get_test_payload(
                     model_type
                 ),  # multipart/form-data
                 data={"model": model},
@@ -245,7 +243,7 @@ def is_model_healthy(url: str, model: str, model_type: str) -> bool:
                 f"{url}{model_url}",
                 headers={"Content-Type": "application/json"},
                 json={"model": model}
-                | ModelType[model_type].get_test_payload(model_type),
+                | ModelType.get_test_payload(model_type),
                 timeout=10,
             )