refactor inference request structure

handler.py

@@ -31,9 +31,10 @@ class EndpointHandler():
         and, if both image and text lists are provided, calculate similarity scores between them.
         
         Args:
-            data (Dict[str, Any]): A dictionary containing the following keys:
-                - "image_list" (List[str]): A list of base64-encoded images.
-                - "text_list" (List[str]): A list of text strings.
+            data (Dict[str, Any]): A dictionary containing the following key:
+                - "inputs" (Dict[str, list]): A dictionary containing the following keys:
+                    - "image_list" (List[str]): A list of base64-encoded images.
+                    - "text_list" (List[str]): A list of text strings.
 
         Returns:
             Dict[str, list]: A dictionary containing the following keys:
@@ -42,8 +43,9 @@ class EndpointHandler():
                 - "similarity_scores" (List[List[float]]): A list of similarity scores between image and text embeddings.
                                                         Empty if either "image_list" or "text_list" is empty.
         """
-        image_list = data.get("image_list", []) # list of b64 images
-        text_list = data.get("text_list", []) # list of texts
+        inputs = data.get("inputs", {})
+        image_list = inputs.get("image_list", []) # list of b64 images
+        text_list = inputs.get("text_list", []) # list of texts
 
         image_features = self.get_image_embeddings(image_list) if len(image_list) > 0 else None
         text_features = self.get_text_embeddings(text_list) if len(text_list) > 0 else None