featu: support passing "extra body" throught to providers

# What does this PR do?
Allows passing through extra_body parameters to inference providers.


closes #2720

## Test Plan
CI and added new test

llama_stack/providers/inline/agents/meta_reference/agent_instance.py

@@ -49,7 +49,7 @@ from llama_stack.apis.inference import (
     Inference,
     Message,
     OpenAIAssistantMessageParam,
-    OpenAIChatCompletionRequest,
+    OpenAIChatCompletionRequestWithExtraBody,
     OpenAIDeveloperMessageParam,
     OpenAIMessageParam,
     OpenAISystemMessageParam,
@@ -583,7 +583,7 @@ class ChatAgent(ShieldRunnerMixin):
                 max_tokens = getattr(sampling_params, "max_tokens", None)
 
                 # Use OpenAI chat completion
-                params = OpenAIChatCompletionRequest(
+                params = OpenAIChatCompletionRequestWithExtraBody(
                     model=self.agent_config.model,
                     messages=openai_messages,
                     tools=openai_tools if openai_tools else None,

llama_stack/providers/inline/agents/meta_reference/responses/streaming.py

@@ -49,7 +49,7 @@ from llama_stack.apis.inference import (
     OpenAIAssistantMessageParam,
     OpenAIChatCompletion,
     OpenAIChatCompletionChunk,
-    OpenAIChatCompletionRequest,
+    OpenAIChatCompletionRequestWithExtraBody,
     OpenAIChatCompletionToolCall,
     OpenAIChoice,
     OpenAIMessageParam,
@@ -169,7 +169,7 @@ class StreamingResponseOrchestrator:
                 # (some providers don't support non-empty response_format when tools are present)
                 response_format = None if self.ctx.response_format.type == "text" else self.ctx.response_format
                 logger.debug(f"calling openai_chat_completion with tools: {self.ctx.chat_tools}")
-                params = OpenAIChatCompletionRequest(
+                params = OpenAIChatCompletionRequestWithExtraBody(
                     model=self.ctx.model,
                     messages=messages,
                     tools=self.ctx.chat_tools,