Fixes; make inference tests pass with newer tool call types

llama_stack/distribution/store/registry.py

@@ -35,7 +35,7 @@ class DistributionRegistry(Protocol):
 
 
 REGISTER_PREFIX = "distributions:registry"
-KEY_VERSION = "v4"
+KEY_VERSION = "v5"
 KEY_FORMAT = f"{REGISTER_PREFIX}:{KEY_VERSION}::" + "{type}:{identifier}"
 
 

llama_stack/providers/utils/inference/openai_compat.py

@@ -142,7 +142,7 @@ async def process_completion_stream_response(
             text = ""
             continue
         yield CompletionResponseStreamChunk(
-            delta=TextDelta(text=text),
+            delta=text,
             stop_reason=stop_reason,
         )
         if finish_reason:
@@ -153,7 +153,7 @@ async def process_completion_stream_response(
             break
 
     yield CompletionResponseStreamChunk(
-        delta=TextDelta(text=""),
+        delta="",
         stop_reason=stop_reason,
     )
 

llama_stack/providers/utils/inference/prompt_adapter.py

@@ -265,6 +265,7 @@ def chat_completion_request_to_messages(
     For eg. for llama_3_1, add system message with the appropriate tools or
     add user messsage for custom tools, etc.
     """
+    assert llama_model is not None, "llama_model is required"
     model = resolve_model(llama_model)
     if model is None:
         log.error(f"Could not resolve model {llama_model}")