fix: React Agent for non-llama models

src/llama_stack_client/lib/agents/agent.py

@@ -162,11 +162,11 @@ def _create_turn_streaming(
         while not is_turn_complete:
             is_turn_complete = True
             for chunk in turn_response:
-                tool_calls = self._get_tool_calls(chunk)
                 if hasattr(chunk, "error"):
                     yield chunk
                     return
-                elif not tool_calls:
+                tool_calls = self._get_tool_calls(chunk)
+                if not tool_calls:
                     yield chunk
                 else:
                     is_turn_complete = False

src/llama_stack_client/lib/agents/client_tool.py

@@ -163,7 +163,8 @@ def get_params_definition(self) -> Dict[str, Parameter]:
                 params[name] = Parameter(
                     name=name,
                     description=param_doc or f"Parameter {name}",
-                    parameter_type=type_hint.__name__,
+                    # Hack: litellm/openai expects "string" for str type
+                    parameter_type=type_hint.__name__ if type_hint.__name__ != "str" else "string",
                     default=(param.default if param.default != inspect.Parameter.empty else None),
                     required=is_required,
                 )

src/llama_stack_client/lib/agents/event_logger.py

@@ -63,7 +63,8 @@ def yield_printable_events(self, chunk: Any) -> Iterator[TurnStreamPrintableEven
         for printable_event in self._yield_printable_events(chunk, self.previous_event_type, self.previous_step_type):
             yield printable_event
 
-        self.previous_event_type, self.previous_step_type = self._get_event_type_step_type(chunk)
+        if not hasattr(chunk, "error"):
+            self.previous_event_type, self.previous_step_type = self._get_event_type_step_type(chunk)
 
     def _yield_printable_events(
         self, chunk: Any, previous_event_type: Optional[str] = None, previous_step_type: Optional[str] = None

src/llama_stack_client/lib/agents/react/agent.py

@@ -3,29 +3,18 @@
 #
 # This source code is licensed under the terms described in the LICENSE file in
 # the root directory of this source tree.
-from typing import Any, Dict, Optional, Tuple
+from typing import Optional, Tuple
 
 from llama_stack_client import LlamaStackClient
 from llama_stack_client.types.agent_create_params import AgentConfig
-from pydantic import BaseModel
+
 
 from ..agent import Agent
 from ..client_tool import ClientTool
 from ..tool_parser import ToolParser
 from .prompts import DEFAULT_REACT_AGENT_SYSTEM_PROMPT_TEMPLATE
 
-from .tool_parser import ReActToolParser
-
-
-class Action(BaseModel):
-    tool_name: str
-    tool_params: Dict[str, Any]
-
-
-class ReActOutput(BaseModel):
-    thought: str
-    action: Optional[Action] = None
-    answer: Optional[str] = None
+from .tool_parser import ReActToolParser, ReActOutput
 
 
 class ReActAgent(Agent):
@@ -97,7 +86,7 @@ def get_tool_defs():
             agent_config = custom_agent_config
 
         if json_response_format:
-            agent_config.response_format = {
+            agent_config["response_format"] = {
                 "type": "json_schema",
                 "json_schema": ReActOutput.model_json_schema(),
             }

src/llama_stack_client/lib/agents/react/tool_parser.py

@@ -5,23 +5,28 @@
 # the root directory of this source tree.
 
 from pydantic import BaseModel, ValidationError
-from typing import Dict, Any, Optional, List
+from typing import Optional, List, Union
 from ..tool_parser import ToolParser
 from llama_stack_client.types.shared.completion_message import CompletionMessage
 from llama_stack_client.types.shared.tool_call import ToolCall
 
 import uuid
 
 
+class Param(BaseModel):
+    name: str
+    value: Union[str, int, float, bool]
+
+
 class Action(BaseModel):
     tool_name: str
-    tool_params: Dict[str, Any]
+    tool_params: List[Param]
 
 
 class ReActOutput(BaseModel):
     thought: str
-    action: Optional[Action] = None
-    answer: Optional[str] = None
+    action: Optional[Action]
+    answer: Optional[str]
 
 
 class ReActToolParser(ToolParser):
@@ -40,8 +45,9 @@ def get_tool_calls(self, output_message: CompletionMessage) -> List[ToolCall]:
         if react_output.action:
             tool_name = react_output.action.tool_name
             tool_params = react_output.action.tool_params
+            params = {param.name: param.value for param in tool_params}
             if tool_name and tool_params:
                 call_id = str(uuid.uuid4())
-                tool_calls = [ToolCall(call_id=call_id, tool_name=tool_name, arguments=tool_params)]
+                tool_calls = [ToolCall(call_id=call_id, tool_name=tool_name, arguments=params)]
 
         return tool_calls