feat(api): fix completion response breakage perhaps?

Author: stainless-app[bot]

.stats.yml

@@ -1,4 +1,4 @@
 configured_endpoints: 107
 openapi_spec_url: https://storage.googleapis.com/stainless-sdk-openapi-specs/llamastack%2Fllama-stack-client-f252873ea1e1f38fd207331ef2621c511154d5be3f4076e59cc15754fc58eee4.yml
 openapi_spec_hash: 10cbb4337a06a9fdd7d08612dd6044c3
-config_hash: 17fe64b23723fc54f2ee61c80223c3e3
+config_hash: 4138b0fa1be4c2807ad4ec538b8b38f4

api.md

@@ -11,19 +11,18 @@ from llama_stack_client.types import (
     InterleavedContent,
     InterleavedContentItem,
     Message,
+    Metric,
     ParamType,
     QueryConfig,
     QueryGeneratorConfig,
     QueryResult,
     ResponseFormat,
-    ReturnType,
     SafetyViolation,
     SamplingParams,
     ScoringResult,
     SharedTokenLogProbs,
     SystemMessage,
     ToolCall,
-    ToolCallOrString,
     ToolParamDefinition,
     ToolResponseMessage,
     UserMessage,
@@ -509,7 +508,6 @@ Types:
 ```python
 from llama_stack_client.types import (
     Event,
-    Metric,
     QueryCondition,
     QuerySpansResponse,
     SpanWithStatus,

src/llama_stack_client/resources/scoring_functions.py

@@ -21,7 +21,6 @@
 from .._base_client import make_request_options
 from ..types.scoring_fn import ScoringFn
 from ..types.scoring_fn_params_param import ScoringFnParamsParam
-from ..types.shared_params.return_type import ReturnType
 from ..types.scoring_function_list_response import ScoringFunctionListResponse
 
 __all__ = ["ScoringFunctionsResource", "AsyncScoringFunctionsResource"]
@@ -107,7 +106,7 @@ def register(
         self,
         *,
         description: str,
-        return_type: ReturnType,
+        return_type: scoring_function_register_params.ReturnType,
         scoring_fn_id: str,
         params: ScoringFnParamsParam | NotGiven = NOT_GIVEN,
         provider_id: str | NotGiven = NOT_GIVEN,
@@ -243,7 +242,7 @@ async def register(
         self,
         *,
         description: str,
-        return_type: ReturnType,
+        return_type: scoring_function_register_params.ReturnType,
         scoring_fn_id: str,
         params: ScoringFnParamsParam | NotGiven = NOT_GIVEN,
         provider_id: str | NotGiven = NOT_GIVEN,

src/llama_stack_client/types/__init__.py

@@ -7,13 +7,12 @@
 from .tool import Tool as Tool
 from .model import Model as Model
 from .trace import Trace as Trace
-from .metric import Metric as Metric
 from .shared import (
+    Metric as Metric,
     Message as Message,
     Document as Document,
     ToolCall as ToolCall,
     ParamType as ParamType,
-    ReturnType as ReturnType,
     AgentConfig as AgentConfig,
     QueryConfig as QueryConfig,
     QueryResult as QueryResult,
@@ -25,7 +24,6 @@
     SamplingParams as SamplingParams,
     BatchCompletion as BatchCompletion,
     SafetyViolation as SafetyViolation,
-    ToolCallOrString as ToolCallOrString,
     CompletionMessage as CompletionMessage,
     InterleavedContent as InterleavedContent,
     SharedTokenLogProbs as SharedTokenLogProbs,

src/llama_stack_client/types/chat_completion_response_stream_chunk.py

@@ -3,8 +3,8 @@
 from typing import List, Optional
 from typing_extensions import Literal
 
-from .metric import Metric
 from .._models import BaseModel
+from .shared.metric import Metric
 from .shared.content_delta import ContentDelta
 from .shared.shared_token_log_probs import SharedTokenLogProbs
 

src/llama_stack_client/types/scoring_fn.py

@@ -5,9 +5,23 @@
 
 from .._models import BaseModel
 from .scoring_fn_params import ScoringFnParams
-from .shared.return_type import ReturnType
 
-__all__ = ["ScoringFn"]
+__all__ = ["ScoringFn", "ReturnType"]
+
+
+class ReturnType(BaseModel):
+    type: Literal[
+        "string",
+        "number",
+        "boolean",
+        "array",
+        "object",
+        "json",
+        "union",
+        "chat_completion_input",
+        "completion_input",
+        "agent_turn_input",
+    ]
 
 
 class ScoringFn(BaseModel):

src/llama_stack_client/types/scoring_function_register_params.py

@@ -2,12 +2,11 @@
 
 from __future__ import annotations
 
-from typing_extensions import Required, TypedDict
+from typing_extensions import Literal, Required, TypedDict
 
 from .scoring_fn_params_param import ScoringFnParamsParam
-from .shared_params.return_type import ReturnType
 
-__all__ = ["ScoringFunctionRegisterParams"]
+__all__ = ["ScoringFunctionRegisterParams", "ReturnType"]
 
 
 class ScoringFunctionRegisterParams(TypedDict, total=False):
@@ -30,3 +29,20 @@ class ScoringFunctionRegisterParams(TypedDict, total=False):
 
     provider_scoring_fn_id: str
     """The ID of the provider scoring function to use for the scoring function."""
+
+
+class ReturnType(TypedDict, total=False):
+    type: Required[
+        Literal[
+            "string",
+            "number",
+            "boolean",
+            "array",
+            "object",
+            "json",
+            "union",
+            "chat_completion_input",
+            "completion_input",
+            "agent_turn_input",
+        ]
+    ]

src/llama_stack_client/types/shared/__init__.py

@@ -1,10 +1,10 @@
 # File generated from our OpenAPI spec by Stainless. See CONTRIBUTING.md for details.
 
+from .metric import Metric as Metric
 from .message import Message as Message
 from .document import Document as Document
 from .tool_call import ToolCall as ToolCall
 from .param_type import ParamType as ParamType
-from .return_type import ReturnType as ReturnType
 from .agent_config import AgentConfig as AgentConfig
 from .query_config import QueryConfig as QueryConfig
 from .query_result import QueryResult as QueryResult
@@ -18,7 +18,6 @@
 from .safety_violation import SafetyViolation as SafetyViolation
 from .completion_message import CompletionMessage as CompletionMessage
 from .interleaved_content import InterleavedContent as InterleavedContent
-from .tool_call_or_string import ToolCallOrString as ToolCallOrString
 from .tool_param_definition import ToolParamDefinition as ToolParamDefinition
 from .tool_response_message import ToolResponseMessage as ToolResponseMessage
 from .query_generator_config import QueryGeneratorConfig as QueryGeneratorConfig

src/llama_stack_client/types/shared/batch_completion.py

@@ -3,7 +3,7 @@
 from typing import List, Optional
 from typing_extensions import Literal
 
-from ..metric import Metric
+from .metric import Metric
 from ..._models import BaseModel
 from .shared_token_log_probs import SharedTokenLogProbs
 

src/llama_stack_client/types/shared/chat_completion_response.py

@@ -2,7 +2,7 @@
 
 from typing import List, Optional
 
-from ..metric import Metric
+from .metric import Metric
 from ..._models import BaseModel
 from .completion_message import CompletionMessage
 from .shared_token_log_probs import SharedTokenLogProbs