chore: api sync, deprecate allow_resume_turn + rename task_config->benchmark_config (Sync updates from stainless branch: yanxi0830/dev) (#176)

# What does this PR do?
- Adapt to API changes in
llamastack/llama-stack#1397 and
llamastack/llama-stack#1377

[//]: # (If resolving an issue, uncomment and update the line below)
[//]: # (Closes #[issue-number])

## Test Plan
```
LLAMA_STACK_BASE_URL=http://localhost:8321 pytest -v tests/integration/agents/test_agents.py--inference-model "meta-llama/Llama-3.3-70B-Instruct" --record-responses
```

```
pytest -v -s --nbval-lax ./docs/notebooks/Llama_Stack_Benchmark_Evals.ipynb
```

[//]: # (## Documentation)
[//]: # (- [ ] Added a Changelog entry if the change is significant)

Author: yanxi0830

docs/cli_reference.md

@@ -182,7 +182,7 @@ Options:
 - `--num-examples`: Optional. Number of examples to evaluate (useful for debugging)
 - `--visualize`: Optional flag. If set, visualizes evaluation results after completion
 
-Example eval_task_config.json:
+Example eval_benchmark_config.json:
 ```json
 {
     "type": "benchmark",

src/llama_stack_client/_base_client.py

@@ -9,7 +9,6 @@
 import inspect
 import logging
 import platform
-import warnings
 import email.utils
 from types import TracebackType
 from random import random
@@ -36,7 +35,7 @@
 import httpx
 import distro
 import pydantic
-from httpx import URL, Limits
+from httpx import URL
 from pydantic import PrivateAttr
 
 from . import _exceptions
@@ -51,19 +50,16 @@
     Timeout,
     NotGiven,
     ResponseT,
-    Transport,
     AnyMapping,
     PostParser,
-    ProxiesTypes,
     RequestFiles,
     HttpxSendArgs,
-    AsyncTransport,
     RequestOptions,
     HttpxRequestFiles,
     ModelBuilderProtocol,
 )
 from ._utils import is_dict, is_list, asyncify, is_given, lru_cache, is_mapping
-from ._compat import model_copy, model_dump
+from ._compat import PYDANTIC_V2, model_copy, model_dump
 from ._models import GenericModel, FinalRequestOptions, validate_type, construct_type
 from ._response import (
     APIResponse,
@@ -207,6 +203,9 @@ def _set_private_attributes(
         model: Type[_T],
         options: FinalRequestOptions,
     ) -> None:
+        if PYDANTIC_V2 and getattr(self, "__pydantic_private__", None) is None:
+            self.__pydantic_private__ = {}
+
         self._model = model
         self._client = client
         self._options = options
@@ -292,6 +291,9 @@ def _set_private_attributes(
         client: AsyncAPIClient,
         options: FinalRequestOptions,
     ) -> None:
+        if PYDANTIC_V2 and getattr(self, "__pydantic_private__", None) is None:
+            self.__pydantic_private__ = {}
+
         self._model = model
         self._client = client
         self._options = options
@@ -331,9 +333,6 @@ class BaseClient(Generic[_HttpxClientT, _DefaultStreamT]):
     _base_url: URL
     max_retries: int
     timeout: Union[float, Timeout, None]
-    _limits: httpx.Limits
-    _proxies: ProxiesTypes | None
-    _transport: Transport | AsyncTransport | None
     _strict_response_validation: bool
     _idempotency_header: str | None
     _default_stream_cls: type[_DefaultStreamT] | None = None
@@ -346,19 +345,13 @@ def __init__(
         _strict_response_validation: bool,
         max_retries: int = DEFAULT_MAX_RETRIES,
         timeout: float | Timeout | None = DEFAULT_TIMEOUT,
-        limits: httpx.Limits,
-        transport: Transport | AsyncTransport | None,
-        proxies: ProxiesTypes | None,
         custom_headers: Mapping[str, str] | None = None,
         custom_query: Mapping[str, object] | None = None,
     ) -> None:
         self._version = version
         self._base_url = self._enforce_trailing_slash(URL(base_url))
         self.max_retries = max_retries
         self.timeout = timeout
-        self._limits = limits
-        self._proxies = proxies
-        self._transport = transport
         self._custom_headers = custom_headers or {}
         self._custom_query = custom_query or {}
         self._strict_response_validation = _strict_response_validation
@@ -794,46 +787,11 @@ def __init__(
         base_url: str | URL,
         max_retries: int = DEFAULT_MAX_RETRIES,
         timeout: float | Timeout | None | NotGiven = NOT_GIVEN,
-        transport: Transport | None = None,
-        proxies: ProxiesTypes | None = None,
-        limits: Limits | None = None,
         http_client: httpx.Client | None = None,
         custom_headers: Mapping[str, str] | None = None,
         custom_query: Mapping[str, object] | None = None,
         _strict_response_validation: bool,
     ) -> None:
-        kwargs: dict[str, Any] = {}
-        if limits is not None:
-            warnings.warn(
-                "The `connection_pool_limits` argument is deprecated. The `http_client` argument should be passed instead",
-                category=DeprecationWarning,
-                stacklevel=3,
-            )
-            if http_client is not None:
-                raise ValueError("The `http_client` argument is mutually exclusive with `connection_pool_limits`")
-        else:
-            limits = DEFAULT_CONNECTION_LIMITS
-
-        if transport is not None:
-            kwargs["transport"] = transport
-            warnings.warn(
-                "The `transport` argument is deprecated. The `http_client` argument should be passed instead",
-                category=DeprecationWarning,
-                stacklevel=3,
-            )
-            if http_client is not None:
-                raise ValueError("The `http_client` argument is mutually exclusive with `transport`")
-
-        if proxies is not None:
-            kwargs["proxies"] = proxies
-            warnings.warn(
-                "The `proxies` argument is deprecated. The `http_client` argument should be passed instead",
-                category=DeprecationWarning,
-                stacklevel=3,
-            )
-            if http_client is not None:
-                raise ValueError("The `http_client` argument is mutually exclusive with `proxies`")
-
         if not is_given(timeout):
             # if the user passed in a custom http client with a non-default
             # timeout set then we use that timeout.
@@ -854,12 +812,9 @@ def __init__(
 
         super().__init__(
             version=version,
-            limits=limits,
             # cast to a valid type because mypy doesn't understand our type narrowing
             timeout=cast(Timeout, timeout),
-            proxies=proxies,
             base_url=base_url,
-            transport=transport,
             max_retries=max_retries,
             custom_query=custom_query,
             custom_headers=custom_headers,
@@ -869,9 +824,6 @@ def __init__(
             base_url=base_url,
             # cast to a valid type because mypy doesn't understand our type narrowing
             timeout=cast(Timeout, timeout),
-            limits=limits,
-            follow_redirects=True,
-            **kwargs,  # type: ignore
         )
 
     def is_closed(self) -> bool:
@@ -1366,45 +1318,10 @@ def __init__(
         _strict_response_validation: bool,
         max_retries: int = DEFAULT_MAX_RETRIES,
         timeout: float | Timeout | None | NotGiven = NOT_GIVEN,
-        transport: AsyncTransport | None = None,
-        proxies: ProxiesTypes | None = None,
-        limits: Limits | None = None,
         http_client: httpx.AsyncClient | None = None,
         custom_headers: Mapping[str, str] | None = None,
         custom_query: Mapping[str, object] | None = None,
     ) -> None:
-        kwargs: dict[str, Any] = {}
-        if limits is not None:
-            warnings.warn(
-                "The `connection_pool_limits` argument is deprecated. The `http_client` argument should be passed instead",
-                category=DeprecationWarning,
-                stacklevel=3,
-            )
-            if http_client is not None:
-                raise ValueError("The `http_client` argument is mutually exclusive with `connection_pool_limits`")
-        else:
-            limits = DEFAULT_CONNECTION_LIMITS
-
-        if transport is not None:
-            kwargs["transport"] = transport
-            warnings.warn(
-                "The `transport` argument is deprecated. The `http_client` argument should be passed instead",
-                category=DeprecationWarning,
-                stacklevel=3,
-            )
-            if http_client is not None:
-                raise ValueError("The `http_client` argument is mutually exclusive with `transport`")
-
-        if proxies is not None:
-            kwargs["proxies"] = proxies
-            warnings.warn(
-                "The `proxies` argument is deprecated. The `http_client` argument should be passed instead",
-                category=DeprecationWarning,
-                stacklevel=3,
-            )
-            if http_client is not None:
-                raise ValueError("The `http_client` argument is mutually exclusive with `proxies`")
-
         if not is_given(timeout):
             # if the user passed in a custom http client with a non-default
             # timeout set then we use that timeout.
@@ -1426,11 +1343,8 @@ def __init__(
         super().__init__(
             version=version,
             base_url=base_url,
-            limits=limits,
             # cast to a valid type because mypy doesn't understand our type narrowing
             timeout=cast(Timeout, timeout),
-            proxies=proxies,
-            transport=transport,
             max_retries=max_retries,
             custom_query=custom_query,
             custom_headers=custom_headers,
@@ -1440,9 +1354,6 @@ def __init__(
             base_url=base_url,
             # cast to a valid type because mypy doesn't understand our type narrowing
             timeout=cast(Timeout, timeout),
-            limits=limits,
-            follow_redirects=True,
-            **kwargs,  # type: ignore
         )
 
     def is_closed(self) -> bool:

src/llama_stack_client/_client.py

@@ -124,7 +124,7 @@ def __init__(
         _strict_response_validation: bool = False,
         provider_data: Mapping[str, Any] | None = None,
     ) -> None:
-        """Construct a new synchronous llama-stack-client client instance.
+        """Construct a new synchronous LlamaStackClient client instance.
 
         This automatically infers the `api_key` argument from the `LLAMA_STACK_API_KEY` environment variable if it is not provided.
         """
@@ -340,7 +340,7 @@ def __init__(
         _strict_response_validation: bool = False,
         provider_data: Mapping[str, Any] | None = None,
     ) -> None:
-        """Construct a new async llama-stack-client client instance.
+        """Construct a new async AsyncLlamaStackClient client instance.
 
         This automatically infers the `api_key` argument from the `LLAMA_STACK_API_KEY` environment variable if it is not provided.
         """

src/llama_stack_client/lib/agents/agent.py

@@ -151,7 +151,6 @@ def _create_turn_streaming(
             stream=True,
             documents=documents,
             toolgroups=toolgroups,
-            allow_turn_resume=True,
         )
 
         # 2. process turn and resume if there's a tool call

src/llama_stack_client/lib/cli/eval/run_benchmark.py

@@ -13,12 +13,7 @@
 from tqdm.rich import tqdm
 
 from ..common.utils import create_bar_chart
-from .utils import (
-    aggregate_accuracy,
-    aggregate_average,
-    aggregate_categorical_count,
-    aggregate_median,
-)
+from .utils import aggregate_accuracy, aggregate_average, aggregate_categorical_count, aggregate_median
 
 
 @click.command("run-benchmark")
@@ -110,7 +105,7 @@ def run_benchmark(
                 benchmark_id=benchmark_id,
                 input_rows=[r],
                 scoring_functions=scoring_functions,
-                task_config={
+                benchmark_config={
                     "type": "benchmark",
                     "eval_candidate": {
                         "type": "model",

src/llama_stack_client/resources/agents/turn.py

@@ -58,7 +58,6 @@ def create(
         *,
         agent_id: str,
         messages: Iterable[turn_create_params.Message],
-        allow_turn_resume: bool | NotGiven = NOT_GIVEN,
         documents: Iterable[turn_create_params.Document] | NotGiven = NOT_GIVEN,
         stream: Literal[False] | NotGiven = NOT_GIVEN,
         tool_config: turn_create_params.ToolConfig | NotGiven = NOT_GIVEN,
@@ -92,7 +91,6 @@ def create(
         agent_id: str,
         messages: Iterable[turn_create_params.Message],
         stream: Literal[True],
-        allow_turn_resume: bool | NotGiven = NOT_GIVEN,
         documents: Iterable[turn_create_params.Document] | NotGiven = NOT_GIVEN,
         tool_config: turn_create_params.ToolConfig | NotGiven = NOT_GIVEN,
         toolgroups: List[turn_create_params.Toolgroup] | NotGiven = NOT_GIVEN,
@@ -125,7 +123,6 @@ def create(
         agent_id: str,
         messages: Iterable[turn_create_params.Message],
         stream: bool,
-        allow_turn_resume: bool | NotGiven = NOT_GIVEN,
         documents: Iterable[turn_create_params.Document] | NotGiven = NOT_GIVEN,
         tool_config: turn_create_params.ToolConfig | NotGiven = NOT_GIVEN,
         toolgroups: List[turn_create_params.Toolgroup] | NotGiven = NOT_GIVEN,
@@ -157,7 +154,6 @@ def create(
         *,
         agent_id: str,
         messages: Iterable[turn_create_params.Message],
-        allow_turn_resume: bool | NotGiven = NOT_GIVEN,
         documents: Iterable[turn_create_params.Document] | NotGiven = NOT_GIVEN,
         stream: Literal[False] | Literal[True] | NotGiven = NOT_GIVEN,
         tool_config: turn_create_params.ToolConfig | NotGiven = NOT_GIVEN,
@@ -178,7 +174,6 @@ def create(
             body=maybe_transform(
                 {
                     "messages": messages,
-                    "allow_turn_resume": allow_turn_resume,
                     "documents": documents,
                     "stream": stream,
                     "tool_config": tool_config,
@@ -412,7 +407,6 @@ async def create(
         *,
         agent_id: str,
         messages: Iterable[turn_create_params.Message],
-        allow_turn_resume: bool | NotGiven = NOT_GIVEN,
         documents: Iterable[turn_create_params.Document] | NotGiven = NOT_GIVEN,
         stream: Literal[False] | NotGiven = NOT_GIVEN,
         tool_config: turn_create_params.ToolConfig | NotGiven = NOT_GIVEN,
@@ -446,7 +440,6 @@ async def create(
         agent_id: str,
         messages: Iterable[turn_create_params.Message],
         stream: Literal[True],
-        allow_turn_resume: bool | NotGiven = NOT_GIVEN,
         documents: Iterable[turn_create_params.Document] | NotGiven = NOT_GIVEN,
         tool_config: turn_create_params.ToolConfig | NotGiven = NOT_GIVEN,
         toolgroups: List[turn_create_params.Toolgroup] | NotGiven = NOT_GIVEN,
@@ -479,7 +472,6 @@ async def create(
         agent_id: str,
         messages: Iterable[turn_create_params.Message],
         stream: bool,
-        allow_turn_resume: bool | NotGiven = NOT_GIVEN,
         documents: Iterable[turn_create_params.Document] | NotGiven = NOT_GIVEN,
         tool_config: turn_create_params.ToolConfig | NotGiven = NOT_GIVEN,
         toolgroups: List[turn_create_params.Toolgroup] | NotGiven = NOT_GIVEN,
@@ -511,7 +503,6 @@ async def create(
         *,
         agent_id: str,
         messages: Iterable[turn_create_params.Message],
-        allow_turn_resume: bool | NotGiven = NOT_GIVEN,
         documents: Iterable[turn_create_params.Document] | NotGiven = NOT_GIVEN,
         stream: Literal[False] | Literal[True] | NotGiven = NOT_GIVEN,
         tool_config: turn_create_params.ToolConfig | NotGiven = NOT_GIVEN,
@@ -532,7 +523,6 @@ async def create(
             body=await async_maybe_transform(
                 {
                     "messages": messages,
-                    "allow_turn_resume": allow_turn_resume,
                     "documents": documents,
                     "stream": stream,
                     "tool_config": tool_config,