scale-gp-beta 0.1.0a29__py3-none-any.whl → 0.1.0a30__py3-none-any.whl

scale_gp_beta/_base_client.py

@@ -532,7 +532,10 @@ class BaseClient(Generic[_HttpxClientT, _DefaultStreamT]):
         is_body_allowed = options.method.lower() != "get"
 
         if is_body_allowed:
-            kwargs["json"] = json_data if is_given(json_data) else None
+            if isinstance(json_data, bytes):
+                kwargs["content"] = json_data
+            else:
+                kwargs["json"] = json_data if is_given(json_data) else None
             kwargs["files"] = files
         else:
             headers.pop("Content-Type", None)

scale_gp_beta/_client.py

@@ -27,6 +27,7 @@ from .resources import (
     datasets,
     inference,
     questions,
+    responses,
     completions,
     evaluations,
     dataset_items,
@@ -61,6 +62,7 @@ ENVIRONMENTS: Dict[str, str] = {
 
 
 class SGPClient(SyncAPIClient):
+    responses: responses.ResponsesResource
     completions: completions.CompletionsResource
     chat: chat.ChatResource
     inference: inference.InferenceResource
@@ -165,6 +167,7 @@ class SGPClient(SyncAPIClient):
             _strict_response_validation=_strict_response_validation,
         )
 
+        self.responses = responses.ResponsesResource(self)
         self.completions = completions.CompletionsResource(self)
         self.chat = chat.ChatResource(self)
         self.inference = inference.InferenceResource(self)
@@ -290,6 +293,7 @@ class SGPClient(SyncAPIClient):
 
 
 class AsyncSGPClient(AsyncAPIClient):
+    responses: responses.AsyncResponsesResource
     completions: completions.AsyncCompletionsResource
     chat: chat.AsyncChatResource
     inference: inference.AsyncInferenceResource
@@ -394,6 +398,7 @@ class AsyncSGPClient(AsyncAPIClient):
             _strict_response_validation=_strict_response_validation,
         )
 
+        self.responses = responses.AsyncResponsesResource(self)
         self.completions = completions.AsyncCompletionsResource(self)
         self.chat = chat.AsyncChatResource(self)
         self.inference = inference.AsyncInferenceResource(self)
@@ -520,6 +525,7 @@ class AsyncSGPClient(AsyncAPIClient):
 
 class SGPClientWithRawResponse:
     def __init__(self, client: SGPClient) -> None:
+        self.responses = responses.ResponsesResourceWithRawResponse(client.responses)
         self.completions = completions.CompletionsResourceWithRawResponse(client.completions)
         self.chat = chat.ChatResourceWithRawResponse(client.chat)
         self.inference = inference.InferenceResourceWithRawResponse(client.inference)
@@ -535,6 +541,7 @@ class SGPClientWithRawResponse:
 
 class AsyncSGPClientWithRawResponse:
     def __init__(self, client: AsyncSGPClient) -> None:
+        self.responses = responses.AsyncResponsesResourceWithRawResponse(client.responses)
         self.completions = completions.AsyncCompletionsResourceWithRawResponse(client.completions)
         self.chat = chat.AsyncChatResourceWithRawResponse(client.chat)
         self.inference = inference.AsyncInferenceResourceWithRawResponse(client.inference)
@@ -550,6 +557,7 @@ class AsyncSGPClientWithRawResponse:
 
 class SGPClientWithStreamedResponse:
     def __init__(self, client: SGPClient) -> None:
+        self.responses = responses.ResponsesResourceWithStreamingResponse(client.responses)
         self.completions = completions.CompletionsResourceWithStreamingResponse(client.completions)
         self.chat = chat.ChatResourceWithStreamingResponse(client.chat)
         self.inference = inference.InferenceResourceWithStreamingResponse(client.inference)
@@ -565,6 +573,7 @@ class SGPClientWithStreamedResponse:
 
 class AsyncSGPClientWithStreamedResponse:
     def __init__(self, client: AsyncSGPClient) -> None:
+        self.responses = responses.AsyncResponsesResourceWithStreamingResponse(client.responses)
         self.completions = completions.AsyncCompletionsResourceWithStreamingResponse(client.completions)
         self.chat = chat.AsyncChatResourceWithStreamingResponse(client.chat)
         self.inference = inference.AsyncInferenceResourceWithStreamingResponse(client.inference)

scale_gp_beta/_files.py

@@ -69,12 +69,12 @@ def _transform_file(file: FileTypes) -> HttpxFileTypes:
         return file
 
     if is_tuple_t(file):
-        return (file[0], _read_file_content(file[1]), *file[2:])
+        return (file[0], read_file_content(file[1]), *file[2:])
 
     raise TypeError(f"Expected file types input to be a FileContent type or to be a tuple")
 
 
-def _read_file_content(file: FileContent) -> HttpxFileContent:
+def read_file_content(file: FileContent) -> HttpxFileContent:
     if isinstance(file, os.PathLike):
         return pathlib.Path(file).read_bytes()
     return file
@@ -111,12 +111,12 @@ async def _async_transform_file(file: FileTypes) -> HttpxFileTypes:
         return file
 
     if is_tuple_t(file):
-        return (file[0], await _async_read_file_content(file[1]), *file[2:])
+        return (file[0], await async_read_file_content(file[1]), *file[2:])
 
     raise TypeError(f"Expected file types input to be a FileContent type or to be a tuple")
 
 
-async def _async_read_file_content(file: FileContent) -> HttpxFileContent:
+async def async_read_file_content(file: FileContent) -> HttpxFileContent:
     if isinstance(file, os.PathLike):
         return await anyio.Path(file).read_bytes()
 

scale_gp_beta/_models.py

@@ -208,14 +208,18 @@ class BaseModel(pydantic.BaseModel):
             else:
                 fields_values[name] = field_get_default(field)
 
+        extra_field_type = _get_extra_fields_type(__cls)
+
         _extra = {}
         for key, value in values.items():
             if key not in model_fields:
+                parsed = construct_type(value=value, type_=extra_field_type) if extra_field_type is not None else value
+
                 if PYDANTIC_V2:
-                    _extra[key] = value
+                    _extra[key] = parsed
                 else:
                     _fields_set.add(key)
-                    fields_values[key] = value
+                    fields_values[key] = parsed
 
         object.__setattr__(m, "__dict__", fields_values)
 
@@ -370,6 +374,23 @@ def _construct_field(value: object, field: FieldInfo, key: str) -> object:
     return construct_type(value=value, type_=type_, metadata=getattr(field, "metadata", None))
 
 
+def _get_extra_fields_type(cls: type[pydantic.BaseModel]) -> type | None:
+    if not PYDANTIC_V2:
+        # TODO
+        return None
+
+    schema = cls.__pydantic_core_schema__
+    if schema["type"] == "model":
+        fields = schema["schema"]
+        if fields["type"] == "model-fields":
+            extras = fields.get("extras_schema")
+            if extras and "cls" in extras:
+                # mypy can't narrow the type
+                return extras["cls"]  # type: ignore[no-any-return]
+
+    return None
+
+
 def is_basemodel(type_: type) -> bool:
     """Returns whether or not the given type is either a `BaseModel` or a union of `BaseModel`"""
     if is_union(type_):
@@ -439,7 +460,7 @@ def construct_type(*, value: object, type_: object, metadata: Optional[List[Any]
         type_ = type_.__value__  # type: ignore[unreachable]
 
     # unwrap `Annotated[T, ...]` -> `T`
-    if metadata is not None:
+    if metadata is not None and len(metadata) > 0:
         meta: tuple[Any, ...] = tuple(metadata)
     elif is_annotated_type(type_):
         meta = get_args(type_)[1:]

scale_gp_beta/_version.py

@@ -1,4 +1,4 @@
 # File generated from our OpenAPI spec by Stainless. See CONTRIBUTING.md for details.
 
 __title__ = "scale_gp_beta"
-__version__ = "0.1.0-alpha.29"  # x-release-please-version
+__version__ = "0.1.0-alpha.30"  # x-release-please-version

scale_gp_beta/lib/CONTRIBUTING.MD

@@ -0,0 +1,53 @@
+# Custom Code Patch
+You can in theory add custom code patches to anywhere in the repo but at the risk of encountering many merge conflicts with Stainless into the future.
+Stainless will never use the `/lib` and `/examples` directories. When possible try to only modify these directories.
+If you have to add custom code elsewhere, please keep the footprint small and create a library for most of the logic.
+
+
+For information on custom code patching with Stainless see [here](https://www.stainless.com/docs/guides/add-custom-code).
+
+# Process for Adding Features
+Checkout the `next` branch and pull, then create a branch from `next`.
+
+> **_NOTE:_** Stainless uses next to "queue up" updates to the SDK.
+> 
+> Stainless will update next with their own logic updates to the SDK along with any changes in the OpenAPI spec and changes to the Stainless
+config on their SaaS platform.
+
+Make any code changes you need. Ensuring all the tests for the library are passing.
+
+There is strict linting in this repo. Use the following commands in order.
+
+```bash
+rye lint --fix
+```
+
+```bash
+rye run lint | grep /specific_file.py
+```
+
+`rye run lint` will not work if there are errors with `rye lint --fix`.
+I am unsure why but I get many errors which are ignorable in the rest of the repo when running `rye run lint` so I usually use it
+with grep to target the file I am developing.
+
+> **_NOTE:_** The strict linting requires all types to be strictly typed. This can be a pain but is worth considering before developing any new solution.
+> Try and avoid using ignore commands when possible, but sometimes it is unavoidable (see OpenAI tracing Plugin).
+
+When commiting, use the [Conventional Commits](https://www.conventionalcommits.org/en/v1.0.0/) spec. Usually this is something like:
+
+```bash
+git commit -m "feat: Added OpenAI tracing processor"
+```
+This allow Stainless to update the release doc with more useful info.
+
+**When creating a PR, you must manually change the destination from `main` to `next`.**
+
+Once merged, Stainless should pick up the changes and update or create a new PR for `main`.
+You will need to accept changes on this PR, Stainless should then auto merge. Note on occasion I have had to manually merge.
+There was no consequence to this.
+
+These PRs Stainless makes have 4 types of changes, all are merged to main via these automatic PR requests:
+1. Custom Code Changes
+2. Stainless SDK logic changes
+3. Changes to the Stainless Config in the SaaS platform
+4. Changes to the OpenAPI schema

scale_gp_beta/lib/tracing/trace_queue_manager.py

@@ -11,6 +11,8 @@ from .util import configure, is_disabled
 from .trace_exporter import TraceExporter
 
 if TYPE_CHECKING:
+    import httpx
+
     from .span import Span
     from .trace import Trace
 
@@ -44,6 +46,7 @@ class TraceQueueManager:
         worker_enabled: Optional[bool] = None,
     ):
         self._client = client
+        self.register_client(client) if client else None
         self._attempted_local_client_creation = False
         self._trigger_queue_size = trigger_queue_size
         self._trigger_cadence = trigger_cadence
@@ -68,6 +71,17 @@ class TraceQueueManager:
         log.info("Registering client")
         self._client = client
 
+        original_prepare_request = self._client._prepare_request
+
+        def custom_prepare_request(request: "httpx.Request") -> None:
+            original_prepare_request(request)
+
+            # TODO: Hook logic here, we should check to see if we are in the scope of a span, if so we should inject
+            # appropriate headers into the request
+            # current_span = Scope.get_current_span()
+
+        self._client._prepare_request = custom_prepare_request  # type: ignore
+
     def shutdown(self, timeout: Optional[float] = None) -> None:
         if not self._worker_enabled:
             log.debug("No worker to shutdown")

scale_gp_beta/resources/__init__.py

@@ -56,6 +56,14 @@ from .questions import (
     QuestionsResourceWithStreamingResponse,
     AsyncQuestionsResourceWithStreamingResponse,
 )
+from .responses import (
+    ResponsesResource,
+    AsyncResponsesResource,
+    ResponsesResourceWithRawResponse,
+    AsyncResponsesResourceWithRawResponse,
+    ResponsesResourceWithStreamingResponse,
+    AsyncResponsesResourceWithStreamingResponse,
+)
 from .completions import (
     CompletionsResource,
     AsyncCompletionsResource,
@@ -90,6 +98,12 @@ from .evaluation_items import (
 )
 
 __all__ = [
+    "ResponsesResource",
+    "AsyncResponsesResource",
+    "ResponsesResourceWithRawResponse",
+    "AsyncResponsesResourceWithRawResponse",
+    "ResponsesResourceWithStreamingResponse",
+    "AsyncResponsesResourceWithStreamingResponse",
     "CompletionsResource",
     "AsyncCompletionsResource",
     "CompletionsResourceWithRawResponse",

scale_gp_beta/resources/chat/completions.py

@@ -522,6 +522,7 @@ class CompletionsResource(SyncAPIResource):
     def models(
         self,
         *,
+        check_availability: bool | NotGiven = NOT_GIVEN,
         ending_before: str | NotGiven = NOT_GIVEN,
         limit: int | NotGiven = NOT_GIVEN,
         model_vendor: Literal[
@@ -569,6 +570,7 @@ class CompletionsResource(SyncAPIResource):
                 timeout=timeout,
                 query=maybe_transform(
                     {
+                        "check_availability": check_availability,
                         "ending_before": ending_before,
                         "limit": limit,
                         "model_vendor": model_vendor,
@@ -1077,6 +1079,7 @@ class AsyncCompletionsResource(AsyncAPIResource):
     async def models(
         self,
         *,
+        check_availability: bool | NotGiven = NOT_GIVEN,
         ending_before: str | NotGiven = NOT_GIVEN,
         limit: int | NotGiven = NOT_GIVEN,
         model_vendor: Literal[
@@ -1124,6 +1127,7 @@ class AsyncCompletionsResource(AsyncAPIResource):
                 timeout=timeout,
                 query=await async_maybe_transform(
                     {
+                        "check_availability": check_availability,
                         "ending_before": ending_before,
                         "limit": limit,
                         "model_vendor": model_vendor,

scale_gp_beta/resources/responses.py

@@ -0,0 +1,314 @@
+# File generated from our OpenAPI spec by Stainless. See CONTRIBUTING.md for details.
+
+from __future__ import annotations
+
+from typing import Any, Dict, List, Union, Iterable, cast
+from typing_extensions import Literal
+
+import httpx
+
+from ..types import response_create_params
+from .._types import NOT_GIVEN, Body, Query, Headers, NotGiven
+from .._utils import maybe_transform, async_maybe_transform
+from .._compat import cached_property
+from .._resource import SyncAPIResource, AsyncAPIResource
+from .._response import (
+    to_raw_response_wrapper,
+    to_streamed_response_wrapper,
+    async_to_raw_response_wrapper,
+    async_to_streamed_response_wrapper,
+)
+from .._base_client import make_request_options
+from ..types.response_create_response import ResponseCreateResponse
+
+__all__ = ["ResponsesResource", "AsyncResponsesResource"]
+
+
+class ResponsesResource(SyncAPIResource):
+    @cached_property
+    def with_raw_response(self) -> ResponsesResourceWithRawResponse:
+        """
+        This property can be used as a prefix for any HTTP method call to return
+        the raw response object instead of the parsed content.
+
+        For more information, see https://www.github.com/scaleapi/sgp-python-beta#accessing-raw-response-data-eg-headers
+        """
+        return ResponsesResourceWithRawResponse(self)
+
+    @cached_property
+    def with_streaming_response(self) -> ResponsesResourceWithStreamingResponse:
+        """
+        An alternative to `.with_raw_response` that doesn't eagerly read the response body.
+
+        For more information, see https://www.github.com/scaleapi/sgp-python-beta#with_streaming_response
+        """
+        return ResponsesResourceWithStreamingResponse(self)
+
+    def create(
+        self,
+        *,
+        input: Union[str, Iterable[response_create_params.InputUnionMember1]],
+        model: str,
+        include: List[str] | NotGiven = NOT_GIVEN,
+        instructions: str | NotGiven = NOT_GIVEN,
+        max_output_tokens: int | NotGiven = NOT_GIVEN,
+        metadata: Dict[str, object] | NotGiven = NOT_GIVEN,
+        parallel_tool_calls: bool | NotGiven = NOT_GIVEN,
+        previous_response_id: str | NotGiven = NOT_GIVEN,
+        reasoning: Dict[str, object] | NotGiven = NOT_GIVEN,
+        store: bool | NotGiven = NOT_GIVEN,
+        stream: bool | NotGiven = NOT_GIVEN,
+        temperature: float | NotGiven = NOT_GIVEN,
+        text: Dict[str, object] | NotGiven = NOT_GIVEN,
+        tool_choice: Union[str, Dict[str, object]] | NotGiven = NOT_GIVEN,
+        tools: Iterable[Dict[str, object]] | NotGiven = NOT_GIVEN,
+        top_p: float | NotGiven = NOT_GIVEN,
+        truncation: Literal["auto", "disabled"] | NotGiven = NOT_GIVEN,
+        # Use the following arguments if you need to pass additional parameters to the API that aren't available via kwargs.
+        # The extra values given here take precedence over values defined on the client or passed to this method.
+        extra_headers: Headers | None = None,
+        extra_query: Query | None = None,
+        extra_body: Body | None = None,
+        timeout: float | httpx.Timeout | None | NotGiven = NOT_GIVEN,
+    ) -> ResponseCreateResponse:
+        """
+        Responses
+
+        Args:
+          model: model specified as `model_vendor/model`, for example `openai/gpt-4o`
+
+          include: Which fields to include in the response
+
+          instructions: Instructions for the response generation
+
+          max_output_tokens: Maximum number of output tokens
+
+          metadata: Metadata for the response
+
+          parallel_tool_calls: Whether to enable parallel tool calls
+
+          previous_response_id: ID of the previous response for chaining
+
+          reasoning: Reasoning configuration for the response
+
+          store: Whether to store the response
+
+          stream: Whether to stream the response
+
+          temperature: Sampling temperature for randomness control
+
+          text: Text configuration parameters
+
+          tool_choice: Tool choice configuration
+
+          tools: Tools available for the response
+
+          top_p: Top-p sampling parameter
+
+          truncation: Truncation configuration
+
+          extra_headers: Send extra headers
+
+          extra_query: Add additional query parameters to the request
+
+          extra_body: Add additional JSON properties to the request
+
+          timeout: Override the client-level default timeout for this request, in seconds
+        """
+        return cast(
+            ResponseCreateResponse,
+            self._post(
+                "/v5/responses",
+                body=maybe_transform(
+                    {
+                        "input": input,
+                        "model": model,
+                        "include": include,
+                        "instructions": instructions,
+                        "max_output_tokens": max_output_tokens,
+                        "metadata": metadata,
+                        "parallel_tool_calls": parallel_tool_calls,
+                        "previous_response_id": previous_response_id,
+                        "reasoning": reasoning,
+                        "store": store,
+                        "stream": stream,
+                        "temperature": temperature,
+                        "text": text,
+                        "tool_choice": tool_choice,
+                        "tools": tools,
+                        "top_p": top_p,
+                        "truncation": truncation,
+                    },
+                    response_create_params.ResponseCreateParams,
+                ),
+                options=make_request_options(
+                    extra_headers=extra_headers, extra_query=extra_query, extra_body=extra_body, timeout=timeout
+                ),
+                cast_to=cast(
+                    Any, ResponseCreateResponse
+                ),  # Union types cannot be passed in as arguments in the type system
+            ),
+        )
+
+
+class AsyncResponsesResource(AsyncAPIResource):
+    @cached_property
+    def with_raw_response(self) -> AsyncResponsesResourceWithRawResponse:
+        """
+        This property can be used as a prefix for any HTTP method call to return
+        the raw response object instead of the parsed content.
+
+        For more information, see https://www.github.com/scaleapi/sgp-python-beta#accessing-raw-response-data-eg-headers
+        """
+        return AsyncResponsesResourceWithRawResponse(self)
+
+    @cached_property
+    def with_streaming_response(self) -> AsyncResponsesResourceWithStreamingResponse:
+        """
+        An alternative to `.with_raw_response` that doesn't eagerly read the response body.
+
+        For more information, see https://www.github.com/scaleapi/sgp-python-beta#with_streaming_response
+        """
+        return AsyncResponsesResourceWithStreamingResponse(self)
+
+    async def create(
+        self,
+        *,
+        input: Union[str, Iterable[response_create_params.InputUnionMember1]],
+        model: str,
+        include: List[str] | NotGiven = NOT_GIVEN,
+        instructions: str | NotGiven = NOT_GIVEN,
+        max_output_tokens: int | NotGiven = NOT_GIVEN,
+        metadata: Dict[str, object] | NotGiven = NOT_GIVEN,
+        parallel_tool_calls: bool | NotGiven = NOT_GIVEN,
+        previous_response_id: str | NotGiven = NOT_GIVEN,
+        reasoning: Dict[str, object] | NotGiven = NOT_GIVEN,
+        store: bool | NotGiven = NOT_GIVEN,
+        stream: bool | NotGiven = NOT_GIVEN,
+        temperature: float | NotGiven = NOT_GIVEN,
+        text: Dict[str, object] | NotGiven = NOT_GIVEN,
+        tool_choice: Union[str, Dict[str, object]] | NotGiven = NOT_GIVEN,
+        tools: Iterable[Dict[str, object]] | NotGiven = NOT_GIVEN,
+        top_p: float | NotGiven = NOT_GIVEN,
+        truncation: Literal["auto", "disabled"] | NotGiven = NOT_GIVEN,
+        # Use the following arguments if you need to pass additional parameters to the API that aren't available via kwargs.
+        # The extra values given here take precedence over values defined on the client or passed to this method.
+        extra_headers: Headers | None = None,
+        extra_query: Query | None = None,
+        extra_body: Body | None = None,
+        timeout: float | httpx.Timeout | None | NotGiven = NOT_GIVEN,
+    ) -> ResponseCreateResponse:
+        """
+        Responses
+
+        Args:
+          model: model specified as `model_vendor/model`, for example `openai/gpt-4o`
+
+          include: Which fields to include in the response
+
+          instructions: Instructions for the response generation
+
+          max_output_tokens: Maximum number of output tokens
+
+          metadata: Metadata for the response
+
+          parallel_tool_calls: Whether to enable parallel tool calls
+
+          previous_response_id: ID of the previous response for chaining
+
+          reasoning: Reasoning configuration for the response
+
+          store: Whether to store the response
+
+          stream: Whether to stream the response
+
+          temperature: Sampling temperature for randomness control
+
+          text: Text configuration parameters
+
+          tool_choice: Tool choice configuration
+
+          tools: Tools available for the response
+
+          top_p: Top-p sampling parameter
+
+          truncation: Truncation configuration
+
+          extra_headers: Send extra headers
+
+          extra_query: Add additional query parameters to the request
+
+          extra_body: Add additional JSON properties to the request
+
+          timeout: Override the client-level default timeout for this request, in seconds
+        """
+        return cast(
+            ResponseCreateResponse,
+            await self._post(
+                "/v5/responses",
+                body=await async_maybe_transform(
+                    {
+                        "input": input,
+                        "model": model,
+                        "include": include,
+                        "instructions": instructions,
+                        "max_output_tokens": max_output_tokens,
+                        "metadata": metadata,
+                        "parallel_tool_calls": parallel_tool_calls,
+                        "previous_response_id": previous_response_id,
+                        "reasoning": reasoning,
+                        "store": store,
+                        "stream": stream,
+                        "temperature": temperature,
+                        "text": text,
+                        "tool_choice": tool_choice,
+                        "tools": tools,
+                        "top_p": top_p,
+                        "truncation": truncation,
+                    },
+                    response_create_params.ResponseCreateParams,
+                ),
+                options=make_request_options(
+                    extra_headers=extra_headers, extra_query=extra_query, extra_body=extra_body, timeout=timeout
+                ),
+                cast_to=cast(
+                    Any, ResponseCreateResponse
+                ),  # Union types cannot be passed in as arguments in the type system
+            ),
+        )
+
+
+class ResponsesResourceWithRawResponse:
+    def __init__(self, responses: ResponsesResource) -> None:
+        self._responses = responses
+
+        self.create = to_raw_response_wrapper(
+            responses.create,
+        )
+
+
+class AsyncResponsesResourceWithRawResponse:
+    def __init__(self, responses: AsyncResponsesResource) -> None:
+        self._responses = responses
+
+        self.create = async_to_raw_response_wrapper(
+            responses.create,
+        )
+
+
+class ResponsesResourceWithStreamingResponse:
+    def __init__(self, responses: ResponsesResource) -> None:
+        self._responses = responses
+
+        self.create = to_streamed_response_wrapper(
+            responses.create,
+        )
+
+
+class AsyncResponsesResourceWithStreamingResponse:
+    def __init__(self, responses: AsyncResponsesResource) -> None:
+        self._responses = responses
+
+        self.create = async_to_streamed_response_wrapper(
+            responses.create,
+        )