From cfb7e308393f2e912e959dd10d68096dd5b3ab9c Mon Sep 17 00:00:00 2001
From: Stainless Bot <107565488+stainless-bot@users.noreply.github.com>
Date: Mon, 18 Dec 2023 09:06:05 -0500
Subject: [PATCH 01/10] chore(internal): fix binary response tests (#983)

---
 tests/api_resources/audio/test_speech.py |  6 ++----
 tests/api_resources/test_files.py        | 12 ++++--------
 2 files changed, 6 insertions(+), 12 deletions(-)

diff --git a/tests/api_resources/audio/test_speech.py b/tests/api_resources/audio/test_speech.py
index 50b00b73b4..23f5303153 100644
--- a/tests/api_resources/audio/test_speech.py
+++ b/tests/api_resources/audio/test_speech.py
@@ -39,8 +39,7 @@ def test_method_create(self, client: OpenAI, respx_mock: MockRouter) -> None:
     @pytest.mark.respx(base_url=base_url)
     def test_method_create_with_all_params(self, client: OpenAI, respx_mock: MockRouter) -> None:
         respx_mock.post("/audio/speech").mock(return_value=httpx.Response(200, json={"foo": "bar"}))
-        speech = respx_mock.post("/audio/speech").mock(return_value=httpx.Response(200, json={"foo": "bar"}))
-        client.audio.speech.create(
+        speech = client.audio.speech.create(
             input="string",
             model="string",
             voice="alloy",
@@ -89,8 +88,7 @@ async def test_method_create(self, client: AsyncOpenAI, respx_mock: MockRouter)
     @pytest.mark.respx(base_url=base_url)
     async def test_method_create_with_all_params(self, client: AsyncOpenAI, respx_mock: MockRouter) -> None:
         respx_mock.post("/audio/speech").mock(return_value=httpx.Response(200, json={"foo": "bar"}))
-        speech = respx_mock.post("/audio/speech").mock(return_value=httpx.Response(200, json={"foo": "bar"}))
-        await client.audio.speech.create(
+        speech = await client.audio.speech.create(
             input="string",
             model="string",
             voice="alloy",
diff --git a/tests/api_resources/test_files.py b/tests/api_resources/test_files.py
index e4cf493319..13ffca9773 100644
--- a/tests/api_resources/test_files.py
+++ b/tests/api_resources/test_files.py
@@ -95,22 +95,20 @@ def test_raw_response_delete(self, client: OpenAI) -> None:
         file = response.parse()
         assert_matches_type(FileDeleted, file, path=["response"])
 
-    @pytest.mark.skip(reason="mocked response isn't working yet")
     @parametrize
     @pytest.mark.respx(base_url=base_url)
     def test_method_content(self, client: OpenAI, respx_mock: MockRouter) -> None:
-        respx_mock.get("/files/{file_id}/content").mock(return_value=httpx.Response(200, json={"foo": "bar"}))
+        respx_mock.get("/files/string/content").mock(return_value=httpx.Response(200, json={"foo": "bar"}))
         file = client.files.content(
             "string",
         )
         assert isinstance(file, BinaryResponseContent)
         assert file.json() == {"foo": "bar"}
 
-    @pytest.mark.skip(reason="mocked response isn't working yet")
     @parametrize
     @pytest.mark.respx(base_url=base_url)
     def test_raw_response_content(self, client: OpenAI, respx_mock: MockRouter) -> None:
-        respx_mock.get("/files/{file_id}/content").mock(return_value=httpx.Response(200, json={"foo": "bar"}))
+        respx_mock.get("/files/string/content").mock(return_value=httpx.Response(200, json={"foo": "bar"}))
         response = client.files.with_raw_response.content(
             "string",
         )
@@ -212,22 +210,20 @@ async def test_raw_response_delete(self, client: AsyncOpenAI) -> None:
         file = response.parse()
         assert_matches_type(FileDeleted, file, path=["response"])
 
-    @pytest.mark.skip(reason="mocked response isn't working yet")
     @parametrize
     @pytest.mark.respx(base_url=base_url)
     async def test_method_content(self, client: AsyncOpenAI, respx_mock: MockRouter) -> None:
-        respx_mock.get("/files/{file_id}/content").mock(return_value=httpx.Response(200, json={"foo": "bar"}))
+        respx_mock.get("/files/string/content").mock(return_value=httpx.Response(200, json={"foo": "bar"}))
         file = await client.files.content(
             "string",
         )
         assert isinstance(file, BinaryResponseContent)
         assert file.json() == {"foo": "bar"}
 
-    @pytest.mark.skip(reason="mocked response isn't working yet")
     @parametrize
     @pytest.mark.respx(base_url=base_url)
     async def test_raw_response_content(self, client: AsyncOpenAI, respx_mock: MockRouter) -> None:
-        respx_mock.get("/files/{file_id}/content").mock(return_value=httpx.Response(200, json={"foo": "bar"}))
+        respx_mock.get("/files/string/content").mock(return_value=httpx.Response(200, json={"foo": "bar"}))
         response = await client.files.with_raw_response.content(
             "string",
         )

From d1e9e8f24df366bb7b796c55a98247c025d229f5 Mon Sep 17 00:00:00 2001
From: Logan Kilpatrick <logan@openai.com>
Date: Tue, 19 Dec 2023 00:53:24 +0900
Subject: [PATCH 02/10] chore(cli): fix typo in completions (#985)

---
 src/openai/cli/_api/completions.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/src/openai/cli/_api/completions.py b/src/openai/cli/_api/completions.py
index ce1036b224..cbdb35bf3a 100644
--- a/src/openai/cli/_api/completions.py
+++ b/src/openai/cli/_api/completions.py
@@ -57,7 +57,7 @@ def register(subparser: _SubParsersAction[ArgumentParser]) -> None:
     )
     sub.add_argument(
         "--logprobs",
-        help="Include the log probabilites on the `logprobs` most likely tokens, as well the chosen tokens. So for example, if `logprobs` is 10, the API will return a list of the 10 most likely tokens. If `logprobs` is 0, only the chosen tokens will have logprobs returned.",
+        help="Include the log probabilities on the `logprobs` most likely tokens, as well the chosen tokens. So for example, if `logprobs` is 10, the API will return a list of the 10 most likely tokens. If `logprobs` is 0, only the chosen tokens will have logprobs returned.",
         type=int,
     )
     sub.add_argument(

From 626bc34d82a7057bac99f8b556f9e5f60c261ee7 Mon Sep 17 00:00:00 2001
From: Stainless Bot <107565488+stainless-bot@users.noreply.github.com>
Date: Mon, 18 Dec 2023 11:31:14 -0500
Subject: [PATCH 03/10] chore(cli): fix typo in completions (#986)

---
 examples/async_demo.py | 2 +-
 examples/streaming.py  | 4 ++--
 2 files changed, 3 insertions(+), 3 deletions(-)

diff --git a/examples/async_demo.py b/examples/async_demo.py
index 793b4e43fb..92c267c38f 100755
--- a/examples/async_demo.py
+++ b/examples/async_demo.py
@@ -10,7 +10,7 @@
 
 async def main() -> None:
     stream = await client.completions.create(
-        model="gpt-3.5-turbo-instruct",
+        model="text-davinci-003",
         prompt="Say this is a test",
         stream=True,
     )
diff --git a/examples/streaming.py b/examples/streaming.py
index 368fa5f911..168877dfc5 100755
--- a/examples/streaming.py
+++ b/examples/streaming.py
@@ -13,7 +13,7 @@
 def sync_main() -> None:
     client = OpenAI()
     response = client.completions.create(
-        model="gpt-3.5-turbo-instruct",
+        model="text-davinci-002",
         prompt="1,2,3,",
         max_tokens=5,
         temperature=0,
@@ -33,7 +33,7 @@ def sync_main() -> None:
 async def async_main() -> None:
     client = AsyncOpenAI()
     response = await client.completions.create(
-        model="gpt-3.5-turbo-instruct",
+        model="text-davinci-002",
         prompt="1,2,3,",
         max_tokens=5,
         temperature=0,

From cedd574e5611f3e71e92b523a72ba87bcfe546f1 Mon Sep 17 00:00:00 2001
From: Stainless Bot <107565488+stainless-bot@users.noreply.github.com>
Date: Mon, 18 Dec 2023 22:33:10 -0500
Subject: [PATCH 04/10] docs: upgrade models in examples to latest version
 (#989)

---
 examples/async_demo.py | 2 +-
 examples/streaming.py  | 4 ++--
 2 files changed, 3 insertions(+), 3 deletions(-)

diff --git a/examples/async_demo.py b/examples/async_demo.py
index 92c267c38f..793b4e43fb 100755
--- a/examples/async_demo.py
+++ b/examples/async_demo.py
@@ -10,7 +10,7 @@
 
 async def main() -> None:
     stream = await client.completions.create(
-        model="text-davinci-003",
+        model="gpt-3.5-turbo-instruct",
         prompt="Say this is a test",
         stream=True,
     )
diff --git a/examples/streaming.py b/examples/streaming.py
index 168877dfc5..368fa5f911 100755
--- a/examples/streaming.py
+++ b/examples/streaming.py
@@ -13,7 +13,7 @@
 def sync_main() -> None:
     client = OpenAI()
     response = client.completions.create(
-        model="text-davinci-002",
+        model="gpt-3.5-turbo-instruct",
         prompt="1,2,3,",
         max_tokens=5,
         temperature=0,
@@ -33,7 +33,7 @@ def sync_main() -> None:
 async def async_main() -> None:
     client = AsyncOpenAI()
     response = await client.completions.create(
-        model="text-davinci-002",
+        model="gpt-3.5-turbo-instruct",
         prompt="1,2,3,",
         max_tokens=5,
         temperature=0,

From 6c3427dac8c414658516aeb4caf5d5fd8b11097b Mon Sep 17 00:00:00 2001
From: Stainless Bot <107565488+stainless-bot@users.noreply.github.com>
Date: Tue, 19 Dec 2023 07:30:56 -0500
Subject: [PATCH 05/10] chore(streaming): update constructor to use direct
 client names (#991)

---
 src/openai/_streaming.py | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/src/openai/_streaming.py b/src/openai/_streaming.py
index e48324fc78..e323c59ac0 100644
--- a/src/openai/_streaming.py
+++ b/src/openai/_streaming.py
@@ -12,7 +12,7 @@
 from ._exceptions import APIError
 
 if TYPE_CHECKING:
-    from ._base_client import SyncAPIClient, AsyncAPIClient
+    from ._client import OpenAI, AsyncOpenAI
 
 
 class Stream(Generic[ResponseT]):
@@ -25,7 +25,7 @@ def __init__(
         *,
         cast_to: type[ResponseT],
         response: httpx.Response,
-        client: SyncAPIClient,
+        client: OpenAI,
     ) -> None:
         self.response = response
         self._cast_to = cast_to
@@ -79,7 +79,7 @@ def __init__(
         *,
         cast_to: type[ResponseT],
         response: httpx.Response,
-        client: AsyncAPIClient,
+        client: AsyncOpenAI,
     ) -> None:
         self.response = response
         self._cast_to = cast_to

From 5ba576ae38d2c4c4d32a21933e0d68e0bc2f0d49 Mon Sep 17 00:00:00 2001
From: Stainless Bot <107565488+stainless-bot@users.noreply.github.com>
Date: Tue, 19 Dec 2023 11:23:44 -0500
Subject: [PATCH 06/10] chore(internal): minor utils restructuring (#992)

---
 src/openai/_response.py         | 17 +++----
 src/openai/_streaming.py        | 71 ++++++++++++++++++++++-------
 src/openai/_types.py            | 14 ++++++
 src/openai/_utils/__init__.py   | 15 ++++---
 src/openai/_utils/_streams.py   | 12 +++++
 src/openai/_utils/_transform.py |  5 +--
 src/openai/_utils/_typing.py    | 80 +++++++++++++++++++++++++++++++++
 src/openai/_utils/_utils.py     | 35 +--------------
 8 files changed, 183 insertions(+), 66 deletions(-)
 create mode 100644 src/openai/_utils/_streams.py
 create mode 100644 src/openai/_utils/_typing.py

diff --git a/src/openai/_response.py b/src/openai/_response.py
index 933c37525e..6b7c86e544 100644
--- a/src/openai/_response.py
+++ b/src/openai/_response.py
@@ -5,12 +5,12 @@
 import datetime
 import functools
 from typing import TYPE_CHECKING, Any, Union, Generic, TypeVar, Callable, cast
-from typing_extensions import Awaitable, ParamSpec, get_args, override, get_origin
+from typing_extensions import Awaitable, ParamSpec, override, get_origin
 
 import httpx
 
 from ._types import NoneType, UnknownResponse, BinaryResponseContent
-from ._utils import is_given
+from ._utils import is_given, extract_type_var_from_base
 from ._models import BaseModel, is_basemodel
 from ._constants import RAW_RESPONSE_HEADER
 from ._exceptions import APIResponseValidationError
@@ -221,12 +221,13 @@ def __init__(self) -> None:
 
 
 def _extract_stream_chunk_type(stream_cls: type) -> type:
-    args = get_args(stream_cls)
-    if not args:
-        raise TypeError(
-            f"Expected stream_cls to have been given a generic type argument, e.g. Stream[Foo] but received {stream_cls}",
-        )
-    return cast(type, args[0])
+    from ._base_client import Stream, AsyncStream
+
+    return extract_type_var_from_base(
+        stream_cls,
+        index=0,
+        generic_bases=cast("tuple[type, ...]", (Stream, AsyncStream)),
+    )
 
 
 def to_raw_response_wrapper(func: Callable[P, R]) -> Callable[P, APIResponse[R]]:
diff --git a/src/openai/_streaming.py b/src/openai/_streaming.py
index e323c59ac0..f1896a242a 100644
--- a/src/openai/_streaming.py
+++ b/src/openai/_streaming.py
@@ -2,12 +2,12 @@
 from __future__ import annotations
 
 import json
-from typing import TYPE_CHECKING, Any, Generic, Iterator, AsyncIterator
-from typing_extensions import override
+from types import TracebackType
+from typing import TYPE_CHECKING, Any, Generic, TypeVar, Iterator, AsyncIterator, cast
+from typing_extensions import Self, override
 
 import httpx
 
-from ._types import ResponseT
 from ._utils import is_mapping
 from ._exceptions import APIError
 
@@ -15,7 +15,10 @@
     from ._client import OpenAI, AsyncOpenAI
 
 
-class Stream(Generic[ResponseT]):
+_T = TypeVar("_T")
+
+
+class Stream(Generic[_T]):
     """Provides the core interface to iterate over a synchronous stream response."""
 
     response: httpx.Response
@@ -23,7 +26,7 @@ class Stream(Generic[ResponseT]):
     def __init__(
         self,
         *,
-        cast_to: type[ResponseT],
+        cast_to: type[_T],
         response: httpx.Response,
         client: OpenAI,
     ) -> None:
@@ -33,18 +36,18 @@ def __init__(
         self._decoder = SSEDecoder()
         self._iterator = self.__stream__()
 
-    def __next__(self) -> ResponseT:
+    def __next__(self) -> _T:
         return self._iterator.__next__()
 
-    def __iter__(self) -> Iterator[ResponseT]:
+    def __iter__(self) -> Iterator[_T]:
         for item in self._iterator:
             yield item
 
     def _iter_events(self) -> Iterator[ServerSentEvent]:
         yield from self._decoder.iter(self.response.iter_lines())
 
-    def __stream__(self) -> Iterator[ResponseT]:
-        cast_to = self._cast_to
+    def __stream__(self) -> Iterator[_T]:
+        cast_to = cast(Any, self._cast_to)
         response = self.response
         process_data = self._client._process_response_data
         iterator = self._iter_events()
@@ -68,8 +71,27 @@ def __stream__(self) -> Iterator[ResponseT]:
         for _sse in iterator:
             ...
 
+    def __enter__(self) -> Self:
+        return self
+
+    def __exit__(
+        self,
+        exc_type: type[BaseException] | None,
+        exc: BaseException | None,
+        exc_tb: TracebackType | None,
+    ) -> None:
+        self.close()
+
+    def close(self) -> None:
+        """
+        Close the response and release the connection.
+
+        Automatically called if the response body is read to completion.
+        """
+        self.response.close()
 
-class AsyncStream(Generic[ResponseT]):
+
+class AsyncStream(Generic[_T]):
     """Provides the core interface to iterate over an asynchronous stream response."""
 
     response: httpx.Response
@@ -77,7 +99,7 @@ class AsyncStream(Generic[ResponseT]):
     def __init__(
         self,
         *,
-        cast_to: type[ResponseT],
+        cast_to: type[_T],
         response: httpx.Response,
         client: AsyncOpenAI,
     ) -> None:
@@ -87,10 +109,10 @@ def __init__(
         self._decoder = SSEDecoder()
         self._iterator = self.__stream__()
 
-    async def __anext__(self) -> ResponseT:
+    async def __anext__(self) -> _T:
         return await self._iterator.__anext__()
 
-    async def __aiter__(self) -> AsyncIterator[ResponseT]:
+    async def __aiter__(self) -> AsyncIterator[_T]:
         async for item in self._iterator:
             yield item
 
@@ -98,8 +120,8 @@ async def _iter_events(self) -> AsyncIterator[ServerSentEvent]:
         async for sse in self._decoder.aiter(self.response.aiter_lines()):
             yield sse
 
-    async def __stream__(self) -> AsyncIterator[ResponseT]:
-        cast_to = self._cast_to
+    async def __stream__(self) -> AsyncIterator[_T]:
+        cast_to = cast(Any, self._cast_to)
         response = self.response
         process_data = self._client._process_response_data
         iterator = self._iter_events()
@@ -123,6 +145,25 @@ async def __stream__(self) -> AsyncIterator[ResponseT]:
         async for _sse in iterator:
             ...
 
+    async def __aenter__(self) -> Self:
+        return self
+
+    async def __aexit__(
+        self,
+        exc_type: type[BaseException] | None,
+        exc: BaseException | None,
+        exc_tb: TracebackType | None,
+    ) -> None:
+        await self.close()
+
+    async def close(self) -> None:
+        """
+        Close the response and release the connection.
+
+        Automatically called if the response body is read to completion.
+        """
+        await self.response.aclose()
+
 
 class ServerSentEvent:
     def __init__(
diff --git a/src/openai/_types.py b/src/openai/_types.py
index 8d543171eb..a20a4b4c1b 100644
--- a/src/openai/_types.py
+++ b/src/openai/_types.py
@@ -353,3 +353,17 @@ def get(self, __key: str) -> str | None:
 IncEx: TypeAlias = "set[int] | set[str] | dict[int, Any] | dict[str, Any] | None"
 
 PostParser = Callable[[Any], Any]
+
+
+@runtime_checkable
+class InheritsGeneric(Protocol):
+    """Represents a type that has inherited from `Generic`
+    The `__orig_bases__` property can be used to determine the resolved
+    type variable for a given base class.
+    """
+
+    __orig_bases__: tuple[_GenericAlias]
+
+
+class _GenericAlias(Protocol):
+    __origin__: type[object]
diff --git a/src/openai/_utils/__init__.py b/src/openai/_utils/__init__.py
index 400ca9b828..a43201d3c7 100644
--- a/src/openai/_utils/__init__.py
+++ b/src/openai/_utils/__init__.py
@@ -9,13 +9,11 @@
 from ._utils import parse_date as parse_date
 from ._utils import is_sequence as is_sequence
 from ._utils import coerce_float as coerce_float
-from ._utils import is_list_type as is_list_type
 from ._utils import is_mapping_t as is_mapping_t
 from ._utils import removeprefix as removeprefix
 from ._utils import removesuffix as removesuffix
 from ._utils import extract_files as extract_files
 from ._utils import is_sequence_t as is_sequence_t
-from ._utils import is_union_type as is_union_type
 from ._utils import required_args as required_args
 from ._utils import coerce_boolean as coerce_boolean
 from ._utils import coerce_integer as coerce_integer
@@ -23,15 +21,20 @@
 from ._utils import parse_datetime as parse_datetime
 from ._utils import strip_not_given as strip_not_given
 from ._utils import deepcopy_minimal as deepcopy_minimal
-from ._utils import extract_type_arg as extract_type_arg
-from ._utils import is_required_type as is_required_type
 from ._utils import get_async_library as get_async_library
-from ._utils import is_annotated_type as is_annotated_type
 from ._utils import maybe_coerce_float as maybe_coerce_float
 from ._utils import get_required_header as get_required_header
 from ._utils import maybe_coerce_boolean as maybe_coerce_boolean
 from ._utils import maybe_coerce_integer as maybe_coerce_integer
-from ._utils import strip_annotated_type as strip_annotated_type
+from ._typing import is_list_type as is_list_type
+from ._typing import is_union_type as is_union_type
+from ._typing import extract_type_arg as extract_type_arg
+from ._typing import is_required_type as is_required_type
+from ._typing import is_annotated_type as is_annotated_type
+from ._typing import strip_annotated_type as strip_annotated_type
+from ._typing import extract_type_var_from_base as extract_type_var_from_base
+from ._streams import consume_sync_iterator as consume_sync_iterator
+from ._streams import consume_async_iterator as consume_async_iterator
 from ._transform import PropertyInfo as PropertyInfo
 from ._transform import transform as transform
 from ._transform import maybe_transform as maybe_transform
diff --git a/src/openai/_utils/_streams.py b/src/openai/_utils/_streams.py
new file mode 100644
index 0000000000..f4a0208f01
--- /dev/null
+++ b/src/openai/_utils/_streams.py
@@ -0,0 +1,12 @@
+from typing import Any
+from typing_extensions import Iterator, AsyncIterator
+
+
+def consume_sync_iterator(iterator: Iterator[Any]) -> None:
+    for _ in iterator:
+        ...
+
+
+async def consume_async_iterator(iterator: AsyncIterator[Any]) -> None:
+    async for _ in iterator:
+        ...
diff --git a/src/openai/_utils/_transform.py b/src/openai/_utils/_transform.py
index 769f7362b9..9117559064 100644
--- a/src/openai/_utils/_transform.py
+++ b/src/openai/_utils/_transform.py
@@ -6,9 +6,8 @@
 
 import pydantic
 
-from ._utils import (
-    is_list,
-    is_mapping,
+from ._utils import is_list, is_mapping
+from ._typing import (
     is_list_type,
     is_union_type,
     extract_type_arg,
diff --git a/src/openai/_utils/_typing.py b/src/openai/_utils/_typing.py
new file mode 100644
index 0000000000..b5e2c2e397
--- /dev/null
+++ b/src/openai/_utils/_typing.py
@@ -0,0 +1,80 @@
+from __future__ import annotations
+
+from typing import Any, cast
+from typing_extensions import Required, Annotated, get_args, get_origin
+
+from .._types import InheritsGeneric
+from .._compat import is_union as _is_union
+
+
+def is_annotated_type(typ: type) -> bool:
+    return get_origin(typ) == Annotated
+
+
+def is_list_type(typ: type) -> bool:
+    return (get_origin(typ) or typ) == list
+
+
+def is_union_type(typ: type) -> bool:
+    return _is_union(get_origin(typ))
+
+
+def is_required_type(typ: type) -> bool:
+    return get_origin(typ) == Required
+
+
+# Extracts T from Annotated[T, ...] or from Required[Annotated[T, ...]]
+def strip_annotated_type(typ: type) -> type:
+    if is_required_type(typ) or is_annotated_type(typ):
+        return strip_annotated_type(cast(type, get_args(typ)[0]))
+
+    return typ
+
+
+def extract_type_arg(typ: type, index: int) -> type:
+    args = get_args(typ)
+    try:
+        return cast(type, args[index])
+    except IndexError as err:
+        raise RuntimeError(f"Expected type {typ} to have a type argument at index {index} but it did not") from err
+
+
+def extract_type_var_from_base(typ: type, *, generic_bases: tuple[type, ...], index: int) -> type:
+    """Given a type like `Foo[T]`, returns the generic type variable `T`.
+
+    This also handles the case where a concrete subclass is given, e.g.
+    ```py
+    class MyResponse(Foo[bytes]):
+        ...
+
+    extract_type_var(MyResponse, bases=(Foo,), index=0) -> bytes
+    ```
+    """
+    cls = cast(object, get_origin(typ) or typ)
+    if cls in generic_bases:
+        # we're given the class directly
+        return extract_type_arg(typ, index)
+
+    # if a subclass is given
+    # ---
+    # this is needed as __orig_bases__ is not present in the typeshed stubs
+    # because it is intended to be for internal use only, however there does
+    # not seem to be a way to resolve generic TypeVars for inherited subclasses
+    # without using it.
+    if isinstance(cls, InheritsGeneric):
+        target_base_class: Any | None = None
+        for base in cls.__orig_bases__:
+            if base.__origin__ in generic_bases:
+                target_base_class = base
+                break
+
+        if target_base_class is None:
+            raise RuntimeError(
+                "Could not find the generic base class;\n"
+                "This should never happen;\n"
+                f"Does {cls} inherit from one of {generic_bases} ?"
+            )
+
+        return extract_type_arg(target_base_class, index)
+
+    raise RuntimeError(f"Could not resolve inner type variable at index {index} for {typ}")
diff --git a/src/openai/_utils/_utils.py b/src/openai/_utils/_utils.py
index c874d3682d..993462a66b 100644
--- a/src/openai/_utils/_utils.py
+++ b/src/openai/_utils/_utils.py
@@ -16,12 +16,11 @@
     overload,
 )
 from pathlib import Path
-from typing_extensions import Required, Annotated, TypeGuard, get_args, get_origin
+from typing_extensions import TypeGuard
 
 import sniffio
 
 from .._types import Headers, NotGiven, FileTypes, NotGivenOr, HeadersLike
-from .._compat import is_union as _is_union
 from .._compat import parse_date as parse_date
 from .._compat import parse_datetime as parse_datetime
 
@@ -166,38 +165,6 @@ def is_list(obj: object) -> TypeGuard[list[object]]:
     return isinstance(obj, list)
 
 
-def is_annotated_type(typ: type) -> bool:
-    return get_origin(typ) == Annotated
-
-
-def is_list_type(typ: type) -> bool:
-    return (get_origin(typ) or typ) == list
-
-
-def is_union_type(typ: type) -> bool:
-    return _is_union(get_origin(typ))
-
-
-def is_required_type(typ: type) -> bool:
-    return get_origin(typ) == Required
-
-
-# Extracts T from Annotated[T, ...] or from Required[Annotated[T, ...]]
-def strip_annotated_type(typ: type) -> type:
-    if is_required_type(typ) or is_annotated_type(typ):
-        return strip_annotated_type(cast(type, get_args(typ)[0]))
-
-    return typ
-
-
-def extract_type_arg(typ: type, index: int) -> type:
-    args = get_args(typ)
-    try:
-        return cast(type, args[index])
-    except IndexError as err:
-        raise RuntimeError(f"Expected type {typ} to have a type argument at index {index} but it did not") from err
-
-
 def deepcopy_minimal(item: _T) -> _T:
     """Minimal reimplementation of copy.deepcopy() that will only copy certain object types:
 

From 3b338a401b206618774291ff8137deb0cc5f6b4c Mon Sep 17 00:00:00 2001
From: Stainless Bot <107565488+stainless-bot@users.noreply.github.com>
Date: Tue, 19 Dec 2023 13:22:14 -0500
Subject: [PATCH 07/10] chore(internal): fix typos (#993)

---
 src/openai/_base_client.py | 2 +-
 src/openai/_streaming.py   | 4 ++--
 2 files changed, 3 insertions(+), 3 deletions(-)

diff --git a/src/openai/_base_client.py b/src/openai/_base_client.py
index 92189617b5..481171a447 100644
--- a/src/openai/_base_client.py
+++ b/src/openai/_base_client.py
@@ -107,7 +107,7 @@
 
 
 class PageInfo:
-    """Stores the necesary information to build the request to retrieve the next page.
+    """Stores the necessary information to build the request to retrieve the next page.
 
     Either `url` or `params` must be set.
     """
diff --git a/src/openai/_streaming.py b/src/openai/_streaming.py
index f1896a242a..85cec70c11 100644
--- a/src/openai/_streaming.py
+++ b/src/openai/_streaming.py
@@ -60,7 +60,7 @@ def __stream__(self) -> Iterator[_T]:
                 data = sse.json()
                 if is_mapping(data) and data.get("error"):
                     raise APIError(
-                        message="An error ocurred during streaming",
+                        message="An error occurred during streaming",
                         request=self.response.request,
                         body=data["error"],
                     )
@@ -134,7 +134,7 @@ async def __stream__(self) -> AsyncIterator[_T]:
                 data = sse.json()
                 if is_mapping(data) and data.get("error"):
                     raise APIError(
-                        message="An error ocurred during streaming",
+                        message="An error occurred during streaming",
                         request=self.response.request,
                         body=data["error"],
                     )

From 0c2da84badf416f8b2213983f68bd2b6f9e52f2b Mon Sep 17 00:00:00 2001
From: Stainless Bot <107565488+stainless-bot@users.noreply.github.com>
Date: Tue, 19 Dec 2023 14:05:05 -0500
Subject: [PATCH 08/10] chore(package): bump minimum typing-extensions to 4.7
 (#994)

---
 pyproject.toml | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/pyproject.toml b/pyproject.toml
index 0cf709a726..24498b18fe 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -10,7 +10,7 @@ authors = [
 dependencies = [
     "httpx>=0.23.0, <1",
     "pydantic>=1.9.0, <3",
-    "typing-extensions>=4.5, <5",
+    "typing-extensions>=4.7, <5",
     "anyio>=3.5.0, <5",
     "distro>=1.7.0, <2",
     "sniffio",

From 7bf9b75067905449e83e828c12eb384022cff6ca Mon Sep 17 00:00:00 2001
From: Stainless Bot <107565488+stainless-bot@users.noreply.github.com>
Date: Tue, 19 Dec 2023 18:17:29 -0500
Subject: [PATCH 09/10] feat(api): add additional instructions for runs (#995)

---
 .../resources/beta/threads/runs/runs.py       | 22 +++++++++++++++----
 .../types/beta/threads/run_create_params.py   | 14 +++++++++---
 tests/api_resources/beta/threads/test_runs.py |  2 ++
 3 files changed, 31 insertions(+), 7 deletions(-)

diff --git a/src/openai/resources/beta/threads/runs/runs.py b/src/openai/resources/beta/threads/runs/runs.py
index 969bfab70a..aea3b8cefc 100644
--- a/src/openai/resources/beta/threads/runs/runs.py
+++ b/src/openai/resources/beta/threads/runs/runs.py
@@ -42,6 +42,7 @@ def create(
         thread_id: str,
         *,
         assistant_id: str,
+        additional_instructions: Optional[str] | NotGiven = NOT_GIVEN,
         instructions: Optional[str] | NotGiven = NOT_GIVEN,
         metadata: Optional[object] | NotGiven = NOT_GIVEN,
         model: Optional[str] | NotGiven = NOT_GIVEN,
@@ -61,8 +62,13 @@ def create(
               [assistant](https://platform.openai.com/docs/api-reference/assistants) to use to
               execute this run.
 
-          instructions: Override the default system message of the assistant. This is useful for
-              modifying the behavior on a per-run basis.
+          additional_instructions: Appends additional instructions at the end of the instructions for the run. This
+              is useful for modifying the behavior on a per-run basis without overriding other
+              instructions.
+
+          instructions: Overrides the
+              [instructions](https://platform.openai.com/docs/api-reference/assistants/createAssistant)
+              of the assistant. This is useful for modifying the behavior on a per-run basis.
 
           metadata: Set of 16 key-value pairs that can be attached to an object. This can be useful
               for storing additional information about the object in a structured format. Keys
@@ -91,6 +97,7 @@ def create(
             body=maybe_transform(
                 {
                     "assistant_id": assistant_id,
+                    "additional_instructions": additional_instructions,
                     "instructions": instructions,
                     "metadata": metadata,
                     "model": model,
@@ -332,6 +339,7 @@ async def create(
         thread_id: str,
         *,
         assistant_id: str,
+        additional_instructions: Optional[str] | NotGiven = NOT_GIVEN,
         instructions: Optional[str] | NotGiven = NOT_GIVEN,
         metadata: Optional[object] | NotGiven = NOT_GIVEN,
         model: Optional[str] | NotGiven = NOT_GIVEN,
@@ -351,8 +359,13 @@ async def create(
               [assistant](https://platform.openai.com/docs/api-reference/assistants) to use to
               execute this run.
 
-          instructions: Override the default system message of the assistant. This is useful for
-              modifying the behavior on a per-run basis.
+          additional_instructions: Appends additional instructions at the end of the instructions for the run. This
+              is useful for modifying the behavior on a per-run basis without overriding other
+              instructions.
+
+          instructions: Overrides the
+              [instructions](https://platform.openai.com/docs/api-reference/assistants/createAssistant)
+              of the assistant. This is useful for modifying the behavior on a per-run basis.
 
           metadata: Set of 16 key-value pairs that can be attached to an object. This can be useful
               for storing additional information about the object in a structured format. Keys
@@ -381,6 +394,7 @@ async def create(
             body=maybe_transform(
                 {
                     "assistant_id": assistant_id,
+                    "additional_instructions": additional_instructions,
                     "instructions": instructions,
                     "metadata": metadata,
                     "model": model,
diff --git a/src/openai/types/beta/threads/run_create_params.py b/src/openai/types/beta/threads/run_create_params.py
index df92f4fd2c..a4f41a9338 100644
--- a/src/openai/types/beta/threads/run_create_params.py
+++ b/src/openai/types/beta/threads/run_create_params.py
@@ -24,10 +24,18 @@ class RunCreateParams(TypedDict, total=False):
     execute this run.
     """
 
-    instructions: Optional[str]
-    """Override the default system message of the assistant.
+    additional_instructions: Optional[str]
+    """Appends additional instructions at the end of the instructions for the run.
 
-    This is useful for modifying the behavior on a per-run basis.
+    This is useful for modifying the behavior on a per-run basis without overriding
+    other instructions.
+    """
+
+    instructions: Optional[str]
+    """
+    Overrides the
+    [instructions](https://platform.openai.com/docs/api-reference/assistants/createAssistant)
+    of the assistant. This is useful for modifying the behavior on a per-run basis.
     """
 
     metadata: Optional[object]
diff --git a/tests/api_resources/beta/threads/test_runs.py b/tests/api_resources/beta/threads/test_runs.py
index d323dfc354..494cae2656 100644
--- a/tests/api_resources/beta/threads/test_runs.py
+++ b/tests/api_resources/beta/threads/test_runs.py
@@ -34,6 +34,7 @@ def test_method_create_with_all_params(self, client: OpenAI) -> None:
         run = client.beta.threads.runs.create(
             "string",
             assistant_id="string",
+            additional_instructions="string",
             instructions="string",
             metadata={},
             model="string",
@@ -180,6 +181,7 @@ async def test_method_create_with_all_params(self, client: AsyncOpenAI) -> None:
         run = await client.beta.threads.runs.create(
             "string",
             assistant_id="string",
+            additional_instructions="string",
             instructions="string",
             metadata={},
             model="string",

From 157d5580f34e365f797ee26bb7c99bbc5f24db10 Mon Sep 17 00:00:00 2001
From: Stainless Bot <107565488+stainless-bot@users.noreply.github.com>
Date: Tue, 19 Dec 2023 18:18:08 -0500
Subject: [PATCH 10/10] release: 1.6.0

---
 .release-please-manifest.json |  2 +-
 CHANGELOG.md                  | 24 ++++++++++++++++++++++++
 pyproject.toml                |  2 +-
 src/openai/_version.py        |  2 +-
 4 files changed, 27 insertions(+), 3 deletions(-)

diff --git a/.release-please-manifest.json b/.release-please-manifest.json
index fbd9082d71..7deae33804 100644
--- a/.release-please-manifest.json
+++ b/.release-please-manifest.json
@@ -1,3 +1,3 @@
 {
-  ".": "1.5.0"
+  ".": "1.6.0"
 }
\ No newline at end of file
diff --git a/CHANGELOG.md b/CHANGELOG.md
index 757d79af62..399e3aaebd 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -1,5 +1,29 @@
 # Changelog
 
+## 1.6.0 (2023-12-19)
+
+Full Changelog: [v1.5.0...v1.6.0](https://github.com/openai/openai-python/compare/v1.5.0...v1.6.0)
+
+### Features
+
+* **api:** add additional instructions for runs ([#995](https://github.com/openai/openai-python/issues/995)) ([7bf9b75](https://github.com/openai/openai-python/commit/7bf9b75067905449e83e828c12eb384022cff6ca))
+
+
+### Chores
+
+* **cli:** fix typo in completions ([#985](https://github.com/openai/openai-python/issues/985)) ([d1e9e8f](https://github.com/openai/openai-python/commit/d1e9e8f24df366bb7b796c55a98247c025d229f5))
+* **cli:** fix typo in completions ([#986](https://github.com/openai/openai-python/issues/986)) ([626bc34](https://github.com/openai/openai-python/commit/626bc34d82a7057bac99f8b556f9e5f60c261ee7))
+* **internal:** fix binary response tests ([#983](https://github.com/openai/openai-python/issues/983)) ([cfb7e30](https://github.com/openai/openai-python/commit/cfb7e308393f2e912e959dd10d68096dd5b3ab9c))
+* **internal:** fix typos ([#993](https://github.com/openai/openai-python/issues/993)) ([3b338a4](https://github.com/openai/openai-python/commit/3b338a401b206618774291ff8137deb0cc5f6b4c))
+* **internal:** minor utils restructuring ([#992](https://github.com/openai/openai-python/issues/992)) ([5ba576a](https://github.com/openai/openai-python/commit/5ba576ae38d2c4c4d32a21933e0d68e0bc2f0d49))
+* **package:** bump minimum typing-extensions to 4.7 ([#994](https://github.com/openai/openai-python/issues/994)) ([0c2da84](https://github.com/openai/openai-python/commit/0c2da84badf416f8b2213983f68bd2b6f9e52f2b))
+* **streaming:** update constructor to use direct client names ([#991](https://github.com/openai/openai-python/issues/991)) ([6c3427d](https://github.com/openai/openai-python/commit/6c3427dac8c414658516aeb4caf5d5fd8b11097b))
+
+
+### Documentation
+
+* upgrade models in examples to latest version ([#989](https://github.com/openai/openai-python/issues/989)) ([cedd574](https://github.com/openai/openai-python/commit/cedd574e5611f3e71e92b523a72ba87bcfe546f1))
+
 ## 1.5.0 (2023-12-17)
 
 Full Changelog: [v1.4.0...v1.5.0](https://github.com/openai/openai-python/compare/v1.4.0...v1.5.0)
diff --git a/pyproject.toml b/pyproject.toml
index 24498b18fe..91d3d79219 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -1,6 +1,6 @@
 [project]
 name = "openai"
-version = "1.5.0"
+version = "1.6.0"
 description = "The official Python library for the openai API"
 readme = "README.md"
 license = "Apache-2.0"
diff --git a/src/openai/_version.py b/src/openai/_version.py
index 9dbb5b1401..9b01b6fcb1 100644
--- a/src/openai/_version.py
+++ b/src/openai/_version.py
@@ -1,4 +1,4 @@
 # File generated from our OpenAPI spec by Stainless.
 
 __title__ = "openai"
-__version__ = "1.5.0"  # x-release-please-version
+__version__ = "1.6.0"  # x-release-please-version