fix(audio): correct types for transcriptions / translations

api.md

@@ -122,24 +122,30 @@ from openai.types import AudioModel, AudioResponseFormat
 Types:
 
 ```python
-from openai.types.audio import Transcription
+from openai.types.audio import (
+    Transcription,
+    TranscriptionSegment,
+    TranscriptionVerbose,
+    TranscriptionWord,
+    TranscriptionCreateResponse,
+)
 ```
 
 Methods:
 
-- <code title="post /audio/transcriptions">client.audio.transcriptions.<a href="./src/openai/resources/audio/transcriptions.py">create</a>(\*\*<a href="src/openai/types/audio/transcription_create_params.py">params</a>) -> <a href="./src/openai/types/audio/transcription.py">Transcription</a></code>
+- <code title="post /audio/transcriptions">client.audio.transcriptions.<a href="./src/openai/resources/audio/transcriptions.py">create</a>(\*\*<a href="src/openai/types/audio/transcription_create_params.py">params</a>) -> <a href="./src/openai/types/audio/transcription_create_response.py">TranscriptionCreateResponse</a></code>
 
 ## Translations
 
 Types:
 
 ```python
-from openai.types.audio import Translation
+from openai.types.audio import Translation, TranslationVerbose, TranslationCreateResponse
 ```
 
 Methods:
 
-- <code title="post /audio/translations">client.audio.translations.<a href="./src/openai/resources/audio/translations.py">create</a>(\*\*<a href="src/openai/types/audio/translation_create_params.py">params</a>) -> <a href="./src/openai/types/audio/translation.py">Translation</a></code>
+- <code title="post /audio/translations">client.audio.translations.<a href="./src/openai/resources/audio/translations.py">create</a>(\*\*<a href="src/openai/types/audio/translation_create_params.py">params</a>) -> <a href="./src/openai/types/audio/translation_create_response.py">TranslationCreateResponse</a></code>
 
 ## Speech
 

src/openai/_utils/_reflection.py

@@ -15,6 +15,7 @@ def assert_signatures_in_sync(
     check_func: Callable[..., Any],
     *,
     exclude_params: set[str] = set(),
+    description: str = "",
 ) -> None:
     """Ensure that the signature of the second function matches the first."""
 
@@ -39,4 +40,6 @@ def assert_signatures_in_sync(
             continue
 
     if errors:
-        raise AssertionError(f"{len(errors)} errors encountered when comparing signatures:\n\n" + "\n\n".join(errors))
+        raise AssertionError(
+            f"{len(errors)} errors encountered when comparing signatures{description}:\n\n" + "\n\n".join(errors)
+        )

src/openai/resources/audio/transcriptions.py

@@ -2,8 +2,9 @@
 
 from __future__ import annotations
 
-from typing import List, Union, Mapping, cast
-from typing_extensions import Literal
+import logging
+from typing import TYPE_CHECKING, List, Union, Mapping, cast
+from typing_extensions import Literal, overload, assert_never
 
 import httpx
 
@@ -24,9 +25,12 @@
 from ...types.audio_model import AudioModel
 from ...types.audio.transcription import Transcription
 from ...types.audio_response_format import AudioResponseFormat
+from ...types.audio.transcription_verbose import TranscriptionVerbose
 
 __all__ = ["Transcriptions", "AsyncTranscriptions"]
 
+log: logging.Logger = logging.getLogger("openai.audio.transcriptions")
+
 
 class Transcriptions(SyncAPIResource):
     @cached_property
@@ -48,14 +52,53 @@ def with_streaming_response(self) -> TranscriptionsWithStreamingResponse:
         """
         return TranscriptionsWithStreamingResponse(self)
 
+    @overload
+    def create(
+        self,
+        *,
+        file: FileTypes,
+        model: Union[str, AudioModel],
+        response_format: Union[Literal["json"], NotGiven] = NOT_GIVEN,
+        language: str | NotGiven = NOT_GIVEN,
+        prompt: str | NotGiven = NOT_GIVEN,
+        temperature: float | NotGiven = NOT_GIVEN,
+        timestamp_granularities: List[Literal["word", "segment"]] | 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,
+    ) -> Transcription: ...
+
+    @overload
+    def create(
+        self,
+        *,
+        file: FileTypes,
+        model: Union[str, AudioModel],
+        response_format: Literal["verbose_json"],
+        language: str | NotGiven = NOT_GIVEN,
+        prompt: str | NotGiven = NOT_GIVEN,
+        temperature: float | NotGiven = NOT_GIVEN,
+        timestamp_granularities: List[Literal["word", "segment"]] | 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,
+    ) -> TranscriptionVerbose: ...
+
+    @overload
     def create(
         self,
         *,
         file: FileTypes,
         model: Union[str, AudioModel],
+        response_format: Literal["text", "srt", "vtt"],
         language: str | NotGiven = NOT_GIVEN,
         prompt: str | NotGiven = NOT_GIVEN,
-        response_format: AudioResponseFormat | NotGiven = NOT_GIVEN,
         temperature: float | NotGiven = NOT_GIVEN,
         timestamp_granularities: List[Literal["word", "segment"]] | NotGiven = NOT_GIVEN,
         # Use the following arguments if you need to pass additional parameters to the API that aren't available via kwargs.
@@ -64,7 +107,25 @@ def create(
         extra_query: Query | None = None,
         extra_body: Body | None = None,
         timeout: float | httpx.Timeout | None | NotGiven = NOT_GIVEN,
-    ) -> Transcription:
+    ) -> str: ...
+
+    def create(
+        self,
+        *,
+        file: FileTypes,
+        model: Union[str, AudioModel],
+        language: str | NotGiven = NOT_GIVEN,
+        prompt: str | NotGiven = NOT_GIVEN,
+        response_format: Union[AudioResponseFormat, NotGiven] = NOT_GIVEN,
+        temperature: float | NotGiven = NOT_GIVEN,
+        timestamp_granularities: List[Literal["word", "segment"]] | 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,
+    ) -> Transcription | TranscriptionVerbose | str:
         """
         Transcribes audio into the input language.
 
@@ -131,7 +192,7 @@ def create(
             options=make_request_options(
                 extra_headers=extra_headers, extra_query=extra_query, extra_body=extra_body, timeout=timeout
             ),
-            cast_to=Transcription,
+            cast_to=_get_response_format_type(response_format),
         )
 
 
@@ -155,14 +216,15 @@ def with_streaming_response(self) -> AsyncTranscriptionsWithStreamingResponse:
         """
         return AsyncTranscriptionsWithStreamingResponse(self)
 
+    @overload
     async def create(
         self,
         *,
         file: FileTypes,
         model: Union[str, AudioModel],
+        response_format: Union[Literal["json"], NotGiven] = NOT_GIVEN,
         language: str | NotGiven = NOT_GIVEN,
         prompt: str | NotGiven = NOT_GIVEN,
-        response_format: AudioResponseFormat | NotGiven = NOT_GIVEN,
         temperature: float | NotGiven = NOT_GIVEN,
         timestamp_granularities: List[Literal["word", "segment"]] | NotGiven = NOT_GIVEN,
         # Use the following arguments if you need to pass additional parameters to the API that aren't available via kwargs.
@@ -171,7 +233,63 @@ async def create(
         extra_query: Query | None = None,
         extra_body: Body | None = None,
         timeout: float | httpx.Timeout | None | NotGiven = NOT_GIVEN,
-    ) -> Transcription:
+    ) -> Transcription: ...
+
+    @overload
+    async def create(
+        self,
+        *,
+        file: FileTypes,
+        model: Union[str, AudioModel],
+        response_format: Literal["verbose_json"],
+        language: str | NotGiven = NOT_GIVEN,
+        prompt: str | NotGiven = NOT_GIVEN,
+        temperature: float | NotGiven = NOT_GIVEN,
+        timestamp_granularities: List[Literal["word", "segment"]] | 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,
+    ) -> TranscriptionVerbose: ...
+
+    @overload
+    async def create(
+        self,
+        *,
+        file: FileTypes,
+        model: Union[str, AudioModel],
+        response_format: Literal["text", "srt", "vtt"],
+        language: str | NotGiven = NOT_GIVEN,
+        prompt: str | NotGiven = NOT_GIVEN,
+        temperature: float | NotGiven = NOT_GIVEN,
+        timestamp_granularities: List[Literal["word", "segment"]] | 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,
+    ) -> str: ...
+
+    async def create(
+        self,
+        *,
+        file: FileTypes,
+        model: Union[str, AudioModel],
+        language: str | NotGiven = NOT_GIVEN,
+        prompt: str | NotGiven = NOT_GIVEN,
+        response_format: Union[AudioResponseFormat, NotGiven] = NOT_GIVEN,
+        temperature: float | NotGiven = NOT_GIVEN,
+        timestamp_granularities: List[Literal["word", "segment"]] | 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,
+    ) -> Transcription | TranscriptionVerbose | str:
         """
         Transcribes audio into the input language.
 
@@ -238,7 +356,7 @@ async def create(
             options=make_request_options(
                 extra_headers=extra_headers, extra_query=extra_query, extra_body=extra_body, timeout=timeout
             ),
-            cast_to=Transcription,
+            cast_to=_get_response_format_type(response_format),
         )
 
 
@@ -276,3 +394,22 @@ def __init__(self, transcriptions: AsyncTranscriptions) -> None:
         self.create = async_to_streamed_response_wrapper(
             transcriptions.create,
         )
+
+
+def _get_response_format_type(
+    response_format: Literal["json", "text", "srt", "verbose_json", "vtt"] | NotGiven,
+) -> type[Transcription | TranscriptionVerbose | str]:
+    if isinstance(response_format, NotGiven) or response_format is None:  # pyright: ignore[reportUnnecessaryComparison]
+        return Transcription
+
+    if response_format == "json":
+        return Transcription
+    elif response_format == "verbose_json":
+        return TranscriptionVerbose
+    elif response_format == "srt" or response_format == "text" or response_format == "vtt":
+        return str
+    elif TYPE_CHECKING:
+        assert_never(response_format)
+    else:
+        log.warn("Unexpected audio response format: %s", response_format)
+        return Transcription