[pull] main from openai:main - #140

- [x] I understand that this repository is auto-generated and my pull
request may not be merged

## Changes being requested

Move the duplicated embeddings response decoder into the SDK-owned
`lib/_parsing/_embeddings.py` module. Both resource methods use the
existing
`post_parser` hook. The decoder logic is unchanged: omitted encoding
requests
base64 and decodes it in place; explicit formats pass through; empty
data keeps
the existing `ValueError`; NumPy and stdlib decoding keep their existing
errors.

Public signatures, exports, request fields and order, sync/async
behavior, and
raw/streaming response wrappers are unchanged. No compiler, schema,
dependency,
or generation-metadata changes are included.

## Additional context & links

New handwritten coverage is in `tests/lib/test_embeddings.py`:

- `test_decode_preserves_response_and_non_string_vectors` checks
in-place
decoding, both omitted-value sentinels, and untouched non-string
vectors.
- `test_explicit_format_is_untouched`,
  `test_missing_data_keeps_existing_error`, and
`test_invalid_data_preserves_decoder_errors` cover passthrough and
failures.
- `test_sync_create_uses_decoder` and `test_async_create_uses_decoder`
cover
normal, raw, and streaming responses with both decoder implementations.

Validation:

- Command: `python -m pytest -q -n 4
tests/api_resources/test_embeddings.py tests/lib/test_embeddings.py`
  passed 77 tests under Pydantic v2 and 77 under Pydantic v1.
- `./scripts/format` and `./scripts/lint` passed, including Ruff,
Pyright, mypy,
and import checks. Unrelated formatter-only reporter edits are excluded.
- `./scripts/build` passed; the wheel and source distribution both
contain the
  new private helper.
- The public custom-code report verifies 41 mixed files, only the
embeddings
customization changed, and 40 other customizations unchanged. The
embeddings
patch shrinks from +73/-22 to +27/-21. `.castiron.stats.yml` is
unchanged.

Command to reproduce the report from this branch:

```sh
$ python3 scripts/castiron/custom_code_report.py report \
    --base 8edd9ae \
    --head 38cf9c9 \
    --fetch --require-head-hash --public \
    --out /tmp/castiron-embedding-decoder
```

- [x] I understand that this repository is auto-generated and my pull
request may not be merged

## Changes being requested

Move the transcription and translation response-format selectors into
the
SDK-owned `lib/_parsing/_audio.py` module. Keep the existing private
resource
functions as thin wrappers, including their signatures and module
loggers.
The selection branches are unchanged, including `None`/omitted handling,
diarized transcription, text formats, and the existing fallback
warnings.

Every resource class is unchanged: public overloads, return annotations,
request fields and order, decorators, and sync/async response wrappers
stay
intact. No compiler, schema, dependency, or generation-metadata changes
are
included.

## Additional context & links

New handwritten coverage is in
`tests/lib/test_audio_response_format.py`:

- `test_supported_formats_keep_exact_response_classes` and
  `test_default_compatibility_values` check the exact response classes.
- `test_fallback_keeps_resource_logger_and_method` and
  `test_fallback_keeps_historical_logger_name_and_warning` preserve the
fallback class, logger, warning text, and existing `warn`/`warning`
behavior.
- `test_sync_create_uses_selected_response_class` and
`test_async_create_uses_selected_response_class` cover every supported
format
through normal, raw, and streaming responses in strict and loose modes.
- `test_sync_transcription_stream_still_yields_events` and
`test_async_transcription_stream_still_yields_events` retain
event-stream
  behavior. The existing overload checks remain in
`tests/lib/test_audio.py::test_translation_create_overloads_in_sync` and
`tests/lib/test_audio.py::test_transcription_create_overloads_in_sync`.

Validation:

- Command: `python -m pytest -n 0
tests/lib/test_audio_response_format.py tests/lib/test_audio.py
tests/api_resources/audio/test_transcriptions.py
tests/api_resources/audio/test_translations.py`
  passed 245 tests under Pydantic v2 and 245 under Pydantic v1.
- `./scripts/format` and `./scripts/lint` passed, including Ruff,
Pyright,
mypy, and import checks. Unrelated formatter-only reporter edits are
excluded.
- `./scripts/build` passed; the wheel and source distribution both
contain
  the new private helper.
- The public custom-code report verifies 41 mixed files, with only the
two
audio customizations changed and 39 others unchanged. Transcriptions
shrink
from +160/-53 to +145/-52; translations shrink from +155/-39 to
+143/-39.
  `.castiron.stats.yml` is unchanged.

Command to reproduce the report from this branch:

```sh
$ python3 scripts/castiron/custom_code_report.py report \
    --base 8edd9ae \
    --head 676b9e7 \
    --fetch --require-head-hash --public \
    --out /tmp/castiron-audio-format-selection
```

- [x] I understand that this repository is auto-generated and my pull
request may not be merged

## Changes being requested

Move the identical webhook header/timestamp/HMAC verification core into
the
SDK-owned `lib/_webhooks.py` module. The public sync and async wrappers
retain
their signatures, client-secret fallback, mismatch errors, and their
existing
exception-chaining differences. Both `unwrap` methods are unchanged.

The helper keeps the same replay-window checks, secret decoding, signed
bytes,
signature order, and `hmac.compare_digest` calls. This is a
behavior-preserving
extraction, not a change to accepted signatures or verification policy.
Generation metadata, dependencies, workflows, and the API reference are
unchanged.

The verified custom-code report keeps 40 mixed files and changes only
the
webhook resource's patch: **+179/-3 → +89/-3**. The other 39
customizations are
unchanged.

## Additional context & links

Please get SDK CODEOWNER review for this verification-boundary change.

The existing 71 cases remain unchanged in

[`tests/lib/test_webhooks.py::TestWebhooks`](https://github.com/openai/openai-python/blob/ece4324da0b96f848b48bdef090a372ff0a1db26/tests/lib/test_webhooks.py#L32)
and

[`tests/lib/test_webhooks.py::TestAsyncWebhooks`](https://github.com/openai/openai-python/blob/ece4324da0b96f848b48bdef090a372ff0a1db26/tests/lib/test_webhooks.py#L170).
The new

[`tests/lib/test_webhook_signature.py`](https://github.com/openai/openai-python/blob/35d955800683b814490c7a583b91611c30df9a71/tests/lib/test_webhook_signature.py)
adds 87 cases for raw/prefixed/empty secrets, text/bytes payloads,
replay-window
boundaries, exact timestamp text, header order, malformed inputs, and
constant-time comparison order. In particular,

[`test_missing_secret_preserves_wrapper_exception_chaining`](https://github.com/openai/openai-python/blob/35d955800683b814490c7a583b91611c30df9a71/tests/lib/test_webhook_signature.py#L146)
and

[`test_mismatch_preserves_wrapper_exception_chaining`](https://github.com/openai/openai-python/blob/35d955800683b814490c7a583b91611c30df9a71/tests/lib/test_webhook_signature.py#L161)
pin the existing sync/async differences.

Validation: exact-source/AST preservation check; `./scripts/format`;
`./scripts/lint` (Ruff, Pyright, mypy, import); wheel and sdist builds
with the
new helper included; **158 tests passed under Pydantic v2 and 158 under
v1**.