v0.27.1

• Adds WordTimeOffset, the ability to get timestamps as part of the response (#3678)

v0.27.0

• Adding a (mostly) autogenerated surface side-by-side with the previously provided Client (#3607)
• Migration guide: https://cloud.google.com/speech/docs/python-client-migration
• Fixing attribute "typo" in HTTP transport (speechContext -> speechContexts) (#3570, H/T to @p-siegel)

PyPI: https://pypi.org/project/google-cloud-speech/0.27.0/

v0.26.0

• Update google-cloud-core dependency to ~= 0.25.

v0.25.1

• Do not force long-running requests to be LINEAR16 only in Speech. (#3291)
• Add new encodings. (#3292)
• Do not validate sample rate. (#3290)

v0.25.0

This updates our manual client library to the Speech v1 API. (#3266)

This entails several backwards incompatible changes:

• The language_code parameter is no longer optional anywhere. It must be explicitly specified, and does not default to 'en-US'.
• The sync_recognize method has been renamed to recognize on every class where it appears.
• The async_recognize method has been renamed to long_running_recognize on every class where it appears.
• The sample_rate parameter and property has been renamed to sample_rate_hertz everywhere it appears.

Additionally, the backend API contains a backwards incompatible change which does not require a code change in the client library, but will likely require one downstream: The START_OF_SPEECH, END_OF_SPEECH, and END_OF_AUDIO events have been removed.

v0.24.0

• Renaming http argument to Client as _http. (#3235)
• Make Client explicitly unpickleable. (#3230)
• Added __all__ to google.speech package. (#3146)
• Docs fix: Fix broken link. (#3198)

PyPI: https://pypi.python.org/pypi/google-cloud-speech/0.24.0

v0.23.0

• Allow bytes and stream-like input for Sample.content (#2850)
• Remove default sample_rate (#2963)
• Changing output of sync_recognize to add an extra "layer" of data (#2962)
• Storing all connection data on Client instead of on private Connection (#2970)
• Fix handling of long-running operations (#2965)
• Changing from Beta to Alpha classifier (#3027)
• Renaming connection module as _http and making Connection class non-public. (#3033)
• google.cloud.speech.__version__ attribute added (#3046)
• On HTTP requests, sending X-Goog-API-Client header with value gl-python/x.y.z gccl/a.b.c where x.y.z is the version of Python that is running and a.b.c is the version of the speech package being used (#3046)
• Sending speech version with gRPC requests (#3048, related to #3046)

PyPI: https://pypi.python.org/pypi/google-cloud-speech/0.23.0

v0.22.0

• change is dependence on new version of google-cloud-core
• Switch from oauth2client to google-auth (#2726, #2828)
• Upgrading GAPIC deps to 0.14 series. (#2795)

PyPI: https://pypi.python.org/pypi/google-cloud-speech/0.22.0

v0.21.0

Speech

• sync_recognize() (#2344 Thanks @Fematich!, #2615)
• async_recognize() (#2663, #2426)
• streaming_recognize() (#2680)