Basic implementation of Speech API. by Fematich · Pull Request #2344 · googleapis/google-cloud-python

Fematich · 2016-09-18T12:16:34Z

I have made a basic implementation of Speech API.
This currently only supports syncrecognize from the REST-API.

I think I followed all the contribution guidelines:

Documented in both the API and narrative documentation (in docs/).
Fully working on the following CPython versions: 2.7, 3.4, and 3.5 on both UNIX and Windows.
Feature added no new dependencies.

However, this is my first contribution, so in case I missed something, please let me know and I will see how to fix it!

tseaver

Thank you very much for the patch -- it is very well done!

Sign in to view

 bucket = client.get_bucket('<your-bucket-name>')
 blob = bucket.blob('my-test-file.txt')
 blob.upload_from_string('this is test content!')
+


Sign in to view

+
+At this moment we only support one method of the Speech API:
+
+- `syncrecognize`_


Sign in to view

+
+- `syncrecognize`_
+
+Synchronous Recognize


Sign in to view

+---------------------
+
+The :meth:`~google.cloud.speech.Client.syncrecognize` method
+does speech to text on a file and returns the text


Sign in to view

+
+The :meth:`~google.cloud.speech.Client.syncrecognize` method
+does speech to text on a file and returns the text
+as a :class:`list` of tuples dicts (each containing a transcript an a confidence value).


Sign in to view

+ for param_name, param in required_params:
+ if param is None:
+ message = '%r cannot be None' % (param_name)
+ raise ValueError(message)


Sign in to view

+ if param is None:
+ message = '%r cannot be None' % (param_name)
+ raise ValueError(message)
+ config = dict(required_params)


Sign in to view

+ ('maxAlternatives', max_alternatives),
+ ('profanityFilter', profanity_filter)]:
+ if param is not None:
+ config[param_name] = param


Sign in to view

+ sample_rate,
+ max_alternatives=2,
+ speech_context=hints)
+ self.assertEqual(speechrecognition_result[0]["transcript"], 'hello')


Sign in to view

+ with self.assertRaises(ValueError):
+ client.syncrecognize(None, None, None, None)
+ with self.assertRaises(ValueError):
+ client.syncrecognize(None, "uri", None, None)


…partly updated docs

Fematich · 2016-09-18T23:42:51Z

Thank you for your feedback! I have implemented your suggestions, except for the additional examples with extra arguments in the docs. I will try to finish this tomorrow.

Sign in to view

- supported, which must be specified in the following
- format: gs://bucket_name/object_name
+ :type content: bytes
+ :param content: Byte stream of audio.


daspecster

Awesome! Thank you!

Sign in to view

+ """Client to bundle configuration needed for API requests.
+
+ :type project: str
+ :param project: the project which the client acts on behalf of. Will be


Sign in to view

+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+"""Basic client for Google Cloud Speech API."""


Sign in to view

 bucket = client.get_bucket('<your-bucket-name>')
 blob = bucket.blob('my-test-file.txt')
- blob.upload_from_string('this is test content!')
+ blob.upload_from_string('this is test content!')


Sign in to view

+ .. code-block:: python
+
+ >>> alternatives = client.sync_recognize(None,"gs://my-bucket/recording.flac",
+ ... "FLAC", 16000, max_alternatives=2):


Sign in to view

+ >>> for alternative in alternatives:
+ ... print('=' * 20)
+ ... print(' transcript: %s' % (alternative["transcript"],))
+ ... print(' confidence: %s' % (alternative["confidence"],))


Sign in to view

+ ... print(' confidence: %s' % (alternative["confidence"],))
+ ====================
+ transcript: Hello, this is a test
+ confidence: 0.81


Sign in to view

+
+"""Google Cloud Speech API wrapper."""
+
+from google.cloud.speech.client import Client, Encoding


Sign in to view

+ response = client.sync_recognize(_AUDIO_CONTENT,
+ encoding,
+ self.SAMPLE_RATE,
+ language_code="EN",


Sign in to view

+
+ self.assertEqual(REQUEST,
+ client.connection._requested[0]['data'])
+ self.assertEqual(response[0]["transcript"], 'hello')


Sign in to view

+ },
+ "audio": {
+ "content": _B64_AUDIO_CONTENT
+ }


Sign in to view

+
+ self.assertEqual(REQUEST,
+ client.connection._requested[0]['data'])
+ self.assertEqual(response[0]["transcript"], 'hello')


Sign in to view

+
+class _Credentials(object):
+
+ _scopes = ('https://www.googleapis.com/auth/cloud-platform')


dhermes · 2016-09-19T16:38:05Z

@Fematich Really great work! Especially impressed that you have Travis 100% green on your first try!

Fematich · 2016-09-20T22:59:14Z

I have updated the code (included source_uri again and updated formatting/unit_tests).
I still have to expand speech-usage.rst and the more comprehensive checking of the contents of client.connection._requested and response

Sign in to view

+
+ :type source_uri: str
+ :param source_uri: URI that points to a file that contains audio
+ data bytes as specified in RecognitionConfig.


Sign in to view

+ between 0 and 1.
+ """
+
+ if (content is None) and (source_uri is None):


Sign in to view

+ raise ValueError('content and source_uri cannot be both equal to\
+ None')
+
+ if (content is not None) and (source_uri is not None):


Sign in to view

+ """
+
+ if (content is None) and (source_uri is None):
+ raise ValueError('content and source_uri cannot be both equal to\


Sign in to view

+ None')
+
+ if (content is not None) and (source_uri is not None):
+ raise ValueError('content and source_uri cannot be both different from\


Sign in to view

+ if sample_rate is None:
+ raise ValueError('sample_rate cannot be None')
+
+ if content is not None:


Sign in to view

+}
+
+SYNC_RECOGNIZE_EMPTY_RESPONSE = {
+ 'results': []


Sign in to view

+
+import unittest
+
+_AUDIO_SOURCE_URI = 'gs://sample-bucket/sample-recording.flac'


tseaver · 2016-09-21T14:46:13Z

@Fematich Thanks for addressing my review issues. I'll let @dhermes finish the review and merge. Thanks again for your work!

dhermes · 2016-09-21T17:22:07Z

I'm going to merge as-is. I'll send a PR addressing most of my concerns. Remaining issues outside of that PR will be:

dhermes · 2016-09-21T17:29:06Z

@Fematich Thanks for your great work! We decided to go ahead and merge to speed up the process.

We'd like to get this library feature-complete ASAP so we didn't want the review cycle to slow us down on getting the first chunks in. From here, a person working on this library full-time will take it to the finish line. You were great!

Fematich · 2016-09-21T18:15:19Z

Thank you it was a pleasure :-)!

Review cleanup for PR #2344.

Updates from #2344 for speech API.

Updates from googleapis#2344 for speech API.

PR created by the Librarian CLI to initialize a release. Merging this PR will auto trigger a release. Librarian Version: v0.7.0 Language Image: us-central1-docker.pkg.dev/cloud-sdk-librarian-prod/images-prod/python-librarian-generator@sha256:c8612d3fffb3f6a32353b2d1abd16b61e87811866f7ec9d65b59b02eb452a620 <details><summary>google-cloud-bigquery: 3.39.0</summary> ## [3.39.0](googleapis/python-bigquery@v3.38.0...v3.39.0) (2025-12-12) ### Features * adds support for Python runtime 3.14 (#2322) ([6065e14c](googleapis/python-bigquery@6065e14c)) * Add ExternalRuntimeOptions to BigQuery routine (#2311) ([fa76e310](googleapis/python-bigquery@fa76e310)) ### Bug Fixes * remove ambiguous error codes from query retries (#2308) ([8bbd3d01](googleapis/python-bigquery@8bbd3d01)) * include `io.Base` in the `PathType` (#2323) ([b11e09cb](googleapis/python-bigquery@b11e09cb)) * honor custom `retry` in `job.result()` (#2302) ([e118b029](googleapis/python-bigquery@e118b029)) ### Documentation * remove experimental annotations from GA features (#2303) ([1f1f9d41](googleapis/python-bigquery@1f1f9d41)) </details> Co-authored-by: Daniel Sanche <d.sanche14@gmail.com>

Fematichand others added 5 commits September 17, 2016 20:20

First basic implementation of Speech API
4920223

unit_test + pep8 + start documentation
dce33a7

Added Speech API to index of docs
bdf141a

Fully tested basic implementation of Speech API
2a4a6b8

Delete build.zip
a542b6d

googlebot added the cla: yes This human has signed the Contributor License Agreement. label Sep 18, 2016

tseaver added the api: speech Issues related to the Speech-to-Text API. label Sep 18, 2016

tseaver suggested changes Sep 18, 2016
View reviewed changes

Fematich added 2 commits September 19, 2016 01:24

sync_recognize rename, drop source_uri support, update unit_test and …
a5f9c0b
…partly updated docs

resolve newlines at end of files (commit from Windows vs. Linux)
c64e5ff

tseaver reviewed Sep 18, 2016
View reviewed changes

daspecster reviewed Sep 19, 2016
View reviewed changes

dhermes suggested changes Sep 19, 2016
View reviewed changes

included source_uri again and updated formatting/unit_tests
988b7d6

dhermes reviewed Sep 20, 2016
View reviewed changes

Comment threadgoogle/cloud/speech/client.py

:type source_uri: str
:param source_uri: URI that points to a file that contains audio
data bytes as specified in RecognitionConfig.
This comment was marked as spam.
Sign in to view