Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Cloud Speech veneer should not throw an exception if there is more than one result. #2947

Closed
pbogle opened this issue Jan 18, 2017 · 9 comments
Assignees
Labels
api: speech Issues related to the Speech-to-Text API. release blocking Required feature/issue must be fixed prior to next release.

Comments

@pbogle
Copy link

pbogle commented Jan 18, 2017

I believe this code in the veneer is incorrect because it throws if there is more than one result.

https://github.com/GoogleCloudPlatform/google-cloud-python/blob/master/speech/google/cloud/speech/_gax.py#L263

As documented in the proto, the results contain a sequential list of transcription results corresponding to sequential portions of audio.

@daspecster
Copy link
Contributor

Thanks @pbogle!

For sync_recognize I was unable to produce any tests that had more than one result.

I was able to confirm tests with multiple results with the other methods. Do you have an example I could try?

@daspecster daspecster added the api: speech Issues related to the Speech-to-Text API. label Jan 18, 2017
@pbogle
Copy link
Author

pbogle commented Jan 18, 2017

Thank you for your help. This will be a beta blocker.

In general, please check with a cloud speech team member before making any assumptions that aren't explicitly documented in the API. (Feel free to contact me, I'm tech lead for the team.) Even if we didn't return multiple results today, we might start doing so tomorrow if allowed by the specification.

An example of a file that results in multiple results is this one gs://pbogle-speech-test/calleng.raw (8khz en-US LINEAR16)

Could you help me identify other places where the veener simplifies or flattens results in ways that aren't explicitly guaranteed to by the spec? We should look at those as potential sources of bugs.

@daspecster
Copy link
Contributor

Awesome! Thanks!

I'll see if I can find other places that flatten out the data. There may be one or two.

Do you happen to know who your Vision API counterpart would be?

@pbogle
Copy link
Author

pbogle commented Jan 18, 2017

@daspecster daspecster self-assigned this Jan 20, 2017
@danoscarmike danoscarmike added the release blocking Required feature/issue must be fixed prior to next release. label Jan 25, 2017
@danoscarmike danoscarmike added this to the Beta Code Complete milestone Jan 25, 2017
@lukesneeringer
Copy link
Contributor

@daspecster I am unable to change assignment on this issue, but am currently handling it.

@daspecster daspecster removed their assignment Jan 25, 2017
@daspecster
Copy link
Contributor

Thanks @lukesneeringer! I cleared the assignees but I can't assign you. Must be a permissions thing for your account?

@lukesneeringer
Copy link
Contributor

PR is #2962, waiting on CI and approval.

@daspecster
Copy link
Contributor

This is actually pretty broken now.

The system tests exhibit the issues.
I'm working on a PR to resolve now.

@daspecster daspecster reopened this Jan 26, 2017
@lukesneeringer
Copy link
Contributor

This has since been dealt with by other PRs; re-closing. :-)

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
api: speech Issues related to the Speech-to-Text API. release blocking Required feature/issue must be fixed prior to next release.
Projects
None yet
Development

No branches or pull requests

4 participants