Query regarding google::cloud::ExperimentalTag{}

[Rathi-G]

Hi,
What is the use of AsyncStreamingRecognize(google::cloud::ExperimentalTag{})?
Does is mean the code/functionality is not reliable yet?

Is it a stable code, if not when can we expect the stable version for this?

Is it safe to use it in the product development in commercial purpose?

Is there any other workaround ?

[dbolduc]

Hey, thanks for the question.

What is the use of AsyncStreamingRecognize(google:cloud:ExperimentalTag{})?

As a preface, this API exists in two places:

• speech::SpeechClient::AsyncStreamingRecognize(...)
• speech_v2::SpeechClient::AsyncStreamingRecognize(...)

The speech_v2::SpeechClient corresponds to the v2 version of the Cloud Speech-to-Text service, which is not GA. I am going to assume we are talking about the speech::SpeechClient, which is GA. (For more details on the different versions of the cloud service, see here).

Does is mean the code/functionality is not reliable yet?

The code is reliable, but it is marked as experimental because it is tricky to use. I opened #10307 to remove the experimental tag, because I think it is confusing.

Aside: The API is tricky to use, but we do have some samples to get you started: https://github.com/GoogleCloudPlatform/cpp-samples/tree/main/speech/api

Is it safe to use it in the product development in commercial purpose?

Yes, but with a caveat. At some point we will change the API signature (e.g. to remove ExperimentalTag as a parameter). You would have to make the same change in your code when you update your version of google-cloud-cpp.

There will always be some way to invoke the StreamingRecognize RPC from the C++ client library.

Is there any other workaround ?

No.

There are other RPCs that do not involve streaming. (e.g. you could use SpeechClient::Recognize to read from a file, instead). But there is only one way to use the streaming API.

[dbolduc]

Update: We dropped the ExperimentalTag from the streaming API (#10340). Hopefully this clears up the confusion. The change is included in our latest release, v2.5.0.

google-cloud-cpp/google/cloud/speech/speech_client.h

Lines 198 to 201 in 45376ec

	std::unique_ptr<::google::cloud::AsyncStreamingReadWriteRpc<
	google::cloud::speech::v1::StreamingRecognizeRequest,
	google::cloud::speech::v1::StreamingRecognizeResponse>>
	AsyncStreamingRecognize(Options opts = {});

Thanks for bringing this to our attention.

[Rathi-G]

Thank you.
Will update the code and api accordingly.