update to v1p1beta1

[ferderml]

Hello,

Great code. Quick question about updating to v1p1beta1 (or is it already there?): how does one do that?
I tried const Speech = require('@google-cloud/speech').v1p1beta1 which worked in the pure node.js solution, but it gives me an error here. Thank you!

[vin-ni]

Hi,
No, i don't think it is updated. what version is it supposed to be?
I guess you should rather do npm update ?

[ferderml]

Should be v1p1beta1 if using the beta, but at least v1. I think this is on v1beta1, but I could be wrong. I have to keep playing with it.

my-MacBook-Pro:src me$ node app.js
Server started on port:1337
Client Connected to server
/Users/ferderml/Google-Cloud-Speech-Node-Socket-Playground/src/app.js:68
        recognizeStream = speech.streamingRecognize(request)
                                 ^

TypeError: Cannot read property 'streamingRecognize' of undefined
    at startRecognitionStream (/Users/me/Google-Cloud-Speech-Node-Socket-Playground/src/app.js:68:34)
    at Socket.<anonymous> (/Users/me/Google-Cloud-Speech-Node-Socket-Playground/src/app.js:53:9)
    at emitOne (events.js:116:13)
    at Socket.emit (events.js:211:7)
    at /Users/me/Google-Cloud-Speech-Node-Socket-Playground/src/node_modules/socket.io/lib/socket.js:528:12
    at _combinedTickCallback (internal/process/next_tick.js:131:7)
    at process._tickCallback (internal/process/next_tick.js:180:9)

Is the error I get when I do:

// Google Cloud
const Speech = require('@google-cloud/speech');
const speech = Speech().v1p1beta1; // Instantiates a client

Gonna keep trying though!

[vin-ni]

Yeah, I'm just trying too :)

[vin-ni]

Ok, I just pushed an update, so it uses gcspeech v 1.5
For the beta version I think you'll somehow need to add the beta version to the package.json.
And then call

const speech = require('speech.v1p1beta1');
var speechClient = new speech.v1p1beta1.SpeechClient({
  // optional auth parameters.
});

instead of

const speech = require('@google-cloud/speech');
const speechClient = new speech.SpeechClient(); // Creates a client

What are you using the beta for?

[ferderml]

You're amazing! I want to be able to add the metadata to improve the transcription.

[vin-ni]

Oh nice!
This doesn't work on 1.5?
Write your solution here, if you found out how to use the beta :)

[ferderml]

I am still not certain it's correct, but into app.js at line 106 I added:

const recognitionMetadata = {
    interactionType: 'DICTATION',
    microphoneDistance: 'NEARFIELD',
    originalMediaType: 'AUDIO',
    recordingDeviceType: 'PC',
    audioTopic: 'Animal names', // this is just an example
  };

It runs with no errors, but doesn't work quite as ideally as wanted. Gonna try adding the dependencies from here. If you run nodejs-speech and talk in comparison to your work, it does have a difference. I am so grateful to have found yours, really saved me the headache of setting up the Socket.io, which is super helpful for a novice like me :D

[vin-ni]

how did it turn out? :)

[ferderml]

Hi! Well!! Was able to add in the SpeechContext necessary for my use case. Thank you so much!!

[vin-ni]

Did the recognitionMetadata work in the end?

[ferderml]

Yes it did!