feat(deepgram): missing STT options

[sachin4840]

Deepgram has lots of feature which are helpful for speech to text

https://developers.deepgram.com/docs/stt-streaming-feature-overview

Below feature are missing in Deepgram STT plugin

Dictation : https://developers.deepgram.com/docs/dictation
Numerals : https://developers.deepgram.com/docs/numerals
Diarization: https://developers.deepgram.com/docs/diarization

#294

[changeset-bot]

🦋 Changeset detected

Latest commit: b67f573

The changes in this PR will be included in the next version bump.

This PR includes changesets to release 1 package

Name	Type
@livekit/agents-plugin-deepgram	Patch

Not sure what this means? Click here to learn what changesets are.

Click here if you're a maintainer who wants to add another changeset to this PR

[CLAassistant]

All committers have signed the CLA.

[arthberman]

@sachin4840 where should we access the speaker when diarize is on true ? my payload look like this
code here
payload:

{
  type: 2,
  alternatives: [
    {
      language: 'en',
      startTime: 17.37,
      endTime: 20.25,
      confidence: 0.9760742,
      text: 'Making stuff more people can participate in design process'
    }
  ]
}

thks!

[sachin4840]

@arthberman i think deepgram will recognize speaker changes. Each word in the transcript will be assigned a speaker number starting at 0

Docs: https://developers.deepgram.com/reference/speech-to-text-api/listen-streaming#request.query.diarize

[arthberman]

@sachin4840 yeah by using the deepgram sdk directly I have:

{
  transcript: 'How are you today?',
  confidence: 1,
  words: [
    {
      word: 'how',
      start: 0,
      end: 0.32,
      confidence: 1,
      speaker: 0,
      punctuated_word: 'How'
    },
    {
      word: 'are',
      start: 0.32,
      end: 0.48,
      confidence: 1,
      speaker: 0,
      punctuated_word: 'are'
    },
    {
      word: 'you',
      start: 0.48,
      end: 0.56,
      confidence: 1,
      speaker: 0,
      punctuated_word: 'you'
    },
    {
      word: 'today',
      start: 0.56,
      end: 0.79999995,
      confidence: 0.99853516,
      speaker: 0,
      punctuated_word: 'today?'
    }
  ]
}

i'm not able to get the speaker through the plugin. will open a pr on this

[sachin4840]

Oh Okay
Got the issue
will try this and update here

[arthberman]

check #331