speech/io.github.maniramezan.kommon.speech/AlignedToken

AlignedToken

data class AlignedToken(val reference: String?, val heard: String?, val confidence: Float?, val status: AlignedTokenStatus)

One position in a TranscriptAlignment: an expected word, a heard word, or both.

Constructors

constructor(reference: String?, heard: String?, confidence: Float?, status: AlignedTokenStatus)

Properties

val confidence: Float?

The recognizer's confidence in heard, when it reported one.

val heard: String?

The heard word as the recognizer wrote it; null for an AlignedTokenStatus.Omission.

val reference: String?

The expected word as written in the reference text; null for an AlignedTokenStatus.Insertion.

val status: AlignedTokenStatus

How the expected and heard words compare.