Reduces a transcript to a comparison form: lowercased, with every run of non-alphanumeric
characters (punctuation, whitespace) collapsed to a single space and the ends trimmed.
This is what makes continuation detection survive an ASR re-punctuating mid-sentence.
Reduces a transcript to a comparison form: lowercased, with every run of non-alphanumeric characters (punctuation, whitespace) collapsed to a single space and the ends trimmed.
This is what makes continuation detection survive an ASR re-punctuating mid-sentence.