Function applyCumulativeUtteranceTranscript

applyCumulativeUtteranceTranscript(
    input: {
        kind: "partial" | "final";
        text: string;
        finalizedPrefix: string;
        sourceLanguage: string;
    },
): { finalizedPrefix: string; events: StreamingAsrEvent[] }
Soft-splits a growing cumulative transcript into caption-sized hub events.

Used by providers that stream one long turn (Modulate, AssemblyAI format_turns) without usable mid-turn endpointing. Finalizes sentence-sized chunks from the growing text and flushes any remainder on the provider final.
Parameters
- input: {
      kind: "partial" | "final";
      text: string;
      finalizedPrefix: string;
      sourceLanguage: string;
  }
  Latest frame kind/text plus text already emitted as finals.
Returns { finalizedPrefix: string; events: StreamingAsrEvent[] }
Updated finalized prefix and zero or more hub events (in order).
- Defined in lib/translation/streaming-asr/utterance-split.ts:730