Build your personal knowledge base with Trilium Notes
TypeScript electronelectron-appknowledge-baseknowledge-graphknowledge-managementknowledge-management-graphlocal-firstnote-managmentnote-takernote-takingnotebooknotesnotes-apppersonal-knowledge-basepersonal-wikiscriptableself-hostedself-hostingwiki
import type { TranscriptionPrecision, TranscriptSource, TranscriptUtterance } from "@triliumnext/commons";
/** Every source and engine works on mono 16 kHz samples. */
export const SAMPLE_RATE = 16000;
/** Where the transcription process reports to, since it has no access to Trilium's log. */
export interface Logger {
info(message: string): void;
error(message: string): void;
}
/** Where the engine's models are. */
export interface ModelPaths {
/** Parakeet: the directory holding `tokens.txt` and the encoder, decoder and joiner. */
recognizer: { directory: string; precision: TranscriptionPrecision };
/** Silero VAD's `.onnx` file. */
vad: string;
/** The speaker embedding model's `.onnx` file, or `null` to label no speakers. */
speaker: string | null;
}
/** What the main process sends the transcription process, in this order. */
export type WorkerRequest =
| {
type: "start";
noteId: string;
models: ModelPaths;
/** Where to keep the audio and the transcript as produced, or `null` to keep neither. */
recordingDirectory: string | null;
}
/** Sent once the recording runs, with the port the capture page posts its samples to. */
| { type: "capture" }
| { type: "stop" }
/** Sent instead of `start`: loads sherpa-onnx without any model, then reports `probed`. */
| { type: "probe" };
/** What the transcription process reports to the main process. */
export type WorkerEvent =
/** The engine is loaded and the recording's start time is set; capturing can begin. */
| { type: "ready"; startedAt: number; settings: Record<string, unknown> }
| { type: "failed"; message: string }
| { type: "pending"; sources: TranscriptSource[] }
| { type: "utterance"; utterance: TranscriptUtterance }
/** The RMS of a source's latest chunk, from 0 to 1; one per chunk, about four a second. */
| { type: "level"; source: TranscriptSource; level: number }
| { type: "log"; level: keyof Logger; message: string }
/** Every utterance captured before `stop` has been sent. */
| { type: "stopped" }
/** Answers `probe`: the version of sherpa-onnx that loaded, or why it did not load. */
| { type: "probed"; version: string } | { type: "probed"; error: string };
/** One chunk of samples, posted by the capture page straight to the transcription process. */
export interface CaptureChunk {
source: TranscriptSource;
samples: Float32Array;
}
/** What the capture page posts: chunks, then once stopping, an end marker after the last one. */
export type CaptureMessage = CaptureChunk | { end: true };