Re-organized functions and added semantic embeddings

src/audio.ts

@@ -0,0 +1,63 @@
+import {spawn} from 'node:child_process';
+import fs from 'node:fs/promises';
+import Path from 'node:path';
+import {Ai} from './ai.ts';
+
+export class Audio {
+	private downloads: {[key: string]: Promise<string>} = {};
+	private whisperModel!: string;
+
+	constructor(private ai: Ai) {
+		if(ai.options.whisper?.binary) {
+			this.whisperModel = ai.options.whisper?.model.endsWith('.bin') ? ai.options.whisper?.model : ai.options.whisper?.model + '.bin';
+			this.downloadAsrModel();
+		}
+	}
+
+	/**
+	 * Convert audio to text using Auditory Speech Recognition
+	 * @param {string} path Path to audio
+	 * @param model Whisper model
+	 * @returns {Promise<any>} Extracted text
+	 */
+	asr(path: string, model: string = this.whisperModel): {abort: () => void, response: Promise<string | null>} {
+		if(!this.ai.options.whisper?.binary) throw new Error('Whisper not configured');
+		let abort: any = () => {};
+		const response = new Promise<string | null>((resolve, reject) => {
+			this.downloadAsrModel(model).then(m => {
+				let output = '';
+				const proc = spawn(<string>this.ai.options.whisper?.binary, ['-nt', '-np', '-m', m, '-f', path], {stdio: ['ignore', 'pipe', 'ignore']});
+				abort = () => proc.kill('SIGTERM');
+				proc.on('error', (err: Error) => reject(err));
+				proc.stdout.on('data', (data: Buffer) => output += data.toString());
+				proc.on('close', (code: number) => {
+					if(code === 0) resolve(output.trim() || null);
+					else reject(new Error(`Exit code ${code}`));
+				});
+			});
+		});
+		return {response, abort};
+	}
+
+	/**
+	 * Downloads the specified Whisper model if it is not already present locally.
+	 *
+	 * @param {string} model Whisper model that will be downloaded
+	 * @return {Promise<string>} Absolute path to model file, resolves once downloaded
+	 */
+	async downloadAsrModel(model: string = this.whisperModel): Promise<string> {
+		if(!this.ai.options.whisper?.binary) throw new Error('Whisper not configured');
+		if(!model.endsWith('.bin')) model += '.bin';
+		const p = Path.join(this.ai.options.whisper.path, model);
+		if(await fs.stat(p).then(() => true).catch(() => false)) return p;
+		if(!!this.downloads[model]) return this.downloads[model];
+		this.downloads[model] = fetch(`https://huggingface.co/ggerganov/whisper.cpp/resolve/main/${model}`)
+			.then(resp => resp.arrayBuffer())
+			.then(arr => Buffer.from(arr)).then(async buffer => {
+				await fs.writeFile(p, buffer);
+				delete this.downloads[model];
+				return p;
+			});
+		return this.downloads[model];
+	}
+}