whisper-onnx-speech-to-text/dist/index.js

Node.js plugin for speech recognition that works with OpenAI's Whisper models using ONNX.

var __awaiter = (this && this.__awaiter) || function (thisArg, _arguments, P, generator) {
    function adopt(value) { return value instanceof P ? value : new P(function (resolve) { resolve(value); }); }
    return new (P || (P = Promise))(function (resolve, reject) {
        function fulfilled(value) { try { step(generator.next(value)); } catch (e) { reject(e); } }
        function rejected(value) { try { step(generator["throw"](value)); } catch (e) { reject(e); } }
        function step(result) { result.done ? resolve(result.value) : adopt(result.value).then(fulfilled, rejected); }
        step((generator = generator.apply(thisArg, _arguments || [])).next());
    });
};
import wavefile from 'wavefile';
import fs from 'fs';
import path from 'path';
import util from 'util';
import { pipeline, env } from '@xenova/transformers';
import { DEFAULT_MODEL, MODELS_LIST, NODE_MODULES_MODELS_PATH } from './constants.js';
env.local_files_only = true;
env.localModelPath = NODE_MODULES_MODELS_PATH;
env.backends.onnx.wasm.numThreads = 1;
const readFile = util.promisify(fs.readFile);
const modelPromise = (modelName) => {
    return new Promise((resolve, reject) => __awaiter(void 0, void 0, void 0, function* () {
        try {
            if (!MODELS_LIST[modelName])
                throw `[whisper-onnx-speech-to-text] modelName "${modelName}" not found in list of models.\n`;
            if (!fs.existsSync(`${NODE_MODULES_MODELS_PATH}/${MODELS_LIST[modelName]}`))
                throw `[whisper-onnx-speech-to-text] '${modelName}' not downloaded! Run 'npx whisper-onnx-speech-to-text download'\n`;
            resolve(yield pipeline("automatic-speech-recognition", MODELS_LIST[modelName], { quantized: false }));
        }
        catch (err) {
            reject(err);
        }
    }));
};
const prepareAudio = (filePath) => __awaiter(void 0, void 0, void 0, function* () {
    const wav = new wavefile.WaveFile(yield readFile(path.normalize(filePath)));
    wav.toBitDepth('32f');
    wav.toSampleRate(16000);
    let audioData = wav.getSamples();
    if (Array.isArray(audioData)) {
        if (audioData.length > 1) {
            const SCALING_FACTOR = Math.sqrt(2);
            for (let i = 0; i < audioData[0].length; ++i) {
                audioData[0][i] = SCALING_FACTOR * (audioData[0][i] + audioData[1][i]) / 2;
            }
        }
        audioData = audioData[0];
    }
    return audioData;
});
class Whisper {
    constructor(model) {
        this.model = model;
    }
    transcribe(filePath, language) {
        return __awaiter(this, void 0, void 0, function* () {
            try {
                const audioData = yield prepareAudio(filePath);
                const lang = language ? { language } : {};
                return this.model(audioData, Object.assign({ chunk_length_s: 30, stride_length_s: 5, return_timestamps: true }, lang));
            }
            catch (error) {
                console.log("[whisper-onnx-speech-to-text] Problem:", error);
            }
        });
    }
    disposeModel() {
        return __awaiter(this, void 0, void 0, function* () {
            return this.model.dispose();
        });
    }
}
export const initWhisper = (modelName) => __awaiter(void 0, void 0, void 0, function* () {
    if (!modelName)
        console.log("[whisper-onnx-speech-to-text] No 'modelName' provided. Trying default model:", DEFAULT_MODEL, "\n");
    const model = yield modelPromise(modelName || DEFAULT_MODEL);
    return new Whisper(model);
});
//# sourceMappingURL=index.js.map