pediatric-ai-scribe-v3/public/js/whisperWorker.js

// ============================================================
// WHISPER WORKER — runs @xenova/transformers in a Web Worker
// Audio never leaves the device. Model cached in IndexedDB.
// ============================================================

console.log('[WhisperWorker] Starting worker initialization');
console.log('[WhisperWorker] Loading transformers from local server...');

try {
  // Load transformers.js from our own server (no CDN dependency!)
  importScripts('/models/transformers.min.js');
  console.log('[WhisperWorker] Transformers library loaded successfully');
} catch (err) {
  console.error('[WhisperWorker] Failed to load transformers library:', err);
  self.postMessage({
    type: 'error',
    message: 'Failed to load transformers library: ' + err.message
  });
  throw err;
}

var T = self.transformers || transformers;
if (!T) {
  console.error('[WhisperWorker] Transformers object not found after import');
  throw new Error('Transformers library did not initialize');
}

// Configure transformers.js to load models from our local server
T.env.allowLocalModels = false;
T.env.useBrowserCache = true;
T.env.allowRemoteModels = false; // Force local-only
T.env.localModelPath = '/models/'; // Our bundled models
console.log('[WhisperWorker] Transformers configured for local models at /models/');

var _pipe = null;
var _loadedModel = null;

async function load(modelName) {
  if (_pipe && _loadedModel === modelName) {
    console.log('[WhisperWorker] Model already loaded:', modelName);
    return;
  }
  console.log('[WhisperWorker] Loading model:', modelName);
  self.postMessage({ type: 'loading', model: modelName });

  try {
    _pipe = await T.pipeline('automatic-speech-recognition', modelName, {
      progress_callback: function(p) {
        console.log('[WhisperWorker] Progress:', p.status, p.file, p.progress);
        if (p.status === 'downloading' || p.status === 'progress') {
          self.postMessage({ type: 'progress', file: p.file || '', progress: Math.round(p.progress || 0) });
        }
        if (p.status === 'done' || p.status === 'ready') {
          self.postMessage({ type: 'progress', file: p.file || '', progress: 100 });
        }
      }
    });
    _loadedModel = modelName;
    console.log('[WhisperWorker] Model loaded successfully');
    self.postMessage({ type: 'ready', model: modelName });
  } catch (err) {
    console.error('[WhisperWorker] Load error:', err);
    self.postMessage({ type: 'error', message: 'Model load failed: ' + err.message });
  }
}

self.addEventListener('message', async function(e) {
  var d = e.data;
  console.log('[WhisperWorker] Message received:', d.type);

  if (d.type === 'load') {
    try {
      await load(d.model || 'Xenova/whisper-tiny.en');
    } catch(err) {
      console.error('[WhisperWorker] Load error:', err);
      self.postMessage({ type: 'error', message: 'Load failed: ' + err.message });
    }
    return;
  }

  if (d.type === 'transcribe') {
    try {
      await load(d.model || 'Xenova/whisper-tiny.en');
      console.log('[WhisperWorker] Starting transcription...');
      var result = await _pipe(d.audio, {
        language: 'english',
        task: 'transcribe',
        chunk_length_s: 30,
        stride_length_s: 5,
        return_timestamps: false
      });
      console.log('[WhisperWorker] Transcription complete');
      self.postMessage({ type: 'result', text: result.text.trim() });
    } catch(err) {
      console.error('[WhisperWorker] Transcribe error:', err);
      self.postMessage({ type: 'error', message: 'Transcription failed: ' + err.message });
    }
  }
});

// Log worker startup
console.log('[WhisperWorker] Worker initialized');