getResponseAsync method

Implementation

@override
Stream<String> getResponseAsync() {
  _controller = StreamController<String>();

  final String fullPrompt = _queryChunks.join("");

  llmInference.generateResponse(
    fullPrompt.toJS,
    ((JSString partialJs, JSAny completeRaw) {
      final complete = completeRaw.parseBool();
      final partial = partialJs.toDart;

      _controller?.add(partial);
      if (complete) {
        // Don't add response back to queryChunks - that's handled by InferenceChat
        _controller?.close();
        _controller = null;
      }
    }).toJS,
  );

  return _controller!.stream;
}