content/renderer/speech_recognition_dispatcher.h

// Copyright (c) 2012 The Chromium Authors. All rights reserved.
// Use of this source code is governed by a BSD-style license that can be
// found in the LICENSE file.

#ifndef CONTENT_RENDERER_SPEECH_RECOGNITION_DISPATCHER_H_
#define CONTENT_RENDERER_SPEECH_RECOGNITION_DISPATCHER_H_

#include <map>
#include <memory>

#include "base/macros.h"
#include "base/memory/shared_memory.h"
#include "base/sync_socket.h"
#include "content/public/common/speech_recognition_result.h"
#include "content/public/renderer/render_view_observer.h"
#include "media/media_features.h"
#include "third_party/WebKit/public/platform/WebMediaStreamTrack.h"
#include "third_party/WebKit/public/platform/WebVector.h"
#include "third_party/WebKit/public/web/WebSpeechRecognitionHandle.h"
#include "third_party/WebKit/public/web/WebSpeechRecognizer.h"

namespace media {
class AudioParameters;
}

namespace content {
class RenderViewImpl;
#if BUILDFLAG(ENABLE_WEBRTC)
class SpeechRecognitionAudioSink;
#endif
struct SpeechRecognitionError;

// SpeechRecognitionDispatcher is a delegate for methods used by WebKit for
// scripted JS speech APIs. It's the complement of
// SpeechRecognitionDispatcherHost (owned by RenderViewHost).
class SpeechRecognitionDispatcher : public RenderViewObserver,
                                    public blink::WebSpeechRecognizer {
 public:
  explicit SpeechRecognitionDispatcher(RenderViewImpl* render_view);
  ~SpeechRecognitionDispatcher() override;

  // Aborts all speech recognitions.
  void AbortAllRecognitions();

 private:
  // RenderViewObserver implementation.
  bool OnMessageReceived(const IPC::Message& message) override;
  void OnDestruct() override;

  // blink::WebSpeechRecognizer implementation.
  void start(const blink::WebSpeechRecognitionHandle&,
             const blink::WebSpeechRecognitionParams&,
             blink::WebSpeechRecognizerClient*) override;
  void stop(const blink::WebSpeechRecognitionHandle&,
            blink::WebSpeechRecognizerClient*) override;
  void abort(const blink::WebSpeechRecognitionHandle&,
             blink::WebSpeechRecognizerClient*) override;

  void OnRecognitionStarted(int request_id);
  void OnAudioStarted(int request_id);
  void OnSoundStarted(int request_id);
  void OnSoundEnded(int request_id);
  void OnAudioEnded(int request_id);
  void OnErrorOccurred(int request_id, const SpeechRecognitionError& error);
  void OnRecognitionEnded(int request_id);
  void OnResultsRetrieved(int request_id,
                          const SpeechRecognitionResults& result);
  void OnAudioReceiverReady(int session_id,
                             const media::AudioParameters& params,
                             const base::SharedMemoryHandle handle,
                             const base::SyncSocket::TransitDescriptor socket);

  void ResetAudioSink();

  int GetOrCreateIDForHandle(const blink::WebSpeechRecognitionHandle& handle);
  bool HandleExists(const blink::WebSpeechRecognitionHandle& handle);
  const blink::WebSpeechRecognitionHandle& GetHandleFromID(int handle_id);

  // The WebKit client class that we use to send events back to the JS world.
  blink::WebSpeechRecognizerClient* recognizer_client_;

#if BUILDFLAG(ENABLE_WEBRTC)
  // Media stream audio track that the speech recognition connects to.
  // Accessed on the render thread.
  blink::WebMediaStreamTrack audio_track_;

  // Audio sink used to provide audio from the track.
  std::unique_ptr<SpeechRecognitionAudioSink> speech_audio_sink_;
#endif

  typedef std::map<int, blink::WebSpeechRecognitionHandle> HandleMap;
  HandleMap handle_map_;
  int next_id_;

  DISALLOW_COPY_AND_ASSIGN(SpeechRecognitionDispatcher);
};

}  // namespace content

#endif  // CONTENT_RENDERER_SPEECH_RECOGNITION_DISPATCHER_H_