Binding media stream audio track to speech recognition [browser]

content/browser/speech/speech_recognition_manager_impl.cc

 // Copyright (c) 2013 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.
 
 #include "content/browser/speech/speech_recognition_manager_impl.h"
 
 #include "base/bind.h"
+#include "base/sync_socket.h"
 #include "content/browser/browser_main_loop.h"
 #include "content/browser/renderer_host/media/media_stream_manager.h"
 #include "content/browser/renderer_host/media/media_stream_ui_proxy.h"
 #include "content/browser/speech/google_one_shot_remote_engine.h"
 #include "content/browser/speech/google_streaming_remote_engine.h"
 #include "content/browser/speech/speech_recognition_engine.h"
 #include "content/browser/speech/speech_recognizer_impl.h"
 #include "content/public/browser/browser_thread.h"
 #include "content/public/browser/content_browser_client.h"
 #include "content/public/browser/resource_context.h"
 #include "content/public/browser/speech_recognition_event_listener.h"
 #include "content/public/browser/speech_recognition_manager_delegate.h"
 #include "content/public/browser/speech_recognition_session_config.h"
 #include "content/public/browser/speech_recognition_session_context.h"
 #include "content/public/common/speech_recognition_error.h"
 #include "content/public/common/speech_recognition_result.h"
 #include "media/audio/audio_manager.h"
 #include "media/audio/audio_manager_base.h"
+#include "media/audio/audio_parameters.h"
 
 #if defined(OS_ANDROID)
 #include "content/browser/speech/speech_recognizer_impl_android.h"
 #endif
 
 using base::Callback;
 
 namespace content {
 
 SpeechRecognitionManager* SpeechRecognitionManager::manager_for_tests_;
@@ skipping 106 matching lines @@
         config.url_request_context_getter.get());
   }
 
   google_remote_engine->SetConfig(remote_engine_config);
 
   session->recognizer = new SpeechRecognizerImpl(
       this,
       session_id,
       config.continuous,
       config.interim_results,
+      config.using_audio_track,
+      config.peer_process_handle,
       google_remote_engine);
 #else
   session->recognizer = new SpeechRecognizerImplAndroid(this, session_id);
 #endif
   return session_id;
 }
 
 void SpeechRecognitionManagerImpl::StartSession(int session_id) {
   DCHECK(BrowserThread::CurrentlyOn(BrowserThread::IO));
   if (!SessionExists(session_id))
@@ skipping 23 matching lines @@
   if (!SessionExists(session_id))
     return;
 
   SessionsTable::iterator iter = sessions_.find(session_id);
   DCHECK(iter != sessions_.end());
   Session* session = iter->second;
 
   if (session->abort_requested)
     return;
 
-  if (ask_user) {
+  if (ask_user && !session->config.using_audio_track) {
     SpeechRecognitionSessionContext& context = session->context;
     context.label = media_stream_manager_->MakeMediaAccessRequest(
         context.render_process_id,
         context.render_frame_id,
         context.request_id,
         StreamOptions(true, false),
         GURL(context.context_name),
         base::Bind(
             &SpeechRecognitionManagerImpl::MediaRequestPermissionCallback,
             weak_factory_.GetWeakPtr(), session_id));
     return;
   }
 
-  if (is_allowed) {
+  if (is_allowed || session->config.using_audio_track) {
     base::MessageLoop::current()->PostTask(
         FROM_HERE,
         base::Bind(&SpeechRecognitionManagerImpl::DispatchEvent,
                    weak_factory_.GetWeakPtr(),
                    session_id,
                    EVENT_START));
   } else {
     OnRecognitionError(session_id, SpeechRecognitionError(
         SPEECH_RECOGNITION_ERROR_NOT_ALLOWED));
     base::MessageLoop::current()->PostTask(
@@ skipping 203 matching lines @@
   if (SpeechRecognitionEventListener* listener = GetListener(session_id))
     listener->OnRecognitionEnd(session_id);
   base::MessageLoop::current()->PostTask(
       FROM_HERE,
       base::Bind(&SpeechRecognitionManagerImpl::DispatchEvent,
                  weak_factory_.GetWeakPtr(),
                  session_id,
                  EVENT_RECOGNITION_ENDED));
 }
 
+void SpeechRecognitionManagerImpl::OnAudioReceiverReady(
+    int session_id,
+    const media::AudioParameters& params,
+    const base::SharedMemoryHandle memory,
+    const base::SyncSocket::TransitDescriptor socket) {
+  DCHECK(BrowserThread::CurrentlyOn(BrowserThread::IO));
+  if (!SessionExists(session_id))
+    return;
+
+  if (SpeechRecognitionEventListener* listener = GetListener(session_id))
+    listener->OnAudioReceiverReady(session_id, params, memory, socket);
+}
+
 int SpeechRecognitionManagerImpl::GetSession(
     int render_process_id, int render_view_id, int request_id) const {
   DCHECK(BrowserThread::CurrentlyOn(BrowserThread::IO));
   SessionsTable::const_iterator iter;
-  for(iter = sessions_.begin(); iter != sessions_.end(); ++iter) {
+  for (iter = sessions_.begin(); iter != sessions_.end(); ++iter) {
     const int session_id = iter->first;
     const SpeechRecognitionSessionContext& context = iter->second->context;
     if (context.render_process_id == render_process_id &&
         context.render_view_id == render_view_id &&
         context.request_id == request_id) {
       return session_id;
     }
   }
   return kSessionIDInvalid;
 }
@@ skipping 128 matching lines @@
 }
 
 // ----------- Contract for all the FSM evolution functions below -------------
 //  - Are guaranteed to be executed in the IO thread;
 //  - Are guaranteed to be not reentrant (themselves and each other);
 
 void SpeechRecognitionManagerImpl::SessionStart(const Session& session) {
   DCHECK_EQ(primary_session_id_, session.id);
   const MediaStreamDevices& devices = session.context.devices;
   std::string device_id;
-  if (devices.empty()) {
-    // From the ask_user=false path, use the default device.
-    // TODO(xians): Abort the session after we do not need to support this path
-    // anymore.
-    device_id = media::AudioManagerBase::kDefaultDeviceId;
-  } else {
-    // From the ask_user=true path, use the selected device.
-    DCHECK_EQ(1u, devices.size());
-    DCHECK_EQ(MEDIA_DEVICE_AUDIO_CAPTURE, devices.front().type);
-    device_id = devices.front().id;
+  if (!session.config.using_audio_track) {
+    if (devices.empty()) {
+      // From the ask_user=false path, use the default device.
+      // TODO(xians): Abort the session after we do not need to support this
+      // path anymore.
+      device_id = media::AudioManagerBase::kDefaultDeviceId;
+    } else {
+      // From the ask_user=true path, use the selected device.
+      DCHECK_EQ(1u, devices.size());
+      DCHECK_EQ(MEDIA_DEVICE_AUDIO_CAPTURE, devices.front().type);
+      device_id = devices.front().id;
+    }
   }
 
   session.recognizer->StartRecognition(device_id);
 }
 
 void SpeechRecognitionManagerImpl::SessionAbort(const Session& session) {
   if (primary_session_id_ == session.id)
     primary_session_id_ = kSessionIDInvalid;
   DCHECK(session.recognizer.get());
   session.recognizer->AbortRecognition();
@@ skipping 85 matching lines @@
 SpeechRecognitionManagerImpl::Session::Session()
   : id(kSessionIDInvalid),
     abort_requested(false),
     listener_is_active(true) {
 }
 
 SpeechRecognitionManagerImpl::Session::~Session() {
 }
 
 }  // namespace content