/// <summary>
/// Constructor of Speech page
/// </summary>
public SpeechSettingsPage()
{
_oSelectedSpeechCommand = null;
InitializeComponent();
App.ShowToast(AppResources.Speech_register);
swEnableSpeech.IsToggled = App.AppSettings.SpeechEnabled;
swEnableSpeech.Toggled += async(sender, args) =>
{
App.AppSettings.SpeechEnabled = swEnableSpeech.IsToggled;
if (swEnableSpeech.IsToggled)
{
if (!await ValidateSpeechRecognition())
{
swEnableSpeech.IsToggled = false;
}
}
};
_oListSource = App.AppSettings.SpeechCommands;
if (_oListSource != null)
{
listView.ItemsSource = _oListSource;
}
}
public IActionResult SpeectTextFromRequest(string text)
{
SpeechModel model = new SpeechModel();
model.Text = text;
return(SpeectTextFromPost(model));
}
public string ConvertTextFromRequest(string text)
{
SpeechModel model = new SpeechModel();
model.Text = text;
return(ConvertTextFromPost(model));
}
/// <summary>
/// Connect device to Speech Command
/// </summary>
private async void btnConnect_Clicked(object sender, EventArgs e)
{
_oSelectedSpeechCommand = (SpeechModel)((TintedCachedImage)sender).BindingContext;
var oSwitchPopup = new SwitchPopup();
oSwitchPopup.DeviceSelectedMethod += DelegateMethod;
await PopupNavigation.Instance.PushAsync(oSwitchPopup);
}
public string ConvertTextFromRequest(string text)
{
var model = new SpeechModel
{
Text = text
};
return(ConvertTextFromPost(model));
}
public IActionResult SpeectTextFromRequest(string text)
{
var model = new SpeechModel
{
Text = text
};
return(SpeectTextFromPost(model));
}
public IActionResult SpeectTextFromPost([FromBody] SpeechModel speech_model)
{
Setting.Lock();
try
{
if (speech_model.SpeakerSetting != null &&
speech_model.SpeakerSetting.VoiceDbName.Length > 0 &&
speech_model.SpeakerSetting.SpeakerName.Length > 0 &&
AitalkWrapper.Parameter.CurrentSpeakerName != speech_model.SpeakerSetting.SpeakerName)
{
var error_message = Setting.ApplySpeakerSetting(speech_model.SpeakerSetting);
if (error_message != null)
{
return(BadRequest($"Saved but {error_message}"));
}
}
// 話者パラメータを設定する
var speaker = speech_model.Speaker ?? new SpeakerModel();
AitalkWrapper.Parameter.VoiceVolume = (0 <= speaker.Volume) ? speaker.Volume : Setting.DefaultSpeakerParameter.Volume;
AitalkWrapper.Parameter.VoiceSpeed = (0 <= speaker.Speed) ? speaker.Speed : Setting.DefaultSpeakerParameter.Speed;
AitalkWrapper.Parameter.VoicePitch = (0 <= speaker.Pitch) ? speaker.Pitch : Setting.DefaultSpeakerParameter.Pitch;
AitalkWrapper.Parameter.VoiceEmphasis = (0 <= speaker.Emphasis) ? speaker.Emphasis : Setting.DefaultSpeakerParameter.Emphasis;
AitalkWrapper.Parameter.PauseMiddle = (0 <= speaker.PauseMiddle) ? speaker.PauseMiddle : Setting.DefaultSpeakerParameter.PauseMiddle;
AitalkWrapper.Parameter.PauseLong = (0 <= speaker.PauseLong) ? speaker.PauseLong : Setting.DefaultSpeakerParameter.PauseLong;
AitalkWrapper.Parameter.PauseSentence = (0 <= speaker.PauseSentence) ? speaker.PauseSentence : Setting.DefaultSpeakerParameter.PauseSentence;
// テキストが与えられた場合は仮名に変換する
string kana = null;
if ((speech_model.Kana != null) && (0 < speech_model.Kana.Length))
{
kana = speech_model.Kana;
}
else if ((speech_model.Text != null) && (0 < speech_model.Text.Length))
{
kana = AitalkWrapper.TextToKana(speech_model.Text, Setting.System.KanaTimeout);
}
if ((kana == null) || (kana.Length <= 0))
{
return(new NoContentResult());
}
// 音声変換して結果を返す
var wave_stream = new MemoryStream();
AitalkWrapper.KanaToSpeech(kana, wave_stream, Setting.System.SpeechTimeout);
return(new FileContentResult(wave_stream.ToArray(), "audio/wav"));
}
catch (Exception)
{
return(new NoContentResult());
}
finally
{
Setting.Unlock();
}
}
public async Task <JsonResult> GetResponseGladysAsync(SpeechModel speech)
{
var reponse = await _client.GetAsync(_urLLuis + speech.Text);
var result = await reponse.Content.ReadAsStringAsync();
var resultObject = JsonConvert.DeserializeObject <WebServices.Models.LuisResult>(result);
return(new JsonResult());
}
private SpeechModel[] ParseGetModelsResponse(byte[] data)
{
string jsonString = Encoding.UTF8.GetString(data);
if (jsonString == null)
{
Log.Error("SpeechToText", "Failed to get JSON string from response.");
return(null);
}
IDictionary json = (IDictionary)Json.Deserialize(jsonString);
if (json == null)
{
Log.Error("SpechToText", "Failed to parse JSON: {0}", jsonString);
return(null);
}
try
{
List <SpeechModel> models = new List <SpeechModel>();
IList imodels = json["models"] as IList;
if (imodels == null)
{
throw new Exception("Expected IList");
}
foreach (var m in imodels)
{
IDictionary imodel = m as IDictionary;
if (imodel == null)
{
throw new Exception("Expected IDictionary");
}
SpeechModel model = new SpeechModel();
model.Name = (string)imodel["name"];
model.Rate = (long)imodel["rate"];
model.Language = (string)imodel["language"];
model.Description = (string)imodel["description"];
model.URL = (string)imodel["url"];
models.Add(model);
}
return(models.ToArray());
}
catch (Exception e)
{
Log.Error("SpeechToText", "Caught exception {0} when parsing GetModels() response: {1}", e.ToString(), jsonString);
}
return(null);
}
public ActionResult Index(SpeechModel speechModel)
{
ViewBag.Key = speechModel.SubscriptionKey;
Authentication obj = new Authentication(ViewBag.Key);
ViewBag.Content = speechModel.Content;
ViewBag.LangCode = speechModel.LanguageCode;
ViewBag.Token = obj.GetAccessToken();
return(View(speechModel));
}
public static LanguageModel CreateLanguageModel(SpeechModel model, byte[] languageFile)
{
var languageModel = new LanguageModel()
{
name = model.name,
description = model.description,
locale = model.locale,
dataImportKind = "Language",
properties = model.properties,
languagedata = languageFile
};
return(languageModel);
}
/// <summary>
/// Create new Speech object
/// </summary>
private void AddNewRecord(string speechID, string speechText)
{
App.ShowToast(AppResources.Speech_saved + " " + speechText);
var speechObject = new SpeechModel()
{
Id = speechID,
Name = speechText,
Enabled = true,
};
_oListSource.Add(speechObject);
SaveAndRefresh();
App.ShowToast(AppResources.noSwitchSelected_explanation_Speech);
}
public static AcousticModel CreateAcousticModel(SpeechModel model, byte[] zipFile, byte[] transcriptionFile)
{
var acousticModel = new AcousticModel()
{
name = model.name,
description = model.description,
locale = model.locale,
dataImportKind = "Acoustic",
properties = model.properties,
audiodata = zipFile,
transcriptions = transcriptionFile
};
return(acousticModel);
}
public ResultModel <SynthesizeSpeechResponse> DownloadWord(SpeechModel model)
{
if (!File.Exists("wwwroot/assets/speeches/"))
{
Directory.CreateDirectory("wwwroot/assets/speeches/");
}
Environment.SetEnvironmentVariable("GOOGLE_APPLICATION_CREDENTIALS", "./speech-key.json");
try
{
var client = TextToSpeechClient.Create();
// The input to be synthesized, can be provided as text or SSML.
var input = new SynthesisInput
{
Text = model.Text
};
// Build the voice request.
var voiceSelection = new VoiceSelectionParams
{
LanguageCode = model.LanguageCode,
SsmlGender = model.Gender
};
// Specify the type of audio file.
var audioConfig = new AudioConfig
{
AudioEncoding = AudioEncoding.Mp3
};
// Perform the text-to-speech request.
return(new ResultModel <SynthesizeSpeechResponse> {
Data = client.SynthesizeSpeech(input, voiceSelection, audioConfig), Success = true
});
}
catch (Exception ex)
{
notFoundLanguageCodes.Add(new NotFoundLanguageCode
{
Code = model.LanguageCode,
ErrorMessage = ex.Message
});
return(new ResultModel <SynthesizeSpeechResponse> {
Success = false, ErrorMessage = ex.Message
});
}
}
public SpeechModel GetVoicFromText(SpeechModel model)
{
model.Code = Guid.NewGuid();
try
{
var response = new ResultModel <SynthesizeSpeechResponse>();
if (notFoundLanguageCodes.Any(x => x.Code == model.LanguageCode))
{
response = new ResultModel <SynthesizeSpeechResponse>
{
Success = false,
ErrorMessage = notFoundLanguageCodes.Find(x => x.Code == model.LanguageCode).ErrorMessage
};
}
else
{
response = DownloadWord(model);
}
if (response.Success)
{
// Write the response to the output file.
using FileStream output = File.Create("wwwroot/assets/speeches/" + model.Code + ".mp3");
response.Data.AudioContent.WriteTo(output);
model.Status = SpeechStatus.Success;
}
else
{
model.ErrorMessage = response.ErrorMessage;
model.Status = SpeechStatus.Error;
}
}
catch (Exception ex)
{
model.Status = SpeechStatus.Error;
model.ErrorMessage = ex.Message;
}
model.AddedDate = DateTime.Now;
model.Id = db.Query <int>(@"
insert into [dbo].[speech] (Code, AddedDate, Status, ErrorMessage)
VALUES (@Code, @AddedDate, @Status, @ErrorMessage);
SELECT CAST(SCOPE_IDENTITY() as int)
", model).SingleOrDefault();
return(model);
}
public string ConvertTextFromPost([FromBody] SpeechModel speech_model)
{
Setting.Lock();
try
{
if ((speech_model.Text == null) || (speech_model.Text.Length <= 0))
{
return(null);
}
return(AitalkWrapper.TextToKana(speech_model.Text, Setting.System.KanaTimeout));
}
catch (Exception)
{
return(null);
}
finally
{
Setting.Unlock();
}
}
private void OnGetModels(SpeechModel [] models)
{
if (models != null)
{
SpeechModel bestModel = null;
foreach (var model in models)
{
if (model.Language.StartsWith(m_Language) &&
(bestModel == null || model.Rate > bestModel.Rate))
{
bestModel = model;
}
}
if (bestModel != null)
{
Log.Status("SpeechToTextWidget", "Selecting Recognize Model: {0} ", bestModel.Name);
m_SpeechToText.RecognizeModel = bestModel.Name;
}
}
}
public ActionResult Index(SpeechModel speechModel, string ArticleUrl)
{
try
{
string accessToken;
//Extract contents from an article
tc.TrackTrace("Extract contents from article");
ArticleExtraction articleExtraction = new ArticleExtraction();
StringBuilder articleExtract = new StringBuilder();
articleExtract.Append(articleExtraction.ExtractArticle(ArticleUrl));
//Add end of article note
articleExtract.AppendLine("This article is over. Thanks for listening.");
//ErrorLogger.Debug("Article extract: " + articleExtract);
//ErrorLogger.Debug("Get access token");
tc.TrackTrace("Article extract: " + articleExtract);
tc.TrackTrace("Get access token");
//Get access token for cognitive speech API
Authentication auth = new Authentication();
accessToken = auth.GetAccessToken();
ViewBag.Content = articleExtract.ToString();
ViewBag.Token = accessToken;
ViewBag.Key = speechModel.SubscriptionKey;
ViewBag.LocaleCode = speechModel.LocaleCode;
//return View("ReadArticle", speechModel);
return(View(speechModel));
}
catch (Exception ex)
{
//return RedirectToAction("Index");
//ErrorLogger.Debug(ex.Message);
//ErrorLogger.Debug(ex.StackTrace);
//ErrorLogger.Debug(ex.InnerException.ToString());
tc.TrackException(ex);
return(View("Error"));
}
}
private static Dictionary <string, object> ModelToDictionary(SpeechModel model)
{
Dictionary <string, object> postParameters = new Dictionary <string, object>();
foreach (var propertyInfo in model.GetType().GetProperties(BindingFlags.Public | BindingFlags.Instance))
{
var param = propertyInfo.GetValue(model);
var fileName = string.Empty;
switch (propertyInfo.Name)
{
case "languagedata":
case "transcriptions":
var textFile = (byte[])param;
fileName = propertyInfo.Name;
var textFileParameter = new FileParameter(textFile, fileName, "text/plain");
postParameters.Add(propertyInfo.Name, textFileParameter);
break;
case "audiodata":
var zipFile = (byte[])param;
fileName = propertyInfo.Name;
var zipFileParameter = new FileParameter(zipFile, fileName, "application/x-zip-compressed");
postParameters.Add(propertyInfo.Name, zipFileParameter);
break;
case "properties":
var jsonProperties = JsonConvert.SerializeObject(param);
postParameters.Add(propertyInfo.Name, jsonProperties);
break;
default: // it's just a string/string key value pair
postParameters.Add(propertyInfo.Name, param);
break;
}
}
return(postParameters);
}
public SpeechModel GetModel(string modelName)
{
if (string.IsNullOrEmpty(modelName))
{
throw new ArgumentNullException("modelName can not be null or empty");
}
SpeechModel result = null;
try
{
result =
this.Client.WithAuthentication(this.UserName, this.Password)
.GetAsync($"{this.Endpoint}{PATH_MODELS}/{modelName}")
.As <SpeechModel>()
.Result;
}
catch (AggregateException ae)
{
throw ae.InnerException as ServiceResponseException;
}
return(result);
}
//ASR: Automatic Speech Recognition
public AutomaticSpeechRecognition(SpeechModel speechModel)
{
this.speechModel = speechModel;
}
public ActionResult Index()
{
SpeechModel speechModel = new SpeechModel();
return(View(speechModel));
}
/// <summary>
/// Create new speech recognition options.
/// </summary>
/// <param name="speechModel">The language model to use for recognition.</param>
/// <param name="audioFormat">The audio format.</param>
/// <param name="language">The language for speech recognition.</param>
public SpeechRecognitionOptions(SpeechModel speechModel, RecognitionAudioFormat audioFormat, RecognitionLanguage language)
{
SpeechModel = speechModel;
AudioFormat = audioFormat;
Language = language;
}
/// <summary>
/// Create new speech recognition session options.
/// </summary>
/// <param name="speechModel">The language model to use for recognition.</param>
/// <param name="audioFormat">The audio format.</param>
/// <param name="connectionMode">Network security settings.</param>
public SpeechRecognitionSessionOptions(SpeechModel speechModel, RecognitionAudioFormat audioFormat, ConnectionMode connectionMode = ConnectionMode.Secure)
{
SpeechModel = speechModel;
AudioFormat = audioFormat;
ConnectionMode = connectionMode;
}
