[UWP] 實作VoiceCommand, 建立語音驅動(Voice Driven)的應用程式

前言:
微軟Window 10 的語音助理Cortana, 實現了用語音去驅動我們應用程式的情境, 開發人員可以在UWP的專案裡, 定義Voice Command, 當使用者的語音輸入符合相對應的指令,(如 Add new Movie on My Library), 程式就會被觸發, 簡單的說, 我們可以用語音把程式叫起來 , 然後讓它去做一些事情

如果想要建立一支能夠被Voice Command觸發的程式, 基本上有幾個步驟

1. 建立一個專案
2. 定義應用程式的Voice Command
3. 修改App.xaml.cs 處理程式被觸發時的動作, 複寫OnLaunched和OnActivated

(比如說做一個顯示使用者的電影資料的程式 )

建立一個新專案

在專案中建立Data Model "Movie" 表示電影資訊

	public class Movie {
	private string name;
	public string Name { get => name; set => name = value; }
	}

view raw Movie.cs hosted with ❤ by GitHub

定義Voice Command

• 這邊定義兩個command
• showMovies, 這個動作會把程式的MainPage叫起來, 然後我們會在頁面上顯示所有電影資訊
• addNewMovie, 這個動作會把語句中的電影名稱抓出來然後加到主頁面上

	<?xml version="1.0" encoding="utf-8" ?>
	<VoiceCommands xmlns="http://schemas.microsoft.com/voicecommands/1.2">
	<CommandSet xml:lang="en-us" Name="MyLibraryCommandSet_en-us">
	<AppName> my library </AppName>
	<Example> Show Movies from my library </Example>
	<Command Name="showMovies">
	<Example> Show Movies </Example>
	<ListenFor RequireAppName="BeforeOrAfterPhrase"> show [my] movies </ListenFor>
	<ListenFor RequireAppName="ExplicitlySpecified"> show {builtin:AppName} movies </ListenFor>
	<Feedback> Showing .. </Feedback>
	<Navigate />
	</Command>
	<Command Name="addNewMovie">
	<Example> Add new movie</Example>
	<ListenFor RequireAppName="BeforeOrAfterPhrase"> add {MovieName} </ListenFor>
	<ListenFor RequireAppName="ExplicitlySpecified"> add {MovieName} to {builtin:AppName} </ListenFor>
	<ListenFor RequireAppName="ExplicitlySpecified"> {builtin:AppName} add {}</ListenFor>
	<Feedback> Updating ... </Feedback>
	<Navigate />
	</Command>
	<PhraseList Label="MovieName">
	<Item>Blood Diamond</Item>
	<Item>God Father</Item>
	<Item>War Horse</Item>
	<Item>Dark Knight</Item>
	</PhraseList>
	</CommandSet>
	</VoiceCommands>

view raw MyVoiceCommand.xml hosted with ❤ by GitHub
• 基本上要做到這件事, 必須要在VoiceCommand中定義PhraseList, 並且在ListFor的語句中標示出來, 列如 add {MovieName}
• Cortana就會根據原先定義在PhraseList中的MovieName item去解析出句子中的電影名稱, 然後傳遞給我們程式

修改App.xaml.cs

• 在OnLaunched 最後面呼叫InstallCommandDefinitionsFromStorageFileAsync去安裝VoiceCommand
• 複寫OnActivated, 我們可以從SpeechRecognitionResult這個類別中知道語意識別的結果, 或是擷取語句中的{MovieName}, 如speechRecognitionResult.SemanticInterpretation.Properties["MovieName"].FirstOrDefault();

	try
	{
	var vcdStorageFile = await Package.Current.InstalledLocation.GetFileAsync(@"MyVoiceCommand.xml");
	await Windows
	.ApplicationModel
	.VoiceCommands
	.VoiceCommandDefinitionManager
	.InstallCommandDefinitionsFromStorageFileAsync(vcdStorageFile);
	}catch(Exception ex)
	{
	System.Diagnostics.Debug.WriteLine(ex.ToString());
	}

view raw App.xaml.cs hosted with ❤ by GitHub

MainPage.xaml

• 在MainPage.xaml 裡定義一個ListBox 然後綁定 MainPage.xaml.cs裡的 mylist

	try
	{
	var vcdStorageFile = await Package.Current.InstalledLocation.GetFileAsync(@"MyVoiceCommand.xml");
	await Windows
	.ApplicationModel
	.VoiceCommands
	.VoiceCommandDefinitionManager
	.InstallCommandDefinitionsFromStorageFileAsync(vcdStorageFile);
	}catch(Exception ex)
	{
	System.Diagnostics.Debug.WriteLine(ex.ToString());
	}

view raw App.xaml.cs hosted with ❤ by GitHub

MainPage.xaml.cs

• 宣告ObservableCollection<string> mylist 當作ListBox的資料來源
• 初始化mylist
• 設定DataContext 為 this
• 複寫OnNavigatedTo, 讀取傳過來的參數NavigationEventArgs e然後更新mylist

	public sealed partial class MainPage : Page
	{
	public ObservableCollection<string> mylist { get; set; }
	public MainPage()
	{
	mylist = new ObservableCollection<string>();
	mylist.Add("Blood Diamond");
	mylist.Add("Superman");
	this.DataContext = this;
	this.InitializeComponent();
	}
	protected override void OnNavigatedTo(NavigationEventArgs e)
	{
	if (e.Parameter is string)
	{
	if (this.mylist!= null)
	this.mylist.Add(e.Parameter.ToString());
	}
	base.OnNavigatedTo(e);
	}
	}

view raw MainPage.xaml.cs hosted with ❤ by GitHub

補充

• 如果要支援多國語言的話, 其實只要在CommandSet中去指定xml:lang的語系, Cortana就會知道要取讀哪個語言的指令

	<CommandSet xml:lang="en-us" Name="MyLibraryCommandSet_en-us">
	...
	</CommandSet >
	<CommandSet xml:lang="zh-cn" Name="MyLibraryCommandSet_zh-cn">
	...
	</CommandSet >
	<CommandSet xml:lang="de-de" Name="MyLibraryCommandSet_de-de">
	...
	</CommandSet >

view raw MyVoiceCommand.xml hosted with ❤ by GitHub

REF:
https://docs.microsoft.com/en-us/uwp/schemas/voicecommands/voice-command-elements-and-attributes-1-2