首页 > 代码库 > 如何将 Cortana 与 Windows Phone 8.1 应用集成 ( Voice command - Natural language recognition )

如何将 Cortana 与 Windows Phone 8.1 应用集成 ( Voice command - Natural language recognition )

随着 Windows Phone 8.1 GDR1 + Cortana 中文版的发布,相信有很多用户或开发者都在调戏 Windows Phone 的语音私人助理 Cortana 吧,在世界杯的时候我亲测 Cortana 预测德国和阿根廷的比赛很准的。(题外话扯远了),可是作为开发者我们怎么将Cortana集成到应用中呢,今天我用一点时间给大家介绍一下如何使用 voice command 集成 Windows Phone 8.1 的应用。

首先要明确两个名词 Voice command & Voice Command Definition 即 VCD文件,相信做过windows Phone 8.0 开发的朋友应该有所了解,通过注册VCD文件 Windows phone 8.0 的应用当中就可以实现 voice command 的功能,如果你不了解请先读一下我之前的文章(这里我就不在过多介绍 8.0 Voice command 的重复内容了),Windows Phone 8 语音 - Speech for Windows Phone 8 快速了解一下Windows Phone 开发语音功能的前期准备工作。

 

简单的说在 Windows Phone 8.0 voice command 功能比较简单,主要是通过 Voice Command Name 判断预制在VCD文件中的几个命令。

在 Windows Phone 8.1 应用中 Cortana 提供了更强的自然语言识别(Natural language recognition)

Cortana1

当然 VCD 文件的中的 grammars  也得到了扩充,并且区别两个OS版本的

http://schemas.microsoft.com/voicecommands/1.0 for Windows Phone 8.0 Voice Command and Cortana compatible.

http://schemas.microsoft.com/voicecommands/1.1 only for Widnows Phone 8.1 Cortnan.

Cortana2

详细内容请参考

Windows Phone 8.0:  Voice command element and attribute reference for Windows Phone 8

Windows Phone 8.1:  Voice command elements and attributes

通过 8.0 和 8.1 VCD 文件属性支持情况来看有一个最主要的区别在8.1 VCD中支持 PhraseTopic 这个属性。

image

文字说的太抽象了还是贴出代码给大家说说吧:

这里我主要强调说一下 ListenFor  结点和 PhraseTopic 结点。 注意在 Listenfor 结点中的中括号 {dictatedSearchTerms} 是对应的 PhraseTopic  结点中的 Label 属性。同时我们可以把 PhraseTopic 理解成任意内容。最后都可以从Cortana回传到我们的应用当中来。

<VoiceCommands xmlns="http://schemas.microsoft.com/voicecommands/1.1">    <!-- The CommandSet Name is used to programmatically access the CommandSet -->    <CommandSet xml:lang="zh-CN" Name="chineseCommands">        <!-- The CommandPrefix provides an alternative to your full app name for invocation -->        <CommandPrefix> 微软 文档 </CommandPrefix>        <!-- The CommandSet Example appears in the global help alongside your app name -->        <Example> 搜索 构造 函数 </Example>        <Command Name="MSDNSearch">            <!-- The Command example appears in the drill-down help page for your app -->            <Example> 搜索 构造 函数‘ </Example>            <!-- ListenFor elements provide ways to say the command, including references to             {PhraseLists} and {PhraseTopics} as well as [optional] words -->            <ListenFor> 查找 {dictatedSearchTerms} </ListenFor>            <ListenFor> 搜 {dictatedSearchTerms} </ListenFor>            <ListenFor> 搜索  {dictatedSearchTerms} </ListenFor>            <ListenFor> 查 {dictatedSearchTerms} </ListenFor>            <ListenFor> 找 {dictatedSearchTerms} </ListenFor>          <!--Feedback provides the displayed and spoken text when your command is triggered -->            <Feedback> 查找 MSDN... </Feedback>            <!-- Navigate specifies the desired page or invocation destination for the Command-->            <Navigate Target="MainPage.xaml" />        </Command>        <Command Name="MSDNNaturalLanguage">            <Example> 我 想 去 Windows 手机 开发 中心 </Example>            <ListenFor> {naturalLanguage} </ListenFor>            <Feedback> 启动 MSDN... </Feedback>            <Navigate Target="MainPage.xaml" />        </Command>        <PhraseTopic Label="dictatedSearchTerms" Scenario="Search">            <Subject> MSDN </Subject>        </PhraseTopic>        <PhraseTopic Label="naturalLanguage" Scenario="Natural Language">            <Subject> MSDN </Subject>        </PhraseTopic>    </CommandSet></VoiceCommands>

 

了解完新的VCD文件,在这里我提醒下大家,其实在Windows Phone 8.0的应用中也可以兼容 Cortana的功能的,在8.0的应用当中我们只需要判断一下操作系统的版本然后选择不同的VCD文件向系统注册即可。

首先我们需要把两个版本的VCD文件都存放到项目中来

image

其次在注册VCD文件的时候根据系统版本进行一下判断即可。

        /// <summary>        /// Installs the Voice Command Definition (VCD) file associated with the application.        /// Based on OS version, installs a separate document based on version 1.0 of the schema or version 1.1.        /// </summary>        private async void InstallVoiceCommands()        {            const string wp80vcdPath = "ms-appx:///VoiceCommandDefinition_8.0.xml";            const string wp81vcdPath = "ms-appx:///VoiceCommandDefinition_8.1.xml";            const string chineseWp80vcdPath = "ms-appx:///ChineseVoiceCommandDefinition_8.0.xml";            const string chineseWp81vcdPath = "ms-appx:///ChineseVoiceCommandDefinition_8.1.xml";            try            {                bool using81orAbove = ((Environment.OSVersion.Version.Major >= 8)                    && (Environment.OSVersion.Version.Minor >= 10));                string vcdPath = using81orAbove ? wp81vcdPath : wp80vcdPath;                if (InstalledSpeechRecognizers.Default.Language.Equals("zh-CN", StringComparison.InvariantCultureIgnoreCase))                {                    vcdPath = using81orAbove ? chineseWp81vcdPath : chineseWp80vcdPath;                }                Uri vcdUri = new Uri(vcdPath);                await VoiceCommandService.InstallCommandSetsFromFileAsync(vcdUri);            }            catch (Exception vcdEx)            {                Dispatcher.BeginInvoke(() =>                {                    MessageBox.Show(String.Format(                        AppResources.VoiceCommandInstallErrorTemplate, vcdEx.HResult, vcdEx.Message));                });            }        }

最后在应用当中获取用户的语音输入方法,注意这里也是需要通过 PhraseTopic 结点的 Label 名称获取的。

        /// <summary>        /// Takes specific action for a retrieved VoiceCommand name.        /// </summary>        /// <param name="voiceCommandName"> the command name triggered to activate the application </param>        private void HandleVoiceCommand(string voiceCommandName)        {            // Voice Commands can be typed into Cortana; when this happens, "voiceCommandMode" is populated with the            // "textInput" value. In these cases, we‘ll want to behave a little differently by not speaking back.            bool typedVoiceCommand = (NavigationContext.QueryString.ContainsKey("commandMode")                 && (NavigationContext.QueryString["commandMode"] == "text"));            string phraseTopicContents = null;            bool doSearch = false;            switch (voiceCommandName)            {                case "MSDNNaturalLanguage":                    if (NavigationContext.QueryString.TryGetValue("naturalLanguage", out phraseTopicContents)                        && !String.IsNullOrEmpty(phraseTopicContents))                    {                        // We‘ll try to process the input as a natural language query; if we‘re successful, we won‘t                        // fall back into searching, since the query will have already been handled.                        doSearch = TryHandleNlQuery(phraseTopicContents, typedVoiceCommand);                    }                    break;                case "MSDNSearch":                    // The user explicitly asked to search, so we‘ll attempt to retrieve the query.                    NavigationContext.QueryString.TryGetValue("dictatedSearchTerms", out phraseTopicContents);                    doSearch = true;                    break;            }            if (doSearch)            {                HandleSearchQuery(phraseTopicContents, typedVoiceCommand);            }        }

image

整个过程就这么简单,心动不如行动,赶快把你的应用加入Cortana 功能让小伙伴儿们调戏一番。

更多参考资料:

Quickstart: Voice commands (XAML)

Speech for Windows Phone 8

快速入门:语音命令 (XAML)

源码下载:

MSDN Voice Search for Windows Phone 8.1