你当前正在访问 Microsoft Azure Global Edition 技术文档网站。 如果需要访问由世纪互联运营的 Microsoft Azure 中国技术文档网站,请访问 https://docs.azure.cn

如何使用 REST API 更新 LUIS 模型

重要

LUIS 将于 2025 年 10 月 1 日停用,从 2023 年 4 月 1 日开始,你将无法创建新的 LUIS 资源。 建议将 LUIS 应用程序迁移对话语言理解,以便从持续的产品支持和多语言功能中受益。

在本文中,你要将示例言语添加到披萨应用并训练该应用。 示例话语是映射到意向的对话式用户文本。 通过提供意向的示例话语,可以教 LUIS 识别用户提供的文本类型属于哪种意向。

参考文档 | 示例

先决条件

示例话语 JSON 文件

示例言语采用特定格式。

text 字段包含示例话语的文本。 intentName 字段必须对应于 LUIS 应用中的现有意向名称。 entityLabels 字段是必填的。 如果不想标记任何实体,请提供一个空数组。

如果 entityLabels 数组不为空,则 startCharIndexendCharIndex 需要标记 entityName 字段中引用的实体。 该索引从零开始。 如果标签的起始或结尾位于文本中的空白处,则添加话语的 API 调用将失败。

[
  {
    "text": "order a pizza",
    "intentName": "ModifyOrder",
    "entityLabels": [
      {
        "entityName": "Order",
        "startCharIndex": 6,
        "endCharIndex": 12
      }
    ]
  },
  {
    "text": "order a large pepperoni pizza",
    "intentName": "ModifyOrder",
    "entityLabels": [
      {
        "entityName": "Order",
        "startCharIndex": 6,
        "endCharIndex": 28
      },
      {
        "entityName": "FullPizzaWithModifiers",
        "startCharIndex": 6,
        "endCharIndex": 28
      },
      {
        "entityName": "PizzaType",
        "startCharIndex": 14,
        "endCharIndex": 28
      },
      {
        "entityName": "Size",
        "startCharIndex": 8,
        "endCharIndex": 12
      }
    ]
  },
  {
    "text": "I want two large pepperoni pizzas on thin crust",
    "intentName": "ModifyOrder",
    "entityLabels": [
      {
        "entityName": "Order",
        "startCharIndex": 7,
        "endCharIndex": 46
      },
      {
        "entityName": "FullPizzaWithModifiers",
        "startCharIndex": 7,
        "endCharIndex": 46
      },
      {
        "entityName": "PizzaType",
        "startCharIndex": 17,
        "endCharIndex": 32
      },
      {
        "entityName": "Size",
        "startCharIndex": 11,
        "endCharIndex": 15
      },
      {
        "entityName": "Quantity",
        "startCharIndex": 7,
        "endCharIndex": 9
      },
      {
        "entityName": "Crust",
        "startCharIndex": 37,
        "endCharIndex": 46
      }
    ]
  }
]

创建 Pizza 应用

创建披萨应用。

  1. 选择 pizza-app-for-luis-v6.json,打开 pizza-app-for-luis.json 文件的 GitHub 页面。
  2. 右键单击或长按“原始”按钮,然后选择“将链接另存为”,将 保存到计算机。
  3. 登录到 LUIS 门户
  4. 选择我的应用
  5. 在“我的应用”页面上,选择“+ 新建用于对话的应用” 。
  6. 选择“导入为 JSON”。
  7. 在“导入新应用”对话框中,选择“选择文件”按钮 。
  8. 选择下载的 pizza-app-for-luis.json 文件,然后选择“打开”。
  9. 在“导入新应用”对话框的“名称”字段中,输入 Pizza 应用的名称,然后选择“完成”按钮 。

随即导入应用。

如果看到一个对话框“如何创建有效的 LUIS 应用”,关闭该对话框。

训练并发布 Pizza 应用

Pizza 应用中应会显示“意向”页面,其中显示了一个意向列表。

  1. 在 LUIS 网站的右上方,选择“训练”按钮。

    “训练”按钮

  2. 当“训练”按钮处于禁用状态时,即表示训练完成。

若要在聊天机器人或其他客户端应用程序中接收 LUIS 预测,需要将应用发布到预测终结点。

  1. 在右上方的导航栏中选择“发布”。

    用于发布到终结点的按钮的屏幕截图。

  2. 选择“生产”槽,然后选择“完成” 。

    LUIS 发布到终结点的屏幕截图。

  3. 在通知中选择“访问终结点 URL”,以转到“Azure 资源”页 。 只有你拥有与应用关联的预测资源时,才能看到 URL。 还可以单击“管理”来找到“Azure 资源”页 。

    显示该应用已发布的消息的屏幕截图。

向披萨应用添加创作资源

  1. 选择“管理”。
  2. 选择“Azure 资源”。
  3. 选择“创作资源”。
  4. 选择“更改创作资源”。

如果你有创作资源,请输入你的创作资源的租户名称、订阅名称以及 LUIS 资源名称。

如果你没有创作资源:

  1. 选择“新建资源”。
  2. 输入租户名称、资源名称、订阅名称和 Azure 资源组名称。

现在 Pizza 应用可以使用了。

记录 Pizza 应用的访问值

若要使用新的披萨应用,你需要该披萨应用的应用 ID、授权密钥和授权终结点。 若要获取预测,需要单独的预测终结点和预测密钥。

若要查找这些值:

  1. 在“意向”页面,选择“管理” 。
  2. 在“应用程序设置”页面,记录“应用 ID” 。
  3. 选择“Azure 资源”。
  4. 选择“创作资源”。
  5. 从“创作资源”和“预测资源”选项卡,记录“主键” 。 此值是你的创作密钥。
  6. 记录“终结点 URL”。 此值是你的创作终结点。

以编程方式更改模型

  1. 使用项目和名为 csharp-model-with-rest 的文件夹创建一个面向 C# 语言的新控制台应用程序。

    dotnet new console -lang C# -n csharp-model-with-rest
    
  2. 更改为创建的 csharp-model-with-rest 目录,并使用以下命令安装所需的依赖项:

    cd csharp-model-with-rest
    dotnet add package System.Net.Http
    dotnet add package JsonFormatterPlus
    
  3. 将 Program.cs 改写为以下代码:

    //
    // This quickstart shows how to add utterances to a LUIS model using the REST APIs.
    //
    
    using System;
    using System.IO;
    using System.Net.Http;
    using System.Text;
    using System.Threading.Tasks;
    using System.Collections.Generic;
    using System.Linq;
    
    // 3rd party NuGet packages
    using JsonFormatterPlus;
    
    namespace AddUtterances
    {
        class Program
        {
            //////////
            // Values to modify.
    
            // YOUR-APP-ID: The App ID GUID found on the www.luis.ai Application Settings page.
            static string appID = "PASTE_YOUR_LUIS_APP_ID_HERE";
    
            // YOUR-AUTHORING-KEY: Your LUIS authoring key, 32 character value.
            static string authoringKey = "PASTE_YOUR_LUIS_AUTHORING_SUBSCRIPTION_KEY_HERE";
    
            // YOUR-AUTHORING-ENDPOINT: Replace this endpoint with your authoring key endpoint.
            // For example, "https://your-resource-name.cognitiveservices.azure.com/"
            static string authoringEndpoint = "PASTE_YOUR_LUIS_AUTHORING_ENDPOINT_HERE";
    
            // NOTE: Replace this your version number.
            static string appVersion = "0.1";
            //////////
    
            static string host = String.Format("{0}luis/authoring/v3.0-preview/apps/{1}/versions/{2}/", authoringEndpoint, appID, appVersion);
    
            // GET request with authentication
            async static Task<HttpResponseMessage> SendGet(string uri)
            {
                using (var client = new HttpClient())
                using (var request = new HttpRequestMessage())
                {
                    request.Method = HttpMethod.Get;
                    request.RequestUri = new Uri(uri);
                    request.Headers.Add("Ocp-Apim-Subscription-Key", authoringKey);
                    return await client.SendAsync(request);
                }
            }
    
            // POST request with authentication
            async static Task<HttpResponseMessage> SendPost(string uri, string requestBody)
            {
                using (var client = new HttpClient())
                using (var request = new HttpRequestMessage())
                {
                    request.Method = HttpMethod.Post;
                    request.RequestUri = new Uri(uri);
    
                    if (!String.IsNullOrEmpty(requestBody))
                    {
                        request.Content = new StringContent(requestBody, Encoding.UTF8, "text/json");
                    }
    
                    request.Headers.Add("Ocp-Apim-Subscription-Key", authoringKey);
                    return await client.SendAsync(request);
                }
            }
    
            // Add utterances as string with POST request
            async static Task AddUtterances(string utterances)
            {
                string uri = host + "examples";
    
                var response = await SendPost(uri, utterances);
                var result = await response.Content.ReadAsStringAsync();
                Console.WriteLine("Added utterances.");
                Console.WriteLine(JsonFormatter.Format(result));
            }
    
            // Train app after adding utterances
            async static Task Train()
            {
                string uri = host  + "train";
    
                var response = await SendPost(uri, null);
                var result = await response.Content.ReadAsStringAsync();
                Console.WriteLine("Sent training request.");
                Console.WriteLine(JsonFormatter.Format(result));
            }
    
            // Check status of training
            async static Task Status()
            {
                var response = await SendGet(host  + "train");
                var result = await response.Content.ReadAsStringAsync();
                Console.WriteLine("Requested training status.");
                Console.WriteLine(JsonFormatter.Format(result));
            }
    
            // Add utterances, train, check status
            static void Main(string[] args)
            {
                string utterances = @"
                [
                    {
                        'text': 'order a pizza',
                        'intentName': 'ModifyOrder',
                        'entityLabels': [
                            {
                                'entityName': 'Order',
                                'startCharIndex': 6,
                                'endCharIndex': 12
                            }
                        ]
                    },
                    {
                        'text': 'order a large pepperoni pizza',
                        'intentName': 'ModifyOrder',
                        'entityLabels': [
                            {
                                'entityName': 'Order',
                                'startCharIndex': 6,
                                'endCharIndex': 28
                            },
                            {
                                'entityName': 'FullPizzaWithModifiers',
                                'startCharIndex': 6,
                                'endCharIndex': 28
                            },
                            {
                                'entityName': 'PizzaType',
                                'startCharIndex': 14,
                                'endCharIndex': 28
                            },
                            {
                                'entityName': 'Size',
                                'startCharIndex': 8,
                                'endCharIndex': 12
                            }
                        ]
                    },
                    {
                        'text': 'I want two large pepperoni pizzas on thin crust',
                        'intentName': 'ModifyOrder',
                        'entityLabels': [
                            {
                                'entityName': 'Order',
                                'startCharIndex': 7,
                                'endCharIndex': 46
                            },
                            {
                                'entityName': 'FullPizzaWithModifiers',
                                'startCharIndex': 7,
                                'endCharIndex': 46
                            },
                            {
                                'entityName': 'PizzaType',
                                'startCharIndex': 17,
                                'endCharIndex': 32
                            },
                            {
                                'entityName': 'Size',
                                'startCharIndex': 11,
                                'endCharIndex': 15
                            },
                            {
                                'entityName': 'Quantity',
                                'startCharIndex': 7,
                                'endCharIndex': 9
                            },
                            {
                                'entityName': 'Crust',
                                'startCharIndex': 37,
                                'endCharIndex': 46
                            }
                        ]
                    }
                ]
                ";
    
                AddUtterances(utterances).Wait();
                Train().Wait();
                Status().Wait();
            }
        }
    }
    
  4. 将以 YOUR- 开头的值替换为你自己的值。

    信息 目的
    YOUR-APP-ID LUIS 应用 ID。
    YOUR-AUTHORING-KEY 32 字符创作密钥。
    YOUR-AUTHORING-ENDPOINT 创作 URL 终结点。 例如,https://replace-with-your-resource-name.api.cognitive.microsoft.com/。 在创建资源时设置资源名称。

    分配的密钥和资源可以在 LUIS 门户的“Azure 资源”页上的“管理”部分中看到。 应用 ID 可以在“应用程序设置”页的同一“管理”部分中找到。

    重要

    完成后,请记住将密钥从代码中删除,并且永远不要公开发布该密钥。 对于生产来说,请使用安全的方式存储和访问凭据,例如 Azure Key Vault。 有关详细信息,请参阅 Azure AI 服务安全性一文。

  5. 生成控制台应用程序。

    dotnet build
    
  6. 运行控制台应用程序。

    dotnet run
    
  7. 查看创作响应:

    Added utterances.
    [
        {
            "value": {
                "ExampleId": 1137150691,
                "UtteranceText": "order a pizza"
            },
            "hasError": false
        },
        {
            "value": {
                "ExampleId": 1137150692,
                "UtteranceText": "order a large pepperoni pizza"
            },
            "hasError": false
        },
        {
            "value": {
                "ExampleId": 1137150693,
                "UtteranceText": "i want two large pepperoni pizzas on thin crust"
            },
            "hasError": false
        }
    ]
    Sent training request.
    {
        "statusId": 9,
        "status": "Queued"
    }
    Requested training status.
    [
        {
            "modelId": "edb46abf-0000-41ab-beb2-a41a0fe1630f",
            "details": {
                "statusId": 9,
                "status": "Queued",
                "exampleCount": 0
            }
        },
        {
            "modelId": "a5030be2-616c-4648-bf2f-380fa9417d37",
            "details": {
                "statusId": 9,
                "status": "Queued",
                "exampleCount": 0
            }
        },
        {
            "modelId": "3f2b1f31-a3c3-4fbd-8182-e9d9dbc120b9",
            "details": {
                "statusId": 9,
                "status": "Queued",
                "exampleCount": 0
            }
        },
        {
            "modelId": "e4b6704b-1636-474c-9459-fe9ccbeba51c",
            "details": {
                "statusId": 9,
                "status": "Queued",
                "exampleCount": 0
            }
        },
        {
            "modelId": "031d3777-2a00-4a7a-9323-9a3280a30000",
            "details": {
                "statusId": 9,
                "status": "Queued",
                "exampleCount": 0
            }
        },
        {
            "modelId": "9250e7a1-06eb-4413-9432-ae132ed32583",
            "details": {
                "statusId": 9,
                "status": "Queued",
                "exampleCount": 0
            }
        }
    ]
    

清理资源

完成本快速入门后,请从文件系统中删除项目文件夹。

后续步骤

应用的最佳实践

参考文档 | 示例

先决条件

示例话语 JSON 文件

示例言语采用特定格式。

text 字段包含示例话语的文本。 intentName 字段必须对应于 LUIS 应用中的现有意向名称。 entityLabels 字段是必填的。 如果不想标记任何实体,请提供一个空数组。

如果 entityLabels 数组不为空,则 startCharIndexendCharIndex 需要标记 entityName 字段中引用的实体。 该索引从零开始。 如果标签的起始或结尾位于文本中的空白处,则添加话语的 API 调用将失败。

[
  {
    "text": "order a pizza",
    "intentName": "ModifyOrder",
    "entityLabels": [
      {
        "entityName": "Order",
        "startCharIndex": 6,
        "endCharIndex": 12
      }
    ]
  },
  {
    "text": "order a large pepperoni pizza",
    "intentName": "ModifyOrder",
    "entityLabels": [
      {
        "entityName": "Order",
        "startCharIndex": 6,
        "endCharIndex": 28
      },
      {
        "entityName": "FullPizzaWithModifiers",
        "startCharIndex": 6,
        "endCharIndex": 28
      },
      {
        "entityName": "PizzaType",
        "startCharIndex": 14,
        "endCharIndex": 28
      },
      {
        "entityName": "Size",
        "startCharIndex": 8,
        "endCharIndex": 12
      }
    ]
  },
  {
    "text": "I want two large pepperoni pizzas on thin crust",
    "intentName": "ModifyOrder",
    "entityLabels": [
      {
        "entityName": "Order",
        "startCharIndex": 7,
        "endCharIndex": 46
      },
      {
        "entityName": "FullPizzaWithModifiers",
        "startCharIndex": 7,
        "endCharIndex": 46
      },
      {
        "entityName": "PizzaType",
        "startCharIndex": 17,
        "endCharIndex": 32
      },
      {
        "entityName": "Size",
        "startCharIndex": 11,
        "endCharIndex": 15
      },
      {
        "entityName": "Quantity",
        "startCharIndex": 7,
        "endCharIndex": 9
      },
      {
        "entityName": "Crust",
        "startCharIndex": 37,
        "endCharIndex": 46
      }
    ]
  }
]

以编程方式更改模型

  1. 创建一个新文件夹以保存 Java 项目,例如 java-model-with-rest

  2. 创建名为 lib 的子目录,并将以下 Java 库中的内容复制到 lib 子目录:

  3. 创建名为 Model.java 的新文件。 添加以下代码:

    //
    // This quickstart shows how to add utterances to a LUIS model using the REST APIs.
    //
    
    import java.io.*;
    import java.net.URI;
    import org.apache.http.HttpEntity;
    import org.apache.http.HttpResponse;
    import org.apache.http.client.HttpClient;
    import org.apache.http.client.methods.HttpGet;
    import org.apache.http.client.methods.HttpPost;
    import org.apache.http.client.utils.URIBuilder;
    import org.apache.http.entity.StringEntity;
    import org.apache.http.impl.client.HttpClients;
    import org.apache.http.util.EntityUtils;
    
    // To compile, execute this command at the console:
    //      Windows: javac -cp ";lib/*" Model.java
    //      macOs: javac -cp ":lib/*" Model.java
    //      Linux: javac -cp ":lib/*" Model.java
    
    // To run, execute this command at the console:
    //      Windows: java -cp ";lib/*" Model
    //      macOs: java -cp ":lib/*" Model
    //      Linux: java -cp ":lib/*" Model
    
    public class Model
    {
        public static void main(String[] args)
        {
            try
            {
                //////////
                // Values to modify.
    
                // YOUR-APP-ID: The App ID GUID found on the www.luis.ai Application Settings page.
                String AppId = "PASTE_YOUR_LUIS_APP_ID_HERE";
    
                // YOUR-AUTHORING-KEY: Your LUIS authoring key, 32 character value.
                String Key = "PASTE_YOUR_LUIS_AUTHORING_SUBSCRIPTION_KEY_HERE";
    
                // YOUR-AUTHORING-ENDPOINT: Replace this with your authoring key endpoint.
                // For example, "https://your-resource-name.cognitiveservices.azure.com/"
                String Endpoint = "PASTE_YOUR_LUIS_AUTHORING_ENDPOINT_HERE";
    
                // NOTE: Replace this your version number. The Pizza app uses a version number of "0.1".
                String Version = "0.1";
                //////////
    
                // The list of utterances to add, in JSON format.
                String Utterances = "[{'text': 'order a pizza', 'intentName': 'ModifyOrder', 'entityLabels': [{'entityName': 'Order', 'startCharIndex': 6, 'endCharIndex': 12}]}, {'text': 'order a large pepperoni pizza', 'intentName': 'ModifyOrder', 'entityLabels': [{'entityName': 'Order', 'startCharIndex': 6, 'endCharIndex': 28}, {'entityName': 'FullPizzaWithModifiers', 'startCharIndex': 6, 'endCharIndex': 28}, {'entityName': 'PizzaType', 'startCharIndex': 14, 'endCharIndex': 28}, {'entityName': 'Size', 'startCharIndex': 8, 'endCharIndex': 12}]}, {'text': 'I want two large pepperoni pizzas on thin crust', 'intentName': 'ModifyOrder', 'entityLabels': [{'entityName': 'Order', 'startCharIndex': 7, 'endCharIndex': 46}, {'entityName': 'FullPizzaWithModifiers', 'startCharIndex': 7, 'endCharIndex': 46}, {'entityName': 'PizzaType', 'startCharIndex': 17, 'endCharIndex': 32}, {'entityName': 'Size', 'startCharIndex': 11, 'endCharIndex': 15}, {'entityName': 'Quantity', 'startCharIndex': 7, 'endCharIndex': 9}, {'entityName': 'Crust', 'startCharIndex': 37, 'endCharIndex': 46}]}]";
    
                // Create the URLs for uploading example utterances and for training.
                URIBuilder addUtteranceURL = new URIBuilder(Endpoint + "luis/authoring/v3.0-preview/apps/" + AppId + "/versions/" + Version + "/examples");
                URIBuilder trainURL = new URIBuilder(Endpoint + "luis/authoring/v3.0-preview/apps/" + AppId + "/versions/" + Version + "/train");
                URI addUtterancesURI = addUtteranceURL.build();
                URI trainURI = trainURL.build();
    
    
                // Add the utterances.
    
                // Create the request.
                HttpClient addUtterancesClient = HttpClients.createDefault();
                HttpPost addUtterancesRequest = new HttpPost(addUtterancesURI);
    
                // Add the headers.
                addUtterancesRequest.setHeader("Ocp-Apim-Subscription-Key",Key);
                addUtterancesRequest.setHeader("Content-type","application/json");
    
                // Add the body.
                StringEntity stringEntity = new StringEntity(Utterances);
                addUtterancesRequest.setEntity(stringEntity);
    
                // Execute the request and obtain the response.
                HttpResponse addUtterancesResponse = addUtterancesClient.execute(addUtterancesRequest);
                HttpEntity addUtterancesEntity = addUtterancesResponse.getEntity();
    
                // Print the response on the console.
                if (addUtterancesEntity != null)
                {
                    System.out.println(EntityUtils.toString(addUtterancesEntity));
                }
    
    
                // Train the model.
    
                // Create the request.
                HttpClient trainClient = HttpClients.createDefault();
                HttpPost trainRequest = new HttpPost(trainURI);
    
                // Add the headers.
                trainRequest.setHeader("Ocp-Apim-Subscription-Key",Key);
                trainRequest.setHeader("Content-type","application/json");
    
                // Execute the request and obtain the response.
                HttpResponse trainResponse = trainClient.execute(trainRequest);
                HttpEntity trainEntity = trainResponse.getEntity();
    
                // Print the response on the console.
                if (trainEntity != null)
                {
                    System.out.println(EntityUtils.toString(trainEntity));
                }
    
    
                // Get the training status.
    
    
                // Create the request.
                HttpClient trainStatusClient = HttpClients.createDefault();
                HttpGet trainStatusRequest = new HttpGet(trainURI);
    
                // Add the headers.
                trainStatusRequest.setHeader("Ocp-Apim-Subscription-Key",Key);
                trainStatusRequest.setHeader("Content-type","application/json");
    
                // Execute the request and obtain the response.
                HttpResponse trainStatusResponse = trainStatusClient.execute(trainStatusRequest);
                HttpEntity trainStatusEntity = trainStatusResponse.getEntity();
    
                // Print the response on the console.
                if (trainStatusEntity != null)
                {
                    System.out.println(EntityUtils.toString(trainStatusEntity));
                }
            }
    
            // Display errors if they occur.
            catch (Exception e)
            {
                System.out.println(e.getMessage());
            }
        }
    }
    
  4. 将以 YOUR- 开头的值替换为你自己的值。

    信息 目的
    YOUR-APP-ID LUIS 应用 ID。
    YOUR-AUTHORING-KEY 32 字符创作密钥。
    YOUR-AUTHORING-ENDPOINT 创作 URL 终结点。 例如,https://replace-with-your-resource-name.api.cognitive.microsoft.com/。 在创建资源时设置资源名称。

    分配的密钥和资源可以在 LUIS 门户的“Azure 资源”页上的“管理”部分中看到。 应用 ID 可以在“应用程序设置”页的同一“管理”部分中找到。

    重要

    完成后,请记住将密钥从代码中删除,并且永远不要公开发布该密钥。 对于生产来说,请使用安全的方式存储和访问凭据,例如 Azure Key Vault。 有关详细信息,请参阅 Azure AI 服务安全性一文。

  5. 在创建 Model.java 文件的同一目录中,在命令提示符下输入以下命令来编译 Java 文件:

    • 如果使用的是 Windows,请使用此命令:javac -cp ";lib/*" Model.java
    • 如果使用的是 macOS 或 Linux,请使用此命令:javac -cp ":lib/*" Model.java
  6. 通过在命令提示符下输入以下文本从命令行运行 Java 应用程序:

    • 如果使用的是 Windows,请使用此命令:java -cp ";lib/*" Model
    • 如果使用的是 macOS 或 Linux,请使用此命令:java -cp ":lib/*" Model
  7. 查看创作响应:

    [{"value":{"ExampleId":1137150691,"UtteranceText":"order a pizza"},"hasError":false},{"value":{"ExampleId":1137150692,"UtteranceText":"order a large pepperoni pizza"},"hasError":false},{"value":{"ExampleId":1137150693,"UtteranceText":"i want two large pepperoni pizzas on thin crust"},"hasError":false}]
    {"statusId":9,"status":"Queued"}
    [{"modelId":"edb46abf-0000-41ab-beb2-a41a0fe1630f","details":{"statusId":9,"status":"Queued","exampleCount":0}},{"modelId":"a5030be2-616c-4648-bf2f-380fa9417d37","details":{"statusId":9,"status":"Queued","exampleCount":0}},{"modelId":"3f2b1f31-a3c3-4fbd-8182-e9d9dbc120b9","details":{"statusId":9,"status":"Queued","exampleCount":0}},{"modelId":"e4b6704b-1636-474c-9459-fe9ccbeba51c","details":{"statusId":9,"status":"Queued","exampleCount":0}},{"modelId":"031d3777-2a00-4a7a-9323-9a3280a30000","details":{"statusId":9,"status":"Queued","exampleCount":0}},{"modelId":"9250e7a1-06eb-4413-9432-ae132ed32583","details":{"statusId":3,"status":"InProgress","exampleCount":0,"progressSubstatus":"CollectingData"}}]
    

    下面是为提高可读性而进行了格式设置的输出:

    [
      {
        "value": {
          "ExampleId": 1137150691,
          "UtteranceText": "order a pizza"
        },
        "hasError": false
      },
      {
        "value": {
          "ExampleId": 1137150692,
          "UtteranceText": "order a large pepperoni pizza"
        },
        "hasError": false
      },
      {
        "value": {
          "ExampleId": 1137150693,
          "UtteranceText": "i want two large pepperoni pizzas on thin crust"
        },
        "hasError": false
      }
    ]
    {
      "statusId": 9,
      "status": "Queued"
    }
    [
      {
        "modelId": "edb46abf-0000-41ab-beb2-a41a0fe1630f",
        "details": {
          "statusId": 9,
          "status": "Queued",
          "exampleCount": 0
        }
      },
      {
        "modelId": "a5030be2-616c-4648-bf2f-380fa9417d37",
        "details": {
          "statusId": 9,
          "status": "Queued",
          "exampleCount": 0
        }
      },
      {
        "modelId": "3f2b1f31-a3c3-4fbd-8182-e9d9dbc120b9",
        "details": {
          "statusId": 9,
          "status": "Queued",
          "exampleCount": 0
        }
      },
      {
        "modelId": "e4b6704b-1636-474c-9459-fe9ccbeba51c",
        "details": {
          "statusId": 9,
          "status": "Queued",
          "exampleCount": 0
        }
      },
      {
        "modelId": "031d3777-2a00-4a7a-9323-9a3280a30000",
        "details": {
          "statusId": 9,
          "status": "Queued",
          "exampleCount": 0
        }
      },
      {
        "modelId": "9250e7a1-06eb-4413-9432-ae132ed32583",
        "details": {
          "statusId": 3,
          "status": "InProgress",
          "exampleCount": 0,
          "progressSubstatus": "CollectingData"
        }
      }
    ]
    

清理资源

完成本快速入门后,请从文件系统中删除项目文件夹。

后续步骤

应用的最佳实践

参考文档 | 示例

先决条件

示例话语 JSON 文件

示例言语采用特定格式。

text 字段包含示例话语的文本。 intentName 字段必须对应于 LUIS 应用中的现有意向名称。 entityLabels 字段是必填的。 如果不想标记任何实体,请提供一个空数组。

如果 entityLabels 数组不为空,则 startCharIndexendCharIndex 需要标记 entityName 字段中引用的实体。 该索引从零开始。 如果标签的起始或结尾位于文本中的空白处,则添加话语的 API 调用将失败。

[
  {
    "text": "order a pizza",
    "intentName": "ModifyOrder",
    "entityLabels": [
      {
        "entityName": "Order",
        "startCharIndex": 6,
        "endCharIndex": 12
      }
    ]
  },
  {
    "text": "order a large pepperoni pizza",
    "intentName": "ModifyOrder",
    "entityLabels": [
      {
        "entityName": "Order",
        "startCharIndex": 6,
        "endCharIndex": 28
      },
      {
        "entityName": "FullPizzaWithModifiers",
        "startCharIndex": 6,
        "endCharIndex": 28
      },
      {
        "entityName": "PizzaType",
        "startCharIndex": 14,
        "endCharIndex": 28
      },
      {
        "entityName": "Size",
        "startCharIndex": 8,
        "endCharIndex": 12
      }
    ]
  },
  {
    "text": "I want two large pepperoni pizzas on thin crust",
    "intentName": "ModifyOrder",
    "entityLabels": [
      {
        "entityName": "Order",
        "startCharIndex": 7,
        "endCharIndex": 46
      },
      {
        "entityName": "FullPizzaWithModifiers",
        "startCharIndex": 7,
        "endCharIndex": 46
      },
      {
        "entityName": "PizzaType",
        "startCharIndex": 17,
        "endCharIndex": 32
      },
      {
        "entityName": "Size",
        "startCharIndex": 11,
        "endCharIndex": 15
      },
      {
        "entityName": "Quantity",
        "startCharIndex": 7,
        "endCharIndex": 9
      },
      {
        "entityName": "Crust",
        "startCharIndex": 37,
        "endCharIndex": 46
      }
    ]
  }
]

创建 Pizza 应用

创建披萨应用。

  1. 选择 pizza-app-for-luis-v6.json,打开 pizza-app-for-luis.json 文件的 GitHub 页面。
  2. 右键单击或长按“原始”按钮,然后选择“将链接另存为”,将 保存到计算机。
  3. 登录到 LUIS 门户
  4. 选择我的应用
  5. 在“我的应用”页面上,选择“+ 新建用于对话的应用” 。
  6. 选择“导入为 JSON”。
  7. 在“导入新应用”对话框中,选择“选择文件”按钮 。
  8. 选择下载的 pizza-app-for-luis.json 文件,然后选择“打开”。
  9. 在“导入新应用”对话框的“名称”字段中,输入 Pizza 应用的名称,然后选择“完成”按钮 。

随即导入应用。

如果看到一个对话框“如何创建有效的 LUIS 应用”,关闭该对话框。

训练并发布 Pizza 应用

Pizza 应用中应会显示“意向”页面,其中显示了一个意向列表。

  1. 在 LUIS 网站的右上方,选择“训练”按钮。

    “训练”按钮

  2. 当“训练”按钮处于禁用状态时,即表示训练完成。

若要在聊天机器人或其他客户端应用程序中接收 LUIS 预测,需要将应用发布到预测终结点。

  1. 在右上方的导航栏中选择“发布”。

    用于发布到终结点的按钮的屏幕截图。

  2. 选择“生产”槽,然后选择“完成” 。

    LUIS 发布到终结点的屏幕截图。

  3. 在通知中选择“访问终结点 URL”,以转到“Azure 资源”页 。 只有你拥有与应用关联的预测资源时,才能看到 URL。 还可以单击“管理”来找到“Azure 资源”页 。

    显示该应用已发布的消息的屏幕截图。

向披萨应用添加创作资源

  1. 选择“管理”。
  2. 选择“Azure 资源”。
  3. 选择“创作资源”。
  4. 选择“更改创作资源”。

如果你有创作资源,请输入你的创作资源的租户名称、订阅名称以及 LUIS 资源名称。

如果你没有创作资源:

  1. 选择“新建资源”。
  2. 输入租户名称、资源名称、订阅名称和 Azure 资源组名称。

现在 Pizza 应用可以使用了。

记录 Pizza 应用的访问值

若要使用新的披萨应用,你需要该披萨应用的应用 ID、授权密钥和授权终结点。 若要获取预测,需要单独的预测终结点和预测密钥。

若要查找这些值:

  1. 在“意向”页面,选择“管理” 。
  2. 在“应用程序设置”页面,记录“应用 ID” 。
  3. 选择“Azure 资源”。
  4. 选择“创作资源”。
  5. 从“创作资源”和“预测资源”选项卡,记录“主键” 。 此值是你的创作密钥。
  6. 记录“终结点 URL”。 此值是你的创作终结点。

以编程方式更改模型

  1. 创建名为 predict.go 的新文件。 添加以下代码:

    //
    // This quickstart shows how to add utterances to a LUIS model using the REST APIs.
    //
    
    // dependencies
    package main
    import (
        "fmt"
        "net/http"
        "io/ioutil"
        "log"
        "strings"
    )
    
    // main function
    func main() {
    
        //////////
        // Values to modify.
    
        // YOUR-APP-ID: The App ID GUID found on the www.luis.ai Application Settings page.
        var appID = "PASTE_YOUR_LUIS_APP_ID_HERE"
    
        // YOUR-AUTHORING-KEY: Your LUIS authoring key, 32 character value.
        var authoringKey = "PASTE_YOUR_LUIS_AUTHORING_SUBSCRIPTION_KEY_HERE"
    
        //  YOUR-AUTHORING-ENDPOINT: Replace this with your authoring key endpoint.
        // For example, "https://your-resource-name.cognitiveservices.azure.com/"
        var endpoint = "PASTE_YOUR_LUIS_AUTHORING_ENDPOINT_HERE"
    
        // NOTE: Replace this your version number. The Pizza app uses a version number of "0.1".
        var version = "0.1"
        //////////
    
        var exampleUtterances = "[{'text': 'order a pizza', 'intentName': 'ModifyOrder', 'entityLabels': [{'entityName': 'Order', 'startCharIndex': 6, 'endCharIndex': 12}]}, {'text': 'order a large pepperoni pizza', 'intentName': 'ModifyOrder', 'entityLabels': [{'entityName': 'Order', 'startCharIndex': 6, 'endCharIndex': 28}, {'entityName': 'FullPizzaWithModifiers', 'startCharIndex': 6, 'endCharIndex': 28}, {'entityName': 'PizzaType', 'startCharIndex': 14, 'endCharIndex': 28}, {'entityName': 'Size', 'startCharIndex': 8, 'endCharIndex': 12}]}, {'text': 'I want two large pepperoni pizzas on thin crust', 'intentName': 'ModifyOrder', 'entityLabels': [{'entityName': 'Order', 'startCharIndex': 7, 'endCharIndex': 46}, {'entityName': 'FullPizzaWithModifiers', 'startCharIndex': 7, 'endCharIndex': 46}, {'entityName': 'PizzaType', 'startCharIndex': 17, 'endCharIndex': 32}, {'entityName': 'Size', 'startCharIndex': 11, 'endCharIndex': 15}, {'entityName': 'Quantity', 'startCharIndex': 7, 'endCharIndex': 9}, {'entityName': 'Crust', 'startCharIndex': 37, 'endCharIndex': 46}]}]"
    
        fmt.Println("add example utterances requested")
        addUtterance(authoringKey, appID, version, exampleUtterances, endpoint)
    
        fmt.Println("training selected")
        requestTraining(authoringKey, appID, version, endpoint)
    
        fmt.Println("training status selected")
        getTrainingStatus(authoringKey, appID, version, endpoint)
    }
    
    // Send the list of utterances to the model.
    func addUtterance(authoringKey string, appID string,  version string, labeledExampleUtterances string, endpoint string){
    
        var authoringUrl = fmt.Sprintf("%sluis/authoring/v3.0-preview/apps/%s/versions/%s/examples", endpoint, appID, version)
    
        httpRequest("POST", authoringUrl, authoringKey, labeledExampleUtterances)
    }
    
    // Request training.
    func requestTraining(authoringKey string, appID string,  version string, endpoint string){
    
        trainApp("POST", authoringKey, appID, version, endpoint)
    }
    
    
    func trainApp(httpVerb string, authoringKey string, appID string,  version string, endpoint string){
    
        var authoringUrl = fmt.Sprintf("%sluis/authoring/v3.0-preview/apps/%s/versions/%s/train", endpoint, appID, version)
    
        httpRequest(httpVerb,authoringUrl, authoringKey, "")
    }
    
    
    func getTrainingStatus(authoringKey string, appID string, version string, endpoint string){
    
        trainApp("GET", authoringKey, appID, version, endpoint)
    }
    
    // generic HTTP request
    // includes setting header with authoring key
    func httpRequest(httpVerb string, url string, authoringKey string, body string){
    
        client := &http.Client{}
    
        request, err := http.NewRequest(httpVerb, url, strings.NewReader(body))
        request.Header.Add("Ocp-Apim-Subscription-Key", authoringKey)
    
        fmt.Println("body")
        fmt.Println(body)
    
        response, err := client.Do(request)
    
        if err != nil {
            log.Fatal(err)
        } else {
            defer response.Body.Close()
    
            contents, err := ioutil.ReadAll(response.Body)
    
            if err != nil {
                log.Fatal(err)
            }
    
            fmt.Println("   ", response.StatusCode)
            fmt.Println(string(contents))
        }
    }
    
  2. 将以 YOUR- 开头的值替换为你自己的值。

    信息 目的
    YOUR-APP-ID LUIS 应用 ID。
    YOUR-AUTHORING-KEY 32 字符创作密钥。
    YOUR-AUTHORING-ENDPOINT 创作 URL 终结点。 例如,https://replace-with-your-resource-name.api.cognitive.microsoft.com/。 在创建资源时设置资源名称。

    分配的密钥和资源可以在 LUIS 门户的“Azure 资源”页上的“管理”部分中看到。 应用 ID 可以在“应用程序设置”页的同一“管理”部分中找到。

    重要

    完成后,请记住将密钥从代码中删除,并且永远不要公开发布该密钥。 对于生产来说,请使用安全的方式存储和访问凭据,例如 Azure Key Vault。 有关详细信息,请参阅 Azure AI 服务安全性一文。

  3. 在创建该文件的同一目录中,在命令提示符下输入以下命令来编译 Go 文件:

    go build model.go
    
  4. 通过在命令提示符下输入以下文本从命令行运行 Go 应用程序:

    go run model.go
    
  5. 查看创作响应:

    add example utterances requested
    body
    [{'text': 'order a pizza', 'intentName': 'ModifyOrder', 'entityLabels': [{'entityName': 'Order', 'startCharIndex': 6, 'endCharIndex': 12}]}, {'text': 'order a large pepperoni pizza', 'intentName': 'ModifyOrder', 'entityLabels': [{'entityName': 'Order', 'startCharIndex': 6, 'endCharIndex': 28}, {'entityName': 'FullPizzaWithModifiers', 'startCharIndex': 6, 'endCharIndex': 28}, {'entityName': 'PizzaType', 'startCharIndex': 14, 'endCharIndex': 28}, {'entityName': 'Size', 'startCharIndex': 8, 'endCharIndex': 12}]}, {'text': 'I want two large pepperoni pizzas on thin crust', 'intentName': 'ModifyOrder', 'entityLabels': [{'entityName': 'Order', 'startCharIndex': 7, 'endCharIndex': 46}, {'entityName': 'FullPizzaWithModifiers', 'startCharIndex': 7, 'endCharIndex': 46}, {'entityName': 'PizzaType', 'startCharIndex': 17, 'endCharIndex': 32}, {'entityName': 'Size', 'startCharIndex': 11, 'endCharIndex': 15}, {'entityName': 'Quantity', 'startCharIndex': 7, 'endCharIndex': 9}, {'entityName': 'Crust', 'startCharIndex': 37, 'endCharIndex': 46}]}]
        201
    [{"value":{"ExampleId":1137150691,"UtteranceText":"order a pizza"},"hasError":false},{"value":{"ExampleId":1137150692,"UtteranceText":"order a large pepperoni pizza"},"hasError":false},{"value":{"ExampleId":1137150693,"UtteranceText":"i want two large pepperoni pizzas on thin crust"},"hasError":false}]
    training selected
    body
    
        202
    {"statusId":9,"status":"Queued"}
    training status selected
    body
    
        200
    [{"modelId":"edb46abf-0000-41ab-beb2-a41a0fe1630f","details":{"statusId":9,"status":"Queued","exampleCount":0}},{"modelId":"a5030be2-616c-4648-bf2f-380fa9417d37","details":{"statusId":9,"status":"Queued","exampleCount":0}},{"modelId":"3f2b1f31-a3c3-4fbd-8182-e9d9dbc120b9","details":{"statusId":9,"status":"Queued","exampleCount":0}},{"modelId":"e4b6704b-1636-474c-9459-fe9ccbeba51c","details":{"statusId":9,"status":"Queued","exampleCount":0}},{"modelId":"031d3777-2a00-4a7a-9323-9a3280a30000","details":{"statusId":9,"status":"Queued","exampleCount":0}},{"modelId":"9250e7a1-06eb-4413-9432-ae132ed32583","details":{"statusId":9,"status":"Queued","exampleCount":0}}]
    

    下面是为提高可读性而进行了格式设置的输出:

    add example utterances requested
    body
    [
      {
        'text': 'order a pizza',
        'intentName': 'ModifyOrder',
        'entityLabels': [
          {
            'entityName': 'Order',
            'startCharIndex': 6,
            'endCharIndex': 12
          }
        ]
      },
      {
        'text': 'order a large pepperoni pizza',
        'intentName': 'ModifyOrder',
        'entityLabels': [
          {
            'entityName': 'Order',
            'startCharIndex': 6,
            'endCharIndex': 28
          },
          {
            'entityName': 'FullPizzaWithModifiers',
            'startCharIndex': 6,
            'endCharIndex': 28
          },
          {
            'entityName': 'PizzaType',
            'startCharIndex': 14,
            'endCharIndex': 28
          },
          {
            'entityName': 'Size',
            'startCharIndex': 8,
            'endCharIndex': 12
          }
        ]
      },
      {
        'text': 'I want two large pepperoni pizzas on thin crust',
        'intentName': 'ModifyOrder',
        'entityLabels': [
          {
            'entityName': 'Order',
            'startCharIndex': 7,
            'endCharIndex': 46
          },
          {
            'entityName': 'FullPizzaWithModifiers',
            'startCharIndex': 7,
            'endCharIndex': 46
          },
          {
            'entityName': 'PizzaType',
            'startCharIndex': 17,
            'endCharIndex': 32
          },
          {
            'entityName': 'Size',
            'startCharIndex': 11,
            'endCharIndex': 15
          },
          {
            'entityName': 'Quantity',
            'startCharIndex': 7,
            'endCharIndex': 9
          },
          {
            'entityName': 'Crust',
            'startCharIndex': 37,
            'endCharIndex': 46
          }
        ]
      }
    ]
    
        201
    [
      {
        "value": {
          "ExampleId": 1137150691,
          "UtteranceText": "order a pizza"
        },
        "hasError": false
      },
      {
        "value": {
          "ExampleId": 1137150692,
          "UtteranceText": "order a large pepperoni pizza"
        },
        "hasError": false
      },
      {
        "value": {
          "ExampleId": 1137150693,
          "UtteranceText": "i want two large pepperoni pizzas on thin crust"
        },
        "hasError": false
      }
    ]
    training selected
    body
    
        202
    {
      "statusId": 9,
      "status": "Queued"
    }
    training status selected
    body
    
        200
    [
      {
        "modelId": "edb46abf-0000-41ab-beb2-a41a0fe1630f",
        "details": {
          "statusId": 9,
          "status": "Queued",
          "exampleCount": 0
        }
      },
      {
        "modelId": "a5030be2-616c-4648-bf2f-380fa9417d37",
        "details": {
          "statusId": 9,
          "status": "Queued",
          "exampleCount": 0
        }
      },
      {
        "modelId": "3f2b1f31-a3c3-4fbd-8182-e9d9dbc120b9",
        "details": {
          "statusId": 9,
          "status": "Queued",
          "exampleCount": 0
        }
      },
      {
        "modelId": "e4b6704b-1636-474c-9459-fe9ccbeba51c",
        "details": {
          "statusId": 9,
          "status": "Queued",
          "exampleCount": 0
        }
      },
      {
        "modelId": "031d3777-2a00-4a7a-9323-9a3280a30000",
        "details": {
          "statusId": 9,
          "status": "Queued",
          "exampleCount": 0
        }
      },
      {
        "modelId": "9250e7a1-06eb-4413-9432-ae132ed32583",
        "details": {
          "statusId": 9,
          "status": "Queued",
          "exampleCount": 0
        }
      }
    ]
    

清理资源

完成本快速入门后,请从文件系统中删除该文件。

后续步骤

应用的最佳实践

参考文档 | 示例

先决条件

示例话语 JSON 文件

示例言语采用特定格式。

text 字段包含示例话语的文本。 intentName 字段必须对应于 LUIS 应用中的现有意向名称。 entityLabels 字段是必填的。 如果不想标记任何实体,请提供一个空数组。

如果 entityLabels 数组不为空,则 startCharIndexendCharIndex 需要标记 entityName 字段中引用的实体。 该索引从零开始。 如果标签的起始或结尾位于文本中的空白处,则添加话语的 API 调用将失败。

[
  {
    "text": "order a pizza",
    "intentName": "ModifyOrder",
    "entityLabels": [
      {
        "entityName": "Order",
        "startCharIndex": 6,
        "endCharIndex": 12
      }
    ]
  },
  {
    "text": "order a large pepperoni pizza",
    "intentName": "ModifyOrder",
    "entityLabels": [
      {
        "entityName": "Order",
        "startCharIndex": 6,
        "endCharIndex": 28
      },
      {
        "entityName": "FullPizzaWithModifiers",
        "startCharIndex": 6,
        "endCharIndex": 28
      },
      {
        "entityName": "PizzaType",
        "startCharIndex": 14,
        "endCharIndex": 28
      },
      {
        "entityName": "Size",
        "startCharIndex": 8,
        "endCharIndex": 12
      }
    ]
  },
  {
    "text": "I want two large pepperoni pizzas on thin crust",
    "intentName": "ModifyOrder",
    "entityLabels": [
      {
        "entityName": "Order",
        "startCharIndex": 7,
        "endCharIndex": 46
      },
      {
        "entityName": "FullPizzaWithModifiers",
        "startCharIndex": 7,
        "endCharIndex": 46
      },
      {
        "entityName": "PizzaType",
        "startCharIndex": 17,
        "endCharIndex": 32
      },
      {
        "entityName": "Size",
        "startCharIndex": 11,
        "endCharIndex": 15
      },
      {
        "entityName": "Quantity",
        "startCharIndex": 7,
        "endCharIndex": 9
      },
      {
        "entityName": "Crust",
        "startCharIndex": 37,
        "endCharIndex": 46
      }
    ]
  }
]

创建 Node.js 项目

  1. 创建一个新文件夹以保存 Node.js 项目,例如 node-model-with-rest

  2. 打开新的命令提示符,导航到你创建的文件夹,并执行以下命令:

    npm init
    

    在每个提示符下按 Enter 以接受默认设置。

  3. 输入以下命令安装“请求-承诺”模块:

    npm install --save request
    npm install --save request-promise
    npm install --save querystring
    

以编程方式更改模型

  1. 创建名为 model.js 的新文件。 添加以下代码:

    //
    // This quickstart shows how to add utterances to a LUIS model using the REST APIs.
    //
    
    var request = require('request-promise');
    
    //////////
    // Values to modify.
    
    // YOUR-APP-ID: The App ID GUID found on the www.luis.ai Application Settings page.
    const LUIS_appId = "PASTE_YOUR_LUIS_APP_ID_HERE";
    
    // YOUR-AUTHORING-KEY: Your LUIS authoring key, 32 character value.
    const LUIS_authoringKey = "PASTE_YOUR_LUIS_AUTHORING_SUBSCRIPTION_KEY_HERE";
    
    // YOUR-AUTHORING-ENDPOINT: Replace this with your authoring key endpoint.
    // For example, "https://your-resource-name.cognitiveservices.azure.com/"
    const LUIS_endpoint = "PASTE_YOUR_LUIS_AUTHORING_ENDPOINT_HERE";
    
    // NOTE: Replace this your version number. The Pizza app uses a version number of "0.1".
    const LUIS_versionId = "0.1";
    //////////
    
    const addUtterancesURI = `${LUIS_endpoint}luis/authoring/v3.0-preview/apps/${LUIS_appId}/versions/${LUIS_versionId}/examples`;
    const addTrainURI = `${LUIS_endpoint}luis/authoring/v3.0-preview/apps/${LUIS_appId}/versions/${LUIS_versionId}/train`;
    
    const utterances = [
        {
            'text': 'order a pizza',
            'intentName': 'ModifyOrder',
            'entityLabels': [
                {
                    'entityName': 'Order',
                    'startCharIndex': 6,
                    'endCharIndex': 12
                }
            ]
        },
        {
            'text': 'order a large pepperoni pizza',
            'intentName': 'ModifyOrder',
            'entityLabels': [
                {
                    'entityName': 'Order',
                    'startCharIndex': 6,
                    'endCharIndex': 28
                },
                {
                    'entityName': 'FullPizzaWithModifiers',
                    'startCharIndex': 6,
                    'endCharIndex': 28
                },
                {
                    'entityName': 'PizzaType',
                    'startCharIndex': 14,
                    'endCharIndex': 28
                },
                {
                    'entityName': 'Size',
                    'startCharIndex': 8,
                    'endCharIndex': 12
                }
            ]
        },
        {
            'text': 'I want two large pepperoni pizzas on thin crust',
            'intentName': 'ModifyOrder',
            'entityLabels': [
                {
                    'entityName': 'Order',
                    'startCharIndex': 7,
                    'endCharIndex': 46
                },
                {
                    'entityName': 'FullPizzaWithModifiers',
                    'startCharIndex': 7,
                    'endCharIndex': 46
                },
                {
                    'entityName': 'PizzaType',
                    'startCharIndex': 17,
                    'endCharIndex': 32
                },
                {
                    'entityName': 'Size',
                    'startCharIndex': 11,
                    'endCharIndex': 15
                },
                {
                    'entityName': 'Quantity',
                    'startCharIndex': 7,
                    'endCharIndex': 9
                },
                {
                    'entityName': 'Crust',
                    'startCharIndex': 37,
                    'endCharIndex': 46
                }
            ]
        }
    ];
    
    // Main function.
    const main = async() =>{
    
        await addUtterances(utterances);
        await train("POST");
        await train("GET");
    
    }
    
    // Adds the utterances to the model.
    const addUtterances = async (utterances) => {
    
        const options = {
            uri: addUtterancesURI,
            method: 'POST',
            headers: {
                'Ocp-Apim-Subscription-Key': LUIS_authoringKey
            },
            json: true,
            body: utterances
        };
    
        const response = await request(options)
        console.log("addUtterance:\n" + JSON.stringify(response, null, 2));
    }
    
    // With verb === "POST", sends a training request.
    // With verb === "GET", obtains the training status.
    const train = async (verb) => {
    
        const options = {
            uri: addTrainURI,
            method: verb,
            headers: {
                'Ocp-Apim-Subscription-Key': LUIS_authoringKey
            },
            json: true,
            body: null // The body can be empty for a training request
        };
    
        const response = await request(options)
        console.log("train " + verb + ":\n" + JSON.stringify(response, null, 2));
    }
    
    // MAIN
    main().then(() => console.log("done")).catch((err)=> console.log(err));
    
  2. 将以 YOUR- 开头的值替换为你自己的值。

    信息 目的
    YOUR-APP-ID LUIS 应用 ID。
    YOUR-AUTHORING-KEY 32 字符创作密钥。
    YOUR-AUTHORING-ENDPOINT 创作 URL 终结点。 例如,https://replace-with-your-resource-name.api.cognitive.microsoft.com/。 在创建资源时设置资源名称。

    分配的密钥和资源可以在 LUIS 门户的“Azure 资源”页上的“管理”部分中看到。 应用 ID 可以在“应用程序设置”页的同一“管理”部分中找到。

    重要

    完成后,请记住将密钥从代码中删除,并且永远不要公开发布该密钥。 对于生产来说,请使用安全的方式存储和访问凭据,例如 Azure Key Vault。 有关详细信息,请参阅 Azure AI 服务安全性一文。

  3. 在命令提示符处,输入下列命令以运行项目:

    node model.js
    
  4. 查看创作响应:

    addUtterance:
    [
      {
        "value": {
          "ExampleId": 1137150691,
          "UtteranceText": "order a pizza"
        },
        "hasError": false
      },
      {
        "value": {
          "ExampleId": 1137150692,
          "UtteranceText": "order a large pepperoni pizza"
        },
        "hasError": false
      },
      {
        "value": {
          "ExampleId": 1137150693,
          "UtteranceText": "i want two large pepperoni pizzas on thin crust"
        },
        "hasError": false
      }
    ]
    train POST:
    {
      "statusId": 9,
      "status": "Queued"
    }
    train GET:
    [
      {
        "modelId": "edb46abf-0000-41ab-beb2-a41a0fe1630f",
        "details": {
          "statusId": 9,
          "status": "Queued",
          "exampleCount": 0
        }
      },
      {
        "modelId": "a5030be2-616c-4648-bf2f-380fa9417d37",
        "details": {
          "statusId": 9,
          "status": "Queued",
          "exampleCount": 0
        }
      },
      {
        "modelId": "3f2b1f31-a3c3-4fbd-8182-e9d9dbc120b9",
        "details": {
          "statusId": 9,
          "status": "Queued",
          "exampleCount": 0
        }
      },
      {
        "modelId": "e4b6704b-1636-474c-9459-fe9ccbeba51c",
        "details": {
          "statusId": 9,
          "status": "Queued",
          "exampleCount": 0
        }
      },
      {
        "modelId": "031d3777-2a00-4a7a-9323-9a3280a30000",
        "details": {
          "statusId": 9,
          "status": "Queued",
          "exampleCount": 0
        }
      },
      {
        "modelId": "9250e7a1-06eb-4413-9432-ae132ed32583",
        "details": {
          "statusId": 9,
          "status": "Queued",
          "exampleCount": 0
        }
      }
    ]
    done
    

清理资源

完成本快速入门后,请从文件系统中删除项目文件夹。

后续步骤

应用的最佳实践

参考文档 | 示例

先决条件

示例话语 JSON 文件

示例言语采用特定格式。

text 字段包含示例话语的文本。 intentName 字段必须对应于 LUIS 应用中的现有意向名称。 entityLabels 字段是必填的。 如果不想标记任何实体,请提供一个空数组。

如果 entityLabels 数组不为空,则 startCharIndexendCharIndex 需要标记 entityName 字段中引用的实体。 该索引从零开始。 如果标签的起始或结尾位于文本中的空白处,则添加话语的 API 调用将失败。

[
  {
    "text": "order a pizza",
    "intentName": "ModifyOrder",
    "entityLabels": [
      {
        "entityName": "Order",
        "startCharIndex": 6,
        "endCharIndex": 12
      }
    ]
  },
  {
    "text": "order a large pepperoni pizza",
    "intentName": "ModifyOrder",
    "entityLabels": [
      {
        "entityName": "Order",
        "startCharIndex": 6,
        "endCharIndex": 28
      },
      {
        "entityName": "FullPizzaWithModifiers",
        "startCharIndex": 6,
        "endCharIndex": 28
      },
      {
        "entityName": "PizzaType",
        "startCharIndex": 14,
        "endCharIndex": 28
      },
      {
        "entityName": "Size",
        "startCharIndex": 8,
        "endCharIndex": 12
      }
    ]
  },
  {
    "text": "I want two large pepperoni pizzas on thin crust",
    "intentName": "ModifyOrder",
    "entityLabels": [
      {
        "entityName": "Order",
        "startCharIndex": 7,
        "endCharIndex": 46
      },
      {
        "entityName": "FullPizzaWithModifiers",
        "startCharIndex": 7,
        "endCharIndex": 46
      },
      {
        "entityName": "PizzaType",
        "startCharIndex": 17,
        "endCharIndex": 32
      },
      {
        "entityName": "Size",
        "startCharIndex": 11,
        "endCharIndex": 15
      },
      {
        "entityName": "Quantity",
        "startCharIndex": 7,
        "endCharIndex": 9
      },
      {
        "entityName": "Crust",
        "startCharIndex": 37,
        "endCharIndex": 46
      }
    ]
  }
]

创建 Pizza 应用

创建披萨应用。

  1. 选择 pizza-app-for-luis-v6.json,打开 pizza-app-for-luis.json 文件的 GitHub 页面。
  2. 右键单击或长按“原始”按钮,然后选择“将链接另存为”,将 保存到计算机。
  3. 登录到 LUIS 门户
  4. 选择我的应用
  5. 在“我的应用”页面上,选择“+ 新建用于对话的应用” 。
  6. 选择“导入为 JSON”。
  7. 在“导入新应用”对话框中,选择“选择文件”按钮 。
  8. 选择下载的 pizza-app-for-luis.json 文件,然后选择“打开”。
  9. 在“导入新应用”对话框的“名称”字段中,输入 Pizza 应用的名称,然后选择“完成”按钮 。

随即导入应用。

如果看到一个对话框“如何创建有效的 LUIS 应用”,关闭该对话框。

训练并发布 Pizza 应用

Pizza 应用中应会显示“意向”页面,其中显示了一个意向列表。

  1. 在 LUIS 网站的右上方,选择“训练”按钮。

    “训练”按钮

  2. 当“训练”按钮处于禁用状态时,即表示训练完成。

若要在聊天机器人或其他客户端应用程序中接收 LUIS 预测,需要将应用发布到预测终结点。

  1. 在右上方的导航栏中选择“发布”。

    用于发布到终结点的按钮的屏幕截图。

  2. 选择“生产”槽,然后选择“完成” 。

    LUIS 发布到终结点的屏幕截图。

  3. 在通知中选择“访问终结点 URL”,以转到“Azure 资源”页 。 只有你拥有与应用关联的预测资源时,才能看到 URL。 还可以单击“管理”来找到“Azure 资源”页 。

    显示该应用已发布的消息的屏幕截图。

向披萨应用添加创作资源

  1. 选择“管理”。
  2. 选择“Azure 资源”。
  3. 选择“创作资源”。
  4. 选择“更改创作资源”。

如果你有创作资源,请输入你的创作资源的租户名称、订阅名称以及 LUIS 资源名称。

如果你没有创作资源:

  1. 选择“新建资源”。
  2. 输入租户名称、资源名称、订阅名称和 Azure 资源组名称。

现在 Pizza 应用可以使用了。

记录 Pizza 应用的访问值

若要使用新的披萨应用,你需要该披萨应用的应用 ID、授权密钥和授权终结点。 若要获取预测,需要单独的预测终结点和预测密钥。

若要查找这些值:

  1. 在“意向”页面,选择“管理” 。
  2. 在“应用程序设置”页面,记录“应用 ID” 。
  3. 选择“Azure 资源”。
  4. 选择“创作资源”。
  5. 从“创作资源”和“预测资源”选项卡,记录“主键” 。 此值是你的创作密钥。
  6. 记录“终结点 URL”。 此值是你的创作终结点。

以编程方式更改模型

  1. 创建名为 model.py 的新文件。 添加以下代码:

    ########### Python 3.6 #############
    
    #
    # This quickstart shows how to add utterances to a LUIS model using the REST APIs.
    #
    
    import requests
    
    try:
    
        ##########
        # Values to modify.
    
        # YOUR-APP-ID: The App ID GUID found on the www.luis.ai Application Settings page.
        appId = "PASTE_YOUR_LUIS_APP_ID_HERE"
    
        # YOUR-AUTHORING-KEY: Your LUIS authoring key, 32 character value.
        authoring_key = "PASTE_YOUR_LUIS_AUTHORING_SUBSCRIPTION_KEY_HERE"
    
        # YOUR-AUTHORING-ENDPOINT: Replace this with your authoring key endpoint.
        # For example, "https://your-resource-name.cognitiveservices.azure.com/"
        authoring_endpoint = "PASTE_YOUR_LUIS_AUTHORING_ENDPOINT_HERE"
    
        # The version number of your LUIS app
        app_version = "0.1"
        ##########
    
        # The headers to use in this REST call.
        headers = {'Ocp-Apim-Subscription-Key': authoring_key}
    
        # The URL parameters to use in this REST call.
        params ={
            #'timezoneOffset': '0',
            #'verbose': 'true',
            #'show-all-intents': 'true',
            #'spellCheck': 'false',
            #'staging': 'false'
        }
    
        # List of example utterances to send to the LUIS app.
        data = """[
        {
            'text': 'order a pizza',
            'intentName': 'ModifyOrder',
            'entityLabels': [
                {
                    'entityName': 'Order',
                    'startCharIndex': 6,
                    'endCharIndex': 12
                }
            ]
        },
        {
            'text': 'order a large pepperoni pizza',
            'intentName': 'ModifyOrder',
            'entityLabels': [
                {
                    'entityName': 'Order',
                    'startCharIndex': 6,
                    'endCharIndex': 28
                },
                {
                    'entityName': 'FullPizzaWithModifiers',
                    'startCharIndex': 6,
                    'endCharIndex': 28
                },
                {
                    'entityName': 'PizzaType',
                    'startCharIndex': 14,
                    'endCharIndex': 28
                },
                {
                    'entityName': 'Size',
                    'startCharIndex': 8,
                    'endCharIndex': 12
                }
            ]
        },
        {
            'text': 'I want two large pepperoni pizzas on thin crust',
            'intentName': 'ModifyOrder',
            'entityLabels': [
                {
                    'entityName': 'Order',
                    'startCharIndex': 7,
                    'endCharIndex': 46
                },
                {
                    'entityName': 'FullPizzaWithModifiers',
                    'startCharIndex': 7,
                    'endCharIndex': 46
                },
                {
                    'entityName': 'PizzaType',
                    'startCharIndex': 17,
                    'endCharIndex': 32
                },
                {
                    'entityName': 'Size',
                    'startCharIndex': 11,
                    'endCharIndex': 15
                },
                {
                    'entityName': 'Quantity',
                    'startCharIndex': 7,
                    'endCharIndex': 9
                },
                {
                    'entityName': 'Crust',
                    'startCharIndex': 37,
                    'endCharIndex': 46
                }
            ]
        }
    ]
    """
    
    
        # Make the REST call to POST the list of example utterances.
        response = requests.post(f'{authoring_endpoint}luis/authoring/v3.0-preview/apps/{appId}/versions/{app_version}/examples',
            headers=headers, params=params, data=data)
    
        # Display the results on the console.
        print('Add the list of utterances:')
        print(response.json())
    
    
        # Make the REST call to initiate a training session.
        response = requests.post(f'{authoring_endpoint}luis/authoring/v3.0-preview/apps/{appId}/versions/{app_version}/train',
            headers=headers, params=params, data=None)
    
        # Display the results on the console.
        print('Request training:')
        print(response.json())
    
    
        # Make the REST call to request the status of training.
        response = requests.get(f'{authoring_endpoint}luis/authoring/v3.0-preview/apps/{appId}/versions/{app_version}/train',
            headers=headers, params=params, data=None)
    
        # Display the results on the console.
        print('Request training status:')
        print(response.json())
    
    
    except Exception as e:
        # Display the error string.
        print(f'{e}')
    
  2. 将以 YOUR- 开头的值替换为你自己的值。

    信息 目的
    YOUR-APP-ID LUIS 应用 ID。
    YOUR-AUTHORING-KEY 32 字符创作密钥。
    YOUR-AUTHORING-ENDPOINT 创作 URL 终结点。 例如,https://replace-with-your-resource-name.api.cognitive.microsoft.com/。 在创建资源时设置资源名称。

    分配的密钥和资源可以在 LUIS 门户的“Azure 资源”页上的“管理”部分中看到。 应用 ID 可以在“应用程序设置”页的同一“管理”部分中找到。

    重要

    完成后,请记住将密钥从代码中删除,并且永远不要公开发布该密钥。 对于生产来说,请使用安全的方式存储和访问凭据,例如 Azure Key Vault。 有关详细信息,请参阅 Azure AI 服务安全性一文。

  3. 在创建该文件的同一目录中,在命令提示符下输入以下命令来运行文件:

    python model.py
    
  4. 查看创作响应:

    Add the list of utterances:
    [{'value': {'ExampleId': 1137150691, 'UtteranceText': 'order a pizza'}, 'hasError': False}, {'value': {'ExampleId': 1137150692, 'UtteranceText': 'order a large pepperoni pizza'}, 'hasError': False}, {'value': {'ExampleId': 1137150693, 'UtteranceText': 'i want two large pepperoni pizzas on thin crust'}, 'hasError': False}]
    Request training:
    {'statusId': 9, 'status': 'Queued'}
    Request training status:
    [{'modelId': 'edb46abf-0000-41ab-beb2-a41a0fe1630f', 'details': {'statusId': 3, 'status': 'InProgress', 'exampleCount': 0, 'progressSubstatus': 'CollectingData'}}, {'modelId': 'a5030be2-616c-4648-bf2f-380fa9417d37', 'details': {'statusId': 3, 'status': 'InProgress', 'exampleCount': 0, 'progressSubstatus': 'CollectingData'}}, {'modelId': '3f2b1f31-a3c3-4fbd-8182-e9d9dbc120b9', 'details': {'statusId': 3, 'status': 'InProgress', 'exampleCount': 0, 'progressSubstatus': 'CollectingData'}}, {'modelId': 'e4b6704b-1636-474c-9459-fe9ccbeba51c', 'details': {'statusId': 3, 'status': 'InProgress', 'exampleCount': 0, 'progressSubstatus': 'CollectingData'}}, {'modelId': '031d3777-2a00-4a7a-9323-9a3280a30000', 'details': {'statusId': 3, 'status': 'InProgress', 'exampleCount': 0, 'progressSubstatus': 'CollectingData'}}, {'modelId': '9250e7a1-06eb-4413-9432-ae132ed32583', 'details': {'statusId': 3, 'status': 'InProgress', 'exampleCount': 0, 'progressSubstatus': 'CollectingData'}}]
    

    下面是为提高可读性而进行了格式设置的输出:

    Add the list of utterances:
    [
      {
        'value': {
          'ExampleId': 1137150691,
          'UtteranceText': 'order a pizza'
        },
        'hasError': False
      },
      {
        'value': {
          'ExampleId': 1137150692,
          'UtteranceText': 'order a large pepperoni pizza'
        },
        'hasError': False
      },
      {
        'value': {
          'ExampleId': 1137150693,
          'UtteranceText': 'i want two large pepperoni pizzas on thin crust'
        },
        'hasError': False
      }
    ]
    
    Request training:
    {
      'statusId': 9,
      'status': 'Queued'
    }
    
    Request training status:
    [
      {
        'modelId': 'edb46abf-0000-41ab-beb2-a41a0fe1630f',
        'details': {
          'statusId': 3,
          'status': 'InProgress',
          'exampleCount': 0,
          'progressSubstatus': 'CollectingData'
        }
      },
      {
        'modelId': 'a5030be2-616c-4648-bf2f-380fa9417d37',
        'details': {
          'statusId': 3,
          'status': 'InProgress',
          'exampleCount': 0,
          'progressSubstatus': 'CollectingData'
        }
      },
      {
        'modelId': '3f2b1f31-a3c3-4fbd-8182-e9d9dbc120b9',
        'details': {
          'statusId': 3,
          'status': 'InProgress',
          'exampleCount': 0,
          'progressSubstatus': 'CollectingData'
        }
      },
      {
        'modelId': 'e4b6704b-1636-474c-9459-fe9ccbeba51c',
        'details': {
          'statusId': 3,
          'status': 'InProgress',
          'exampleCount': 0,
          'progressSubstatus': 'CollectingData'
        }
      },
      {
        'modelId': '031d3777-2a00-4a7a-9323-9a3280a30000',
        'details': {
          'statusId': 3,
          'status': 'InProgress',
          'exampleCount': 0,
          'progressSubstatus': 'CollectingData'
        }
      },
      {
        'modelId': '9250e7a1-06eb-4413-9432-ae132ed32583',
        'details': {
          'statusId': 3,
          'status': 'InProgress',
          'exampleCount': 0,
          'progressSubstatus': 'CollectingData'
        }
      }
    ]
    

清理资源

完成本快速入门后,请从文件系统中删除该文件。

后续步骤

应用的最佳实践