簡易回應 (Dialogflow)

在 Dialogflow 中探索

按一下 [繼續]，將回應範例匯入 Dialogflow。接著，請按照 部署及測試範例的步驟：

  1. 輸入虛擬服務專員名稱，並為範例建立新的 Dialogflow 代理程式。
  2. 代理程式匯入完畢後，按一下「Go to agent」。
  3. 在主要導覽選單中，前往「Fulfillment」(執行要求)。
  4. 啟用「Inline Editor」(內嵌編輯器)，然後按一下 [Deploy] (部署)。編輯器包含範例 再也不是件繁重乏味的工作
  5. 在主要導覽選單中，前往「Integrations」(整合)，然後按一下「Google」 Google 助理。
  6. 在出現的互動視窗中，啟用「自動預覽變更」，然後按一下「測試」 以開啟動作模擬工具
  7. 在模擬工具中輸入 Talk to my test app 即可測試範例！
繼續

簡單的回覆會以即時通訊泡泡形式顯示，並使用文字轉語音 (TTS) 或語音合成標記語言 (SSML)。

根據預設，系統會使用 TTS 文字做為即時通訊泡泡內容。如果畫面顯示 文字符合您的需求，因此在即時通訊中無須指定任何顯示文字 泡泡。

或是參閱對話設計指南，瞭解 如何將這些視覺元素整合至動作中。

屬性

圖 1. 簡易回應範例 (智慧型手機)

簡易回應具有下列規定和選填 您可以設定下列屬性：

  • 支援具有 actions.capability.AUDIO_OUTPUT 或 actions.capability.SCREEN_OUTPUT 功能。

  • 每個即時通訊泡泡上限為 640 個字元。超過這個上限的字串 在第一個文字斷行 (或空白字元) 的最前 640 個字元。

  • 即時通訊泡泡內容必須拼音，或是完整轉錄稿 TTS/SSML 輸出內容。幫助使用者清楚掌握您的發言內容，並提高回覆次數 不同條件的理解能力

  • 每個回合最多兩個即時通訊泡泡。

  • 提交給 Google 的即時通訊頭像 (標誌) 必須為 192x192 像素， 且不可為動畫。

圖 2. 簡易回應範例 (智慧螢幕)

程式碼範例

Node.js

app.intent('Simple Response', (conv) => {
  conv.ask(new SimpleResponse({
    speech: `Here's an example of a simple response. ` +
      `Which type of response would you like to see next?`,
    text: `Here's a simple response. ` +
      `Which response would you like to see next?`,
  }));
});
index.js

Java

@ForIntent("Simple Response")
public ActionResponse welcome(ActionRequest request) {
  ResponseBuilder responseBuilder = getResponseBuilder(request);
  responseBuilder.add(
      new SimpleResponse()
          .setTextToSpeech(
              "Here's an example of a simple response. "
                  + "Which type of response would you like to see next?")
          .setDisplayText(
              "Here's a simple response. Which response would you like to see next?"));
  return responseBuilder.build();
}
MyActionsApp.java

Node.js

conv.ask(new SimpleResponse({
  speech: `Here's an example of a simple response. ` +
    `Which type of response would you like to see next?`,
  text: `Here's a simple response. ` +
    `Which response would you like to see next?`,
}));
index.js

Java

ResponseBuilder responseBuilder = getResponseBuilder(request);
responseBuilder.add(
    new SimpleResponse()
        .setTextToSpeech(
            "Here's an example of a simple response. "
                + "Which type of response would you like to see next?")
        .setDisplayText(
            "Here's a simple response. Which response would you like to see next?"));
return responseBuilder.build();
MyActionsApp.java

JSON

請注意，下列 JSON 會說明 Webhook 回應。

{
  "payload": {
    "google": {
      "expectUserResponse": true,
      "richResponse": {
        "items": [
          {
            "simpleResponse": {
              "textToSpeech": "Here's an example of a simple response. Which type of response would you like to see next?",
              "displayText": "Here's a simple response. Which response would you like to see next?"
            }
          }
        ]
      }
    }
  }
}

JSON

請注意，下列 JSON 會說明 Webhook 回應。

{
  "expectUserResponse": true,
  "expectedInputs": [
    {
      "possibleIntents": [
        {
          "intent": "actions.intent.TEXT"
        }
      ],
      "inputPrompt": {
        "richInitialPrompt": {
          "items": [
            {
              "simpleResponse": {
                "textToSpeech": "Here's an example of a simple response. Which type of response would you like to see next?",
                "displayText": "Here's a simple response. Which response would you like to see next?"
              }
            }
          ]
        }
      }
    }
  ]
}

SSML 和聲音

在回應中使用 SSML 和音效，就能讓模型更漂亮、 使用者體驗下列程式碼片段說明如何建立回應 模型使用 SSML：

Node.js

app.intent('SSML', (conv) => {
  conv.ask(`<speak>` +
    `Here are <say-as interpet-as="characters">SSML</say-as> examples.` +
    `Here is a buzzing fly ` +
    `<audio src="https://actions.google.com/sounds/v1/animals/buzzing_fly.ogg"></audio>` +
    `and here's a short pause <break time="800ms"/>` +
    `</speak>`);
  conv.ask('Which response would you like to see next?');
});
index.js

Java

@ForIntent("SSML")
public ActionResponse ssml(ActionRequest request) {
  ResponseBuilder responseBuilder = getResponseBuilder(request);
  responseBuilder.add(
      "<speak>"
          + "Here are <say-as interpet-as=\"characters\">SSML</say-as> examples."
          + "Here is a buzzing fly "
          + "<audio src=\"https://actions.google.com/sounds/v1/animals/buzzing_fly.ogg\"></audio>"
          + "and here's a short pause <break time=\"800ms\"/>"
          + "</speak>");
  return responseBuilder.build();
}
MyActionsApp.java

Node.js

conv.ask(`<speak>` +
  `Here are <say-as interpet-as="characters">SSML</say-as> examples.` +
  `Here is a buzzing fly ` +
  `<audio src="https://actions.google.com/sounds/v1/animals/buzzing_fly.ogg"></audio>` +
  `and here's a short pause <break time="800ms"/>` +
  `</speak>`);
conv.ask('Which response would you like to see next?');
index.js

Java

ResponseBuilder responseBuilder = getResponseBuilder(request);
responseBuilder.add(
    "<speak>"
        + "Here are <say-as interpet-as=\"characters\">SSML</say-as> examples."
        + "Here is a buzzing fly "
        + "<audio src=\"https://actions.google.com/sounds/v1/animals/buzzing_fly.ogg\"></audio>"
        + "and here's a short pause <break time=\"800ms\"/>"
        + "</speak>");
return responseBuilder.build();
MyActionsApp.java

JSON

請注意，下列 JSON 會說明 Webhook 回應。

{
  "payload": {
    "google": {
      "expectUserResponse": true,
      "richResponse": {
        "items": [
          {
            "simpleResponse": {
              "textToSpeech": "<speak>Here are <say-as interpet-as=\"characters\">SSML</say-as> examples.Here is a buzzing fly <audio src=\"https://actions.google.com/sounds/v1/animals/buzzing_fly.ogg\"></audio>and here's a short pause <break time=\"800ms\"/></speak>"
            }
          },
          {
            "simpleResponse": {
              "textToSpeech": "Which response would you like to see next?"
            }
          }
        ]
      }
    }
  }
}

JSON

請注意，下列 JSON 會說明 Webhook 回應。

{
  "expectUserResponse": true,
  "expectedInputs": [
    {
      "possibleIntents": [
        {
          "intent": "actions.intent.TEXT"
        }
      ],
      "inputPrompt": {
        "richInitialPrompt": {
          "items": [
            {
              "simpleResponse": {
                "textToSpeech": "<speak>Here are <say-as interpet-as=\"characters\">SSML</say-as> examples.Here is a buzzing fly <audio src=\"https://actions.google.com/sounds/v1/animals/buzzing_fly.ogg\"></audio>and here's a short pause <break time=\"800ms\"/></speak>"
              }
            },
            {
              "simpleResponse": {
                "textToSpeech": "Which response would you like to see next?"
              }
            }
          ]
        }
      }
    }
  ]
}

詳情請參閱 SSML 參考說明文件。

音效庫

YouTube 音效庫提供了各式各樣的免費簡短音效。這些 音訊會為您代管，因此只要將音訊加入 SSML 即可。