官术网_书友最值得收藏!

Using entities

In the First-President intent, you noticed that the intent had a problem answering the question who is the second president of USA? because the response was static, only replying with the first president, George Washington, and not the second president, John Adams, regardless of which ordinal positions were given by the user. Let's address this issue by dynamically creating the response using an entity:

  1. Let's start by creating an entity called president-number. In the first column, add the presidents' names, and in the second column, add the ordinals that match the president. The following screenshot shows the president-number entity:
Creating a president-number entity
  1. Now, create a new intent called Find President. Enter who is the first president of USA? in the User says textbox and then highlight the words first president and you will see that a drop-down pops out, where you can search and select the entity that you created, called president-number. The following screenshot shows the process of mapping the entity to the highlighted word:
Mapping an entity to a word

By accessing the entity value within the response, you have the ability to create a dynamic response that can map the ordinal value to the president's name. $president-number.original refers to the user's ordinal input, first president, and $president-number refers to the ordinal value mapped to the name George Washington in the president-name entity. The response The $president-number.original of $geo-country is $president-number creates a proper answer to the user's question about who was the first or second President of the USA. When you test this in Google Assistant in Dialogflow, you will see that the proper response gets returned: The second president of USA is John Adams.

  1. Now, create a second question that does not have a geo-country, USA: Who is first president?, and create an equivalent response The $president-number.original is $president-number. The FaqChatBot agent will be smart enough to figure out that if the geo-country is omitted in the question, the agent will use the response that does not have $geo-country. The following screenshot shows the completed Find President intent:
Find President intent
主站蜘蛛池模板: 潢川县| 青田县| 华蓥市| 沈丘县| 阳江市| 左权县| 鄄城县| 仙居县| 彩票| 崇左市| 石首市| 曲阜市| 平遥县| 安顺市| 彰化县| 隆德县| 吴江市| 弥勒县| 晋城| 宾阳县| 乐陵市| 三河市| 巫山县| 阿勒泰市| 轮台县| 隆尧县| 安泽县| 木里| 廊坊市| 佛坪县| 通海县| 封开县| 晋江市| 曲水县| 兴安县| 龙井市| 广西| 古田县| 龙胜| 舞钢市| 翼城县|