Testing

This article is a part of a tutorial about creating a bot which reports the current weather in any city of the world.

Let’s test the bot before deploying it to make sure we haven’t made any mistakes in the script.

.xml tests

Create a file called test.xml in the test folder:

<test integration="false">

    <test-case id="1">
        <q>/start</q>
        <a state="/Start"/>
    </test-case>

    <test-case id="2">
    <mockData>
        <query method="get">http://api.openweathermap.org/data/2.5/weather?APPID=***&amp;units=metric&amp;lang=en&amp;q=London</query>
        <response type="json" status = "200">{"coord":{"lon":37.62,"lat":55.75},"weather":[{"id":800,"main":"Clear","description":"clear sky","icon":"01n"}],"base":"stations","main":{"temp":9.82,"feels_like":4.17,"temp_min":9.44,"temp_max":10,"pressure":1001,"humidity":31},"visibility":10000,"wind":{"speed":4.13,"deg":185},"clouds":{"all":0},"dt":1603894949,"sys":{"type":3,"id":2018597,"country":"RU","sunrise":1603859255,"sunset":1603893527},"timezone":10800,"id":524901,"name":"London","cod":200}</response>
    </mockData>

        <q>/start</q>
        <a>Hello! I am a virtual assistant. I can tell you the current weather in any city. Just type the city.</a>
        <q>London</q>
        <a>Today in London: clear sky, 10°C</a>
    </test-case>


    <test-case id="3">
        <q>/start</q>
        <a state="/Start"/>
        <q>Tell me something else</q>
        <a state="/CatchAll"/>
        <q>London</q>
        <a state="/GetWeather"/>
    </test-case>

</test>

Click the Save the script icon on the right side of the upper panel to save your script.

Each test-case checks if the actual response of the bot to a specific question meets client expectations. A user reply is usually identified with the <q> tag, a reaction is identified as <a>.

We can assign id to each test-case:

<test-case id="1"> — checks the main script of the bot, or rather, whether the bot gets into the GetWeather state after launch.
<test-case id="2"> — checks the accuracy of the function that handles the API response.

We will use <mockData> tag for this test case. It allows simulating a response from an external service.
Inside <mockData> tag we will make an API call using <query method="get"></query> tag. URL inside <mockData> must be exactly the same as the URL written in the script. Instead of asterisks in APPID = ***, insert your API key.
Copy the example of API response from Logs tab or use log(toPrettyString(response)) method to log the response in the script. Insert the response inside <response type="json" status="200"></response> tag.
To prohibit test scripts from sending real HTTP requests, set the integration="false" flag in <test> flag.

caution

In xml-tests, it is important to remember to escape special characters: &, '', and <>. For example, in our URL ampersand & is escaping to &.

<test-case id="3"> — checks that the bot switches to CatchAll state when a user sends a text that differs from all the considered options.

tip

Learn more about bot script testing and <mockData> tag

Let’s start the bot using the test widget and ask it to report current weather in New York:

Test widget

Now let’s check that the CatchAll state is triggered by a message that differs from options provided for in the script

Test widget

Now you can connect the channel yourself and ask the bot about current weather in any city.

.xml tests​

Test widget​

.xml tests

Test widget