Xatkit's Bot Testing Tools

This repository contains utility classes to do advanced testing on Xatkit chatbots.

Installation

• You need to have Xatkit installed locally to use this project.

• Install this project in your local maven repository:

  git clone https://github.com/xatkit-bot-platform/labs-bot-testing-tools
  cd labs-bot-testing-tools
  mvn clean install

• Add the dependency to your bot project:

  <dependency>
      <groupId>com.xatkit</groupId>
      <artifactId>labs-bot-testing-tools</artifactId>
      <version>0.0.1-SNAPSHOT</version>
  </dependency>

How to use it

IntentMatchingTesting

This test is used to check if the detected intent for a given utterance is the expected one (i.e. to test if a chatbot detects the correct intents). This code snippet can be used as an example:

public void intentsTest() {
    testIntentMatching(
            xatkitBot,
            "myFile.csv",
            state);
}

• xatkitBot must be a valid com.xatkit.core.XatkitBot

• state must be a valid com.xatkit.execution.State, containing a set of intents that can be recignized through it (alternatively, instead of passing a state as a parameter, you can pass a list of com.xatkit.intent.IntentDefinition directly)

• "myFile.csv must be a csv file (located in the src/test/resources/ folder of your project) containing a table with this structure:

utterance	expected_intent	detected_intent	expected_parameters	detected_parameters
Hello	Greetings
How are you?	HowAreYou
My name is John and I live in Barcelona	Presentation		name = John; city = Barcelona

Fill the expected_parameters column if there is any parameter to be matched, so you can later compare them with detected_parameters

After running this code, a file myFileResult.csv will be created in the src/test/resources/ folder of your project, filling the missing values of the detected_intent column. If the detected intents match with the expected intents, then your bot does what you want!

📚 Note that the intent detection will depend primarily on the intent recognition provider you use within your bot (e.g. Dialogflow or NLP.js)