How to create and execute a decisiontree

Goal

You want to select and run a decision tree.

Use the AQ_DecisionTreeSelector to search for decision tree symptoms and use the AQ_DecisiontreeEvaluator to execute a decision tree.

This recipe will start by creating the various containers and decision tree before creating the decision tree page flow.

Create a decision tree

Create a new decision tree in Studio.
Enter name and display name.
Close the general tab.
Start building the decision tree by selecting an attribute for the first question ( condition node ).
Please note attributes used in decision trees must belong to a singleton entity
For each possible answer you can select another node via the resource centric controls.
A decision tree must end with a result node (container, asset, tree or action).
In order to be able to search for a decision tree you must enter one or more symptoms in the symptoms tab on the right.
Save and close the decision tree when it is finished.

Create a decision tree selector

Create a new container.
Enter the name.
Select AQ_DecisionTreeSelector as the container type.
Close the general tab.
Enter the name of an attribute which holds the search query.
This attribute must belong to a singleton entity and will be placed on the page by the AQ_DecisionTreeSelector container.
Enter the names of the buttons that will be used to search for (matchbutton) and select (selectbutton) a decision tree.
The AQ_DecisionTreeSelector uses a trigram matching algorithm to determine which decision tree symptom matches with the search query and calculates an overall score for each decision tree.
Optionally you can enter a threshold ( percentage ) or a max number of results that are displayed.
Save and close the container when it is finished.

Create a decision tree evaluator

Create a new container.
Enter the name.
Select AQ_DecisionTreeEvaluator as the container type.
Close the general tab.
Enter the names of the buttons that will be used to display the continue and retry button.
Optionally you can enter the name of the decision tree if you would like to start a decision tree immediately without using a AQ_DecisiontreeSelector.
Optionally you can choose to not ask prefilled questions by setting the skip-filled-question parameter.
Optionally you can show derived answers by setting the show-implicit-answers parameter.
Optionally you can display answered questions as read-only with the answers-readonly parameter.
Optionally you can display action attributes with the show-attribute-solution parameter.
Save and close the container when it is finished.

Decision tree flow

Now that you have created all individual components you can create the following flow, to search for and execute a flow.
Create a page containing the decision tree selector. In our example we named this page advice.
Create a sub flow to evaluate the tree.
In the sub flow create a page (on which the decision tree evaluator is placed) and a condition node to fetch the system.event.
Please note: the decision tree evaluator still uses actions to navigate a flow. Which actions are available are described in the reference guide.
Save the flow.
Test your decision tree in the runtime.