Testing a deployed model

How to test a deployed model

As soon as you deploy one of your models, you can manually test it.

To test a deployed model, proceed as follows:

Navigate to NLU → Deployments section.
Select a deployed model and click on it. The following page opens:

Click on the Test NLU Deployment button.

You can only test running models. If a deployed model is stopped, the Test NLU Deployment button will be greyed out.

The following window opens.

In the Request tab, make necessary changes in the data by directly editing the values:
1. Hover over the value you want to edit and click the Edit icon.
2. Make changes and click the Tick icon.
Click the Test button to start testing.
The response is returned as JSON and displayed in the Response tab. You can copy the response body by clicking the Copy icon.

Below you can find the description of both request and response properties.

JSON request description

Property			Description
`flow_id`			The NLU workflow ID that indicates the engine running an annotation.
`input`			User input.
	`entries`		Utterances input by the user.
`dialog_context`			Dialog context information used to evaluate user input and decide on the next system actions. Part of this context can be referenced in the creation of context sensitive rules.
	`max_nbest`		Reserved.
	`nbest`		Reserved.
	`ambiguous_utterance`		User’s previous utterance(s) in cases the status of the previous step’s intent was ambiguous.
	`system_actions`		System actions.
		`name`	The name of the system action.
		`type`	The type of the system action. The following values are applicable: Ask, Suggestion, Disambiguation, Confirmation, Announcement, BackEndCall
		`subtype`	Defines an explicit intent request. The following subtypes are possible: GetIntent: System requests, confirms, disambiguates an intent. GetEntity: System requests, confirms, disambiguates an entity.
		`tag`	Describes the system action. Typically used to group similar actions together.
		`level`	The error level used to identify the prompt that was used. Applicable values are: normal, error, unrecovered.
		`prompt`	The prompt used for the system action.

JSON response description

Property		Description
`utterance`		Phrase under query.
`intents`		List of intents and their parameters.
	`name`	intent the queried phrase is associated with.
	`source`	The source the intent is originated from. Valid values: ml is Machine Learning, used to expand your model’s understanding using your custom data. rb is rule-based NLU, the out-of-the-box understanding based on the model’s domain.
	`confidence`	The Machine Learning confidence for the intent.
	`score`	A numerical value determining the validity of a constraint for an intent.
	`constraint`	Details about the constraint that matches this intent. Null for ml intents.
	`segment_index`	An indicator used in understanding utterances: if there are no connecting words in the utterance, it returns Null. if there are connecting words in the utterance, depending on the part of the utterance, it returns 1 or 2.
`nbest_intents`		List of nbest intents and their parameters.
	`name`	intent the queried phrase is associated with.
	`source`	The source the intent is originated from. Valid values: ml is Machine Learning, used to expand your model’s understanding using your custom data. rb is rule-based NLU, the out-of-the-box understanding based on the model’s domain.
	`confidence`	The Machine Learning confidence for the intent.
	`score`	A numerical value determining the validity of a constraint for an intent.
	`constraint`	Details about the constraint that matches this intent. Null for ml intents.
	`matched_selection_entities`	Entities returned based on given utterance.