Identify objects using Insight

Relevant for: GUI tests and components

Insight enables OpenText Functional Testing to recognize objects in your application based on what they look like, instead of properties that are part of their design. This can be useful if you are working with an application whose technology is not supported by OpenText Functional Testing, or with an application running on a remote computer.

In this topic:

Overview

An image of the object is stored with the Insight test object (along with ordinal identifiers, if necessary). This image is used as the main description property to identify the object in the application.

You can create InsightObject test objects during recording sessions, or when adding test objects to an object repository manually.

When adding objects to an object repository, you can add an object directly from the application, or even from a picture of the object displayed on your screen.

In the object repository, you can edit the object's properties, such as its name and image, and the default location to click in the control when performing test object methods. You can also define visual relation identifiers to more accurately describe the object.

Back to top

Creating Insight test objects

OpenText Functional Testing learns Insight objects using the following elements:

Description property	An image of the object is stored and used as the main description property to identify the object in the application. The following additional description properties are also used, to ensure unique identification: similarity: Specifies how similar a control in the application has to be to the test object image for it to be considered a match. skipmakeobjvisible: Specifies whether to skip the step that makes sure the object is visible before trying to identify it. For example, if you see that an attempt to make the object visible causes the application window to close before the object can be identified. methodoptionorder: Specifies the order in which the image matching algorithms should be used. ratiotestthreshold: The threshold of the ratio of a feature's distance from its closest match to the distance of its second closest match.
Ignore areas	If parts of the object do not always look the same, you can specify that those areas be ignored when the image is used to identify the object.
Ordinal identifier	Use an ordinal identifier to create a unique description for the object. Other aspects of object configuration, such as mandatory and assistive properties, and smart identification, are not relevant for Insight test objects.
Visual relation identifiers	After a description is created for an Insight test object, you can add visual relation identifiers to improve identification of the object.

The Insight object is always added to the object repository as a child of the test object that represents its containing application, such as a Window or Browser object. (The new object's parent object is also added if it does not already exist in the object repository.)

Back to top

Creating steps with Insight Test Objects

Create steps using Insight test objects in much the same way as you would with other types of test objects.

In the Editor, the test object image is displayed in the step instead of the test object name. When you hold the cursor over the image, an enlarged view of the image is displayed. Double-click the image to open the object repository with the InsightObject selected.

For details on the methods and properties supported by Insight test objects, see the Insight section of the Object Model Reference for GUI Testing.

Back to top

Running tests with Insight

During a run session, OpenText Functional Testing searches for the Insight test object that matches the image stored with the test object.

OpenText Functional Testing searches for the matching object within the Insight object's parent test object. You can help focus the search on a smaller area by creating smaller parent objects and building a hierarchy of Insight test objects in your object repository.

OpenText Functional Testing's image matching algorithm allows for some variation, enabling OpenText Functional Testing to recognize the object even if it changed slightly. However, the algorithm is not based on object properties, and therefore does not use the smart identification mechanism.

Back to top

See also:

• Record a GUI test or component
• Work with Insight or Text test objects
• Test object descriptions