Configuring Index Units

The Storage menu of Data Factory Studio lets you configure how to store data in the Semantic Graph Index.

This task shows you how to:

Add a New Index Unit

The first step in the Data Factory Studio configuration is to create an index unit. Index Units are storage units in the Semantic Graph Index.

You can consider the Index Unit as the container of your Data Model classes that you define next, in the Data Model tab.

Note: The 3DSpace crawler and the 3DPlan crawler have their own index unit in read-only mode.

From the left panel, select Storage.
The Add new storage dialog box opens to let you create a new storage or import an existing storage configuration (see Exporting and Importing a Storage Configuration.)
Click Create Storage.
Enter a Storage name.

It is possible to edit this name later.
For Storage type, select Semantic Graph Index Unit.
Click Create.
The Index Unit Storage configuration menu opens.
Tip: In the newly created index unit, it is often useful to add a description to indicate what is its use and goal.

Recommendation: Click Save after any editing operation.

The second step of the Storage configuration consists in the definition of classes.

Note: You can also import classes from an ontology defined in Ontology Editor as described in the next section.

Select an Index unit and go to the Data Model tab.
Optional: If there is no existing Package, click to create one.
Define the class.
1. Enter the Class Name, for example car.
2. Select a Parent class, for example, core.Item.
3. Select a Package Name, for example, org.common

Define the class properties, with pairs of Name and Type, and optionally, Annotation.

For example, brand (string), number of wheels (integer), creation (date).


Annotation	Description
Default value	Defines the default value of an attribute if nothing is specified. For example, in a CSV file, if an attribute value is not valuated, the index unit item uses this default value.
Origin	Defines the origin of the attribute or the class with the fields: Source: The name of the source (for example, ontology). URI: The Uniform Resource Identifier of the current attribute or class.
Full Text Search	Pushes the attribute value in one or multiple attributes in the `text` field.
Index Exact	Defines how to index the tokens of the current attribute. With this annotation, the index takes the exact value of the attribute.
Index Lowercase	Defines how to index the tokens of the current attribute. With this annotation, the index takes the lowercase value of the attribute. In other words, it removes capital letters.
Index Normalized	Defines how to index the tokens of the current attribute. With this annotation, the index takes the normalized value of the attribute. In other words, it removes accents and capital letters.
Precision	Defines the number of precision digits for a decimal attribute.

You can import classes from an ontology defined in Ontology Editor.

Before you begin: An ontology is defined in Ontology Editor.

Select an Index unit and go to the Data Model tab.
Under Packages, click New from Ontology.
The New from Ontology dialog box appears.
Expand Ontology and select the ontology you want to import classes from.
The dialog box refreshes to display all the classes contained in the ontology.
Select a class.
The dialog box refreshes to display the properties contained in the selected class.
Select the class properties you want to import.
Click Create Class.
Data Factory Studio adds the class and its selected properties to the Data Model.
Optional: Repeat the previous steps to add more classes.

After configuration changes, you can clear all indexed documents from the Index Unit before reindexing them.