Ontology Builder

Ontology Builder is a workspace where you can visually define and manage entities and relationships.

Screen composition

View mode

You can switch between the two screen modes in the top header.

• Graph: A pictorial canvas screen represented by nodes and connecting lines. It is suitable for examining or designing the entire structure.
• Table: A screen that shows entities and relationships as a list. Used when managing a large number of items or viewing properties in detail.

Navigation panel

It is on the left and provides the template used to create entities.

• You can quickly create entities by dragging and dropping pre-made templates such as Person, Organization, Product onto the canvas.

Canvas

This is the central work area where you can freely place and join entity nodes.

• Move: Drag the node to move it to the desired location.
• Connection: When you drag the connection point (handle) of a node and drop it on another node, the Create Relationship window opens.
• Zoom in/Zoom out: Use the mouse wheel or the adjustment buttons on the screen.
• Auto Layout: You can use the function to automatically arrange nodes in the top menu.
• Relationship Highlight: When you hover or click on a relationship line (connection line), only the relationship and the entities at both ends are clearly highlighted. Neighbor relationships attached to the same entity are displayed in light color, and the remaining nodes and relationships are displayed dimly. This is convenient when focusing on only one specific connection (1-hop) in a tangled graph.
• Entity Group View (Partial Graph): When you assign a group to an entity, a group selection menu appears at the top left of the canvas. When you select a specific group, the entities in that group are displayed as central nodes, and the connected relationships and counterpart entities are displayed together as background context nodes. Select All groups to return to the original screen. Groups are assigned only to entities, not to relationships themselves.

Properties panel

Located on the right, edit the details of the selected entity or relationship.

• Name notation: Alias is displayed in large letters as the representative name in the property panel and node name tag, and the system name is displayed as a small subtitle. If there is no alias, only the system name appears.

Search

• Current screen search: Quickly find entities and relationships in the canvas you are currently viewing with Ctrl/Cmd + F.
• Open Global Search: Click the search icon in the modeling list header to immediately open Hub Global Search (Cmd+K). You can find not only entities and relationships from other models, but also data assets and agents at once, and clicking on an item takes you directly to the corresponding modeling location (deep link).

Validation

Check the validation panel for errors or warnings in the ontology definition. Catch in advance where necessary attributes are missing or relationships are incorrectly defined.

Create entity

1. Use a template: Drag and drop the desired template from the left panel onto the canvas.
2. Create it yourself: Right-click on an empty space on the canvas or click the + Entity button at the top.
3. Settings: Define and save name, description, properties (columns), etc.

Relationship settings

1. Hover the mouse over the connection point (handle) on the border of the starting entity (Source).
2. Press and drag and place it on the target entity (Target).
3. In the Create Relationship window, enter a relationship name (e.g. PURCHASED, WORKS_FOR) and save.

Schema information — identification key & display column

The Identity Keys and Display Column fields appear in the property panel of an entity/relationship.

• Identity Keys: A key column that distinguishes each item without overlapping. The entity/relationship input/output of the pipeline must be specified for it to work.
• Display Column: This is the column to be used as the representative name tag in the global search/trace/property panel. If not specified, the system name is used as a fallback.

For detailed settings, refer to the Schema Editor document.

Load data

Upload a CSV, JSON, or Parquet file from the Data tab of an entity or relationship. To load data with a pipeline, connect a dataset to a code node and connect the code output to the entity or relationship. Pipeline fields are not mapped automatically, so make the code output's column names and types match the ontology schema.

AI automatic generation

Descriptions of entities and relationships can be automatically generated with AI. Select an entity or relationship and click the AI ​​button in the properties panel to create a description based on the name and properties. The builder's auto-suggestion panel works the same way.

Versions and collection membership

• Version history: Every change is recorded and displayed in the version history. The current timeline is read-only and cannot be restored like a general asset. See Version history for differences by asset.
• Collection affiliation required: When creating or moving an entity or relationship, the collection affiliation must be specified. Builders opened and accessed from the collection tree will automatically have the corresponding collection applied.

Export/Import

Ontology definitions can be exported to a file or imported from an external file.

• Export (JSON): Saves the entities, relationships, and properties (schema) of the current collection as a JSON file.
• Import (JSON · TTL/Turtle): You can import not only JSON definition files but also Turtle(TTL/RDF) files. OWL classes in the TTL are replaced with entities, data properties are replaced with entity columns, object properties are replaced with relationships, rdfs:label becomes an alias, and rdfs:comment becomes a description.
• Automatically specify the collection to import: Imported items will go into the now selected collection regardless of the value written in the file. If an entity/relationship with the same name already exists, it is skipped.
• Progress display: During import, the following steps are displayed in real time: Validation → Entity → Relationship → Refresh along with Processed / Total and the name of the item being processed now.

Self-referential relationship

An entity can also define a relationship with itself. For example, it is used when the MANAGES relationship of the Person entity points to the same Person entity.

table view

In table view, you can manage entities and relationships in the form of a list.

Entity Table

• Search: Search by entity name.
• Sort: Click on the column header to sort.
• Batch Edit: Select multiple entities at once to edit or delete them.

Relationship table

• Source/Target Filter: Shows only relationships associated with a specific entity.
• Relationship Type Filter: Shows only certain types of relationships.

Keyboard shortcuts

Shortcut keys	Action
Delete / Backspace	Delete selected item
Ctrl/Cmd + Z	Undo
Ctrl/Cmd + Shift + Z	Redo
Ctrl/Cmd + S	Save
Ctrl/Cmd + A	Select all
Space (hold)	Pan Mode (Screen Movement)

Click the ? icon at the top to see the full list of shortcuts.

Practical example

• covid19 Ontology Tutorial — A 20-minute walkthrough that follows the functions of this page from beginning to end using Nowon-gu data.