Create and manage knowledge graph spaces

Permissions

• Super administrators and system administrators can create and manage knowledge graph spaces.

• Space administrators can manage the knowledge graph spaces assigned to them.

Limits

A maximum of 50 knowledge graph spaces can be created per tenant.

Create a space

1. On the Dataphin home page, choose Knowledge Graph > Space Management from the top menu bar.

2. On the Space Management page, click Create Knowledge Graph Space.

3. In the Create Space dialog box, configure the parameters.

   Parameter	Description
   Knowledge Graph Space Name	The name of the knowledge graph space. The name must be unique within the tenant and contain 2 to 128 characters.
   Logo	Select an icon for the space to help with business classification.
   Description	Enter a brief description of the space to provide information such as its background. The description can contain up to 512 characters.
   Datasource	Select a data source to store entity and relational data. You can select Neo4j.
   Database	Select a database that corresponds to the data source. A database can be attached to only one space.
   Vector Model	This model is used for semantic search and intelligent recognition in smart applications. It converts data, such as text and images, into data vectors that a computer can understand. This process enables precise retrieval based on meaning instead of keywords. The system creates a system property named __system_embedding__ in each entity and relationship. This property is used for subsequent search matching and AI chat. You can select a vector-supported model that has been configured in the knowledge graph model configuration assistant. To find this, go to the Super X > Smart Application Management > Smart Application page.
   Vector Dimensions	The number of components required to fully express a vector. For example, if you view a vector as an ordered list of numbers, the number of items in the list is the vector's dimension. Select the vector dimensions based on the vector model.
   Vector Index Configuration	The system displays the vector index configuration. It cannot be modified. Distance Measure: An algorithm used in vector retrieval to calculate the similarity between two vectors. In a vector space, a smaller distance or a smaller angle indicates greater semantic similarity. The default value is Cosine. HNSW parameter m/HNSW parameter ef_construction: Parameters used for index building. `m` determines the maximum number of connections a single node can have in the graph. `ef_construction` determines the size of the candidate set for the nearest connections. The values are fixed at 16 and 200. This configuration uses less memory and builds indexes faster, but the recall rate is slightly lower.

4. Click OK to complete the knowledge graph space configuration.

   Click OK and Configure Graph Model to go to the model design page to configure entities and relationships.

Manage spaces

The system displays the spaces that the current user has permission to view.

1. Each space is displayed as a card that shows its name, logo, description, space administrator, and creation time.

2. You can search for spaces by keywords such as their names.

3. You can perform the following operations on a target space.

   Operation	Description
   Manage Knowledge Graph Model	Click Knowledge Graph Model to go to the model design page, where you can create entities and relationships. For more information, see Entity Types and Relationship Types.
   Manage Knowledge Graph	Click Knowledge Graph to go to the Knowledge Management page, where you can manage entities and relationships. For more information, see Knowledge Management.
   Space Configuration	Click Space Configuration to modify the configured space information. The vector model cannot be modified. Important If you change the data source, existing data is not automatically migrated to the new data source. To keep the data, manually export and migrate it to the new data source before you modify the configuration.
   Member Management	Click Member Management to add members to the space for daily maintenance, management, and other operations. For more information, see Add and manage space members.
   Delete	Click the Delete icon in the upper-right corner of the card. In the Delete Space dialog box, enter the name of the current space and click OK to delete the space. Important If you select Also delete the associated data in the graph engine, the models, relationships, and other data stored in the data source are also deleted. Deleting a space also deletes all associated ontology models, mapping tasks, version histories, and member permissions. This operation cannot be undone. Proceed with caution.