Ingestion, Ontology and Network Creation

The Ingestion, Ontology and Network Creation configurations in integrally linked. It is helpful to understand how they interact.

The key rules to remember are summarised below:

• Collection default Ingestion configuration: Every Collection has a default Ingestion configuration. When you select a Collection when harvesting or loading documents, the Collection default Ingestion configuration will be applied.

• Collection Ontology: The Collection default Ingestion configuration identifies the Ontology applied to the Collection. The Ontology controls how Text References, Entities and Links are displayed in Documents.

• Networks: The Ingestion configuration identifies the Network(s) created/updated when loading data into the Collection.

• Network Ontology: The Ingestion configuration first used to create a Network sets the Network default Ontology configuration. You cannot change the Network default Ontology. Instead, you need to Re-create the Network.

• Ontology Best Practice: To avoid conflicts, have one Ontology per Project.

The diagram below shows a simple structure with one Collection loading data into one Network.

The diagram below shows loading structured data into a Collection which populates one Network. When loading structured data, you select a Load Structured Data configuration which incorporates an Ingestion configuration.

You can structure an analysis in different ways. For example, you can have:

Understanding how the Ingestion, Ontology and Network Creation configurations interact between Collections and Networks can help you decide how best to structure your analysis.

• Create: When you create a Collection, you choose an Ingestion configuration. This becomes the default Ingestion configuration.

• Load Unstructured Data: When you load unstructured data into an existing Collection (Add Documents or Harvest from the web), the Collection default Ingestion configuration is used.

• Reprocess: When you select the Reprocess option within a Collection, the Text Reference, Entities and Links are removed from the Documents and the Network(s). Text References, Entities and Links identified are regenerated and populated into the Network(s). Useful for updating a Collection and associated Networks with changes to the configurations.

• Ontology: The Ontology configuration linked to the default Ingestion configuration is used to display and edit Entities, Text References and Links.

• Change: You can change the Collection default Ingestion configuration, which will be applied when harvesting or loading documents.

• Load Structured Data (Exception): When loading structured data, an Ingestion configuration is linked in the Load Structured Data configuration. It is applied to any data added to Documents stored in the Collection. It can be different to the Collection default Ingestion configuration.

• Create: The Network Creation configuration linked to the default Ingestion configuration for a Collection is used to create and reprocess a Network.

• Default Ontology: Every Network has a default Ontology configuration. The Ingestion configuration first used to create a Network sets the Network's default Ontology configuration. You cannot change the Network default Ontology configuration. Instead, you need to Re-create the Network.

• Warning: A warning is displayed if you choose to load data into a Network using an Ingestion configuration which links to a different Ontology to the Network default Ontology. Data loaded into a Network using a different Ontology to the Network default Ontology will be displayed using the default Ontology, which may give unexpected results.

• No Collection: When harvesting or loading data, you can choose not to store Documents in the Collection, instead data is published directly to the Network. In this case, you will need to select the Ingestion configuration to apply when loading the data.

Since the Ontology controls how data is visualised in both Collections and Networks, it is important to have a single Ontology applied to one analysis task.

A simple way to avoid conflicts, is to have one Ontology per Project.

The following diagrams illustrate how the configurations interact when you apply different ways to structure your Collections and Networks.

The diagram below shows a simple structure with one Collection loading data into one Network.

The diagram below shows a simple structure with one Collection loading data into one Network.

The diagram below shows two Collections loading data into one Network.

This can be useful when loading data from different sources, which you may want to apply different Ingestion rules to, but still want to analyse in one Network.

The diagram below shows a single Collection loading data into two Networks.

The diagram below shows a more complex structure with multiple Collections loading data into multiple Networks.

The diagram below shows a more complex structure with multiple Collections loading data into multiple Networks.  However, the two Collections have different Ingestion configurations linking to different Ontologies.

The default Ontology applied to Network B will depend on which Collection is used to first create the Network.

Every Collection has a default Ingestion configuration assigned.

You can change a Collection's default Ingestion configuration.