ETA networksA network is a visual summary, generated by ETA, of the information in one or more documents. can be exported to an xml file then imported into the following IBM i2 applications:
The broad steps are to:
By default, automatic mapping is used. ETA matches its entity types with IBM i2 entity types. You can improve the mapping by ‘influencing’ automatic mapping. This involves changing the ETA network schema to be more similar to the i2 schema. If you are not satisfied with automatic mapping you can manually map entities. However, influencing automatic mapping is the preferred method as it will be more robust when the i2 schema changes.
Once you have created a configuration you can test it using different networks and different IBM i2 schemas. Both the ETA network and i2 Analyze Schema can be changed at any time and the i2 Mappings configuration will remain valid.
The IBM i2 schema can specify that some properties, or a group of properties, are mandatory. Entities cannot be created if a mandatory property has no value. If a configuration maps to such an i2 entity type, but mandatory properties are not mapped, the configuration will display an error and the entity mapping will not be valid.
Note: Even though mandatory properties are mapped, the ETA entity might still not have any data in the corresponding entity property. These entities are still not valid and will be skipped during the export process.
To map ETA network to an IBM i2 application:
Note: The configuration is not network-dependent, but a reference network is needed to confirm that all the network’s types are mapped as expected. If no reference network was selected then only entities in the ETA default schema are listed.
The result of automatic mapping between the entities in the ETA network and those in i2 schema is shown. The entities are listed in alphabetical order.
Each ETA entity has a name and a semantic typeA term used when mapping a ETA network to an IBM i2 schema to describe the type of information an entity represents, for example, Document, Gender, Organisation, Relationship Type or Latitude.. Automatic mapping attempts to match the name and semantic type (first both, then either) with i2 entities. If exactly one i2 entity matches, automatic mapping will map to that entity type.
Once an entity and its properties are mapped, you can test the mapping to see if the export works as intended on the Reference Network.
The ETA network entities that will be mapped to entities in IBM i2 are listed at the top of the pane on the right. The entity you want to preview is selected and the details are shown in the table below. If any warnings or errors are generated by exporting the entities, they are shown.
If you are not satisfied with an entity that was automatically mapped, or an entity was not automatically mapped, you can influence the mapping in several ways.
Note: Renaming a type might reorder the entity list. Scroll to the new name to continue editing.
Once a Sintelix entity is mapped to the correct IBM i2 entity, properties need to be matched. The process is identical to entity matching: if the automatic mapping is not your intended mapping, either force the correct property by selecting it from the list, or modify the semantic type and/or property name to influence the automatic matching.
By default, all fields of the network entity will be listed as Sintelix properties. You can create more properties by using the Sintelix ‘path’ syntax, which gives access to documents and network neighbours in addition to simple data transformations. The same syntax is used in Network Table View configuration. When you create a new property, click the ‘Field Path Usage’ help button to display more information.
You can specify a default value for any property. If it is not empty, this value will be used whenever the Sintelix entity does not have any value to be mapped.
It is particularly useful for mandatory properties because it prevents entities with missing mandatory property from being invalid.
To specify a default value, click the ‘Default Value’ box under the mapped IBM i2 property.
Links are listed in the centre configuration panel after entities, in alphabetical order.
Links are mapped just like entities. However, the IBM i2 schema defines an additional constraint on links by specifying which entity types are allowed on both ends of a link. As a result, one Sintelix link type might not map to one i2 link type, because the Sintelix link type could have unacceptable entities on its ends.
Additional configuration options allow splitting one Sintelix network link type into multiple link types, and then mapping each type separately. There is also additional feedback on the configuration panel to confirm if mapping is acceptable, given expected and allowed entity types.
When editing a Sintelix schema link you can modify which entities are allowed as ‘From’ and ‘To’ targets. For example, the default Association link joins Person to either Organisation or Location, and therefore appears as shown:
However, an IBM i2 Analyze Schema can have two link types: one allowing only Person to Organisation, and another only allowing Person to Location. When trying to map Association to such a link type, a warning is shown:
In this example, ‘From’ End Type is accepted (‘Person’ (green). ‘Network Identity’ would also be acceptable) but ‘To’ End Type accepts ‘Organization’ and rejects ‘Location’ (red).
It's still possible to leave such mapping. Links that are not acceptable by IBM i2 will be not be exported. To see which links will be exported and which cannot be (marked in red, with red icon providing a reason), click ‘Data Preview’.
To export all links you must split Association Links into two different link types. To do this:
At this point, Association is a link type that only links from Person to Organization so it can be mapped to an IBM i2 link type without warnings.
At this point there are two Sintelix link types and both can be mapped independently. Both are derived from the underlying ‘Association’ network type but have different targets.
When splitting a link by its targets the following considerations apply:
In this example, an ‘Employment’ link is between ‘Person’ and ‘Person’, so Entity Types do not imply direction. However, if the two roles are of pattern ‘lawyer-of’ and ‘has-lawyer’, as is typical for Sintelix network links, then patterns ‘*-of’ and ‘has-*’ can be used to define direction.
If direction still cannot be implied, targets are assigned arbitrarily and a warning is produced in the Data Preview table.
The properties of links are mapped identically to the properties of an entity.