Omega Ontology: `.bulk` format

"Bulk" format is the standard format for inputting new knowledge into the Omega3 ontology. When we use the term "Bulk" format, we are really using shorthand to mean data that is ready to be indexed relative to preexisting Omega3 entities. The format of this data is tab-separated database-ready files, conformant to the advertised Omega3 schema. For best results, preconvert these data to UTF-8 format.

Bulk format data can be used to directly populate the AttributeType, DataType, EntityType, Keyword, LinkType, and Namespace tables. In practice, however, any new namespaces, datatypes, keywords, link types, entity types, or attribute types can more easily just be added into the global bulk files for those tables. Whether added to the global files or partitioned out, the column format for the data closely follows the database schema for the associated table.

The more common use of bulk format data is to populate the Entity, Link, and Attribute tables, albeit indirectly. All three of these tables are populated by processing a single data file, hereafter referred to as a bulkfile. This data file contains symbolic references to entities, links, and attributes. Processing of the file resolves the entity references, possibly by creating new entities; then it adds appropriate rows to the Link and Attribute tables.

For illustration, consider some rows in a bulkfile that could help to define the concept O3@::|BANANA|, Omega's representation for a banana tree. o3 CL-USER(5): (describe 'o3@::O3@::|BANANA|) O3@::O3@::|BANANA| is a SYMBOL. It is unbound. It is INTERNAL in the O3@ package. Its property list has these indicator/value pairs: SENSOR3::HAS-SUBJECT-DOMAIN (D@::|botany|) SENSOR3::CONCEPT-SENSE (O3@SNS::|banana_tree%1:20:00::| O3@SNS::|banana%1:20:00::|) SENSOR3:DEFINITION ("any of several tropical and subtropical treelike herbs of the genus Musa having a terminal crown of large entire leaves and usually bearing hanging clusters of elongated fruits") SENSOR3:SOURCE (:WORDNET2.0) SENSOR3:DIRECT-SUPERCLASS (O3@::|herb|) SENSOR3:DIRECT-MEMBER-OF (O3@::|Musa|) SENSOR3:DIRECT-SUBCLASS (O3@::|plantain

(a) O3@::|banana| HAS-SUBJECT-DOMAIN D@::|botany|
or, informally, "the concept banana tree belongs to the subject domain botany"

(b) O3@::|banana| SOURCE :WORDNET2.0
or, informally, "the concept banana tree came from the source WordNet 2.0"

The bulkfile has five columns:

Relation Name. This column should contain either a defined linkType name or a defined attributeType name. Recall that in Omega3, a "link" relates one entity (concept, lexical item, or sense object) to a second entity; an "attribute" is a relation between an entity and a simple literal, such as a number, string, keyword, etc. In fact (a), the relation name is "HAS-SUBJECT-DOMAIN"; for fact (b), it is "SOURCE"
Source Entity name. This column is the entity name of the source entity. For both facts above, this is "banana".
Source Namespace. This column represents the namespace of the source object of interest. Our concept "banana" (= banana tree) is not the same as the word "banana". We use namespaces to differentiate these two (as well as any other distinct entity that might be also called "banana"). The source namespace for this concept is "O3".
Target Entity name. This column is polymorphic. When the relation is a linkType, and thus defines an entity/entity relationship, this field contains the entity name of the relationship target. In fact (a) above, the TargetEntity name is "botany", i.e., the entity name of the relationship target. However, when the relationship is an attributeType, and thus defines an association between the entity and a literal value, this field contains (a representation of) that literal value. In fact (b) above, the TargetEntity name is "WORDNET2.0", the string representation of the keyword.
Target Namespace. This column's interpretation works in parallel with the Target Entity name above. If the relation is a linkType, then this field contains the namespace, as would be expected. In fact (a), the target namespace is D, so this would contain "D". However, if the relation is an attributeType, this the row defines an association to a literal value with no namespace, and this column is thus ignored. By convention, the dummy values "X" or "NULL" in such cases.

Omega Ontology: .bulk format

Omega Ontology: `.bulk` format