entities.tsv
Provides the list of known entity types.
Name | A machine-readable identifier token, preferably lower snakecase. |
Label | A UI-displayable label. |
Definitional reference | A centrally-registered identifier (CRID) referring to a class in an external ontology that is highly pertinent to the entity type. For example, OBI:0000097 . The regular expression ^[A-Z]{2,12}(:|_)[\dA-Z]+$ is used to match these CRIDs. |
Definition | A natural-language sentence or sentences explaining how, in principle, to identify all instances of this entity type in a given context. |
properties.tsv
Properties can generally be understood as a binary relation between two entities, the entity that “has” the property, and the entity that is the “value”. (That is, we have in mind what are sometimes called functional properties.)
In practice properties can be divided into primitives and more general relations. Relations have general Entity values, of an entity type that is a first-class member of the schema.
For primitives, the entity type for the “value” is simple and universal enough that the character strings representing them leave little doubt as to their basic meaning, as in ordinalities or raw quantities. This framework supports Integer, Float, and String primitives.
String primitives are also often used in cases where:
- The “value” entity will be specified by name or disambiguating description text.
- Other information about the entities of the “value” type is explicitly not part of the schema.
These cases seem to cover all the ground we could reasonable expect from our ADI. So the disjoint list of property “value types” is:
- Integer
- Float
- String
- Entity
properties.tsv
lists the properties that may appear as fields.
Name | A machine-readable identifier token, preferably lower snakecase. |
Label | A UI-displayable label. |
Entity | The entity type to which this property applies. |
Value type | One of “Integer”, “Float”, “String”, “Entity”. |
Related entity | The Name or Label of the entity type which applies to the targets of this property, in case this property is a relation. Otherwise the empty string '' . |
Definitional reference | A centrally-registered identifier (CRID) referring to a property (or class) in an external ontology that is highly pertinent to the property type. For example, OBI:0000097 . The regular expression ^[A-Z]{2,12}(:|_)[\dA-Z]+$ is used to match these CRIDs. |
Definition | A natural-language sentence or sentences explaining how, in principle, to ascertain the value of this property for a given entity. |
values.tsv
Lists “controlled vocabulary” values or terms in the case of properties that admit such values.
Name | A machine-readable identifier token, preferably lower snakecase. |
Label | A UI-displayable label. |
Enumeration | A non-negative integer (i.e. starting with 0, 1, 2, …) identifying the value within the scope of the property. |
Parent property | The Name or Label of the property associated with this value. |
Definitional reference | A centrally-registered identifier (CRID) referring to a class or property in an external ontology that is highly pertinent to the value. For example, OBI:0000097 . The regular expression ^[A-Z]{2,12}(:|_)[\dA-Z]+$ is used to match these CRIDs. |
Definition | A natural-language sentence or sentences explaining how, in principle, to ascertain whether this specific value of the property is the one that a given entity has. |
verbalizations.txt.jinja
Use this file to specify a natural-language sentence, involving the fields of a given record, expressing the factual content of the record.
The document format is illustrated below:
Table name 1:
Sentence ... , ... ... .
Table name 2:
Sentence ... , ... ... .
...
For an example, here is verbalizations.txt.jinja from the pathology schema.