Operations preset

The operations preset adds three entity types to the default schema: Facility, Service and Occupation. It fits schemas about sites, equipment, service delivery and staffing. The preset adds no relations.

Import and name

OPERATIONS is a constant in agrag.common.data_models. Its name is operations, which is the name that load_schema accepts.

from agrag.common.data_models import OPERATIONS

print(OPERATIONS.name)

operations

Entity types added

Type	Description
Facility	A place, equipment, or service that supports a specific function.
Service	A service provided by an organization, such as delivery or printing.
Occupation	A profession, which may need long training or a formal qualification.

The three types have no parents and no properties.

Apply the preset

load_schema("operations") and OPERATIONS.apply(DEFAULT_SCHEMA) return the default schema with the three types added. The name of the result is default+operations. load_schema({}, extends=OPERATIONS) builds a schema named custom on top of the preset. The relations example below uses the same extends form.

from agrag.common.data_models import DEFAULT_SCHEMA, OPERATIONS
from agrag.schema import load_schema

by_name = load_schema("operations")
by_apply = OPERATIONS.apply(DEFAULT_SCHEMA)
by_extends = load_schema({}, extends=OPERATIONS)

for schema in (by_name, by_apply, by_extends):
    print(schema.name, len(schema.entities), len(schema.relations))

default+operations 12 8
default+operations 12 8
custom 12 8

apply returns a new schema and does not change DEFAULT_SCHEMA. It raises ValueError when the schema already declares one of the preset labels, so apply the preset once. Use a preset lists the same limits.

Size of the resulting schema

The default schema has nine entity types and eight relation types. With the operations preset, the schema has 12 entity types and 8 relation types.

Relations for the new types

The default relations do not accept the new types. A LOCATED_IN pair from Facility to Location is rejected, and the same pair from Organization is accepted because LOCATED_IN already allows Organization to Location.

from agrag.common.data_models import DEFAULT_SCHEMA, OPERATIONS

schema = OPERATIONS.apply(DEFAULT_SCHEMA)
print(schema.allows("LOCATED_IN", "Facility", "Location"))
print(schema.allows("LOCATED_IN", "Organization", "Location"))

False
True

add_patterns adds Facility to the source types of LOCATED_IN and Service to the targets of SUPPLIES. HOLDS_OCCUPATION is a new relation from Person to Occupation.

from agrag.common.data_models import OPERATIONS
from agrag.schema import load_schema

schema = load_schema(
    {
        "relations": [
            {"label": "LOCATED_IN", "add_patterns": [["Facility", "Location"]]},
            {"label": "SUPPLIES", "add_patterns": [["Organization", "Service"]]},
            {
                "label": "HOLDS_OCCUPATION",
                "description": "A person has a profession.",
                "patterns": [["Person", "Occupation"]],
            },
        ]
    },
    extends=OPERATIONS,
)

print(schema.name, len(schema.entities), len(schema.relations))
for relation in schema.relations:
    if relation.label in ("LOCATED_IN", "SUPPLIES", "HOLDS_OCCUPATION"):
        print(relation.label, relation.patterns)
print(schema.allows("HOLDS_OCCUPATION", "Person", "Occupation"))
print(schema.allows("SUPPLIES", "Organization", "Service"))

custom 12 9
LOCATED_IN [('Organization', 'Location'), ('Event', 'Location'), ('Project', 'Location'), ('Facility', 'Location')]
SUPPLIES [('Organization', 'Product'), ('Organization', 'Service')]
HOLDS_OCCUPATION [('Person', 'Occupation')]
True
True

add_patterns adds pairs to a base relation, and it keeps the patterns that the base already declares. A new relation label needs patterns, and the schema rejects a relation that names a type it does not declare.

Related

• Use a preset shows how to add the operations types to a graph with Graph.open.
• Default schema lists the default entity types and relation types that the preset extends.
• Ontologies explains how an ontology maps to a schema and the formats that define one.