Validator Interface

The Validator Interface provides access to a number of different validation operations over ontologies.

The notion of validation in OAK is intentionally very flexible, and may encompass:

  • Schema validation, for example, checking definitions are strings and have 0..1 cardinality.

  • Logical validation, using a reasoner.

  • Lexical validation, for example, ensuring there are no spelling errors

  • Stylistic validation, against a style guide

  • Content validation, checking the content of the ontology against domain knowledge or other ontologies.

Different adapters may implement different portions of this.

Schema Validation

The core validate method is configured using a metadata schema. The default one used is:

This is specified using LinkML which provides an expressive way to state constraints on metadata elements, such as AnnotationProperty assertions in ontologies. For example, this schema states that definition is recommended (not required), and that it is single-valued.

Different projects may wish to configure this - it is possible to pass in a different or modified schema

For more details see this howto guide

Warning

Currently only implemented for :ref`sql_implementation`

LLM-based validation

See use_llms

class oaklib.interfaces.validator_interface.ValidatorInterface(resource: ~oaklib.resource.OntologyResource | None = None, strict: bool = False, _multilingual: bool | None = None, autosave: bool = <factory>, exclude_owl_top_and_bottom: bool = <factory>, ontology_metamodel_mapper: ~oaklib.mappers.ontology_metadata_mapper.OntologyMetadataMapper | None = None, _converter: ~curies.api.Converter | None = None, auto_relax_axioms: bool | None = None, cache_lookups: bool = False, property_cache: ~oaklib.utilities.keyval_cache.KeyValCache = <factory>, _edge_index: ~oaklib.indexes.edge_index.EdgeIndex | None = None, _entailed_edge_index: ~oaklib.indexes.edge_index.EdgeIndex | None = None)[source]

Multi-faceted ontology validator.

This interface defines a number of different validation processes

The overall goal is to support the following

  • validating against OntologyMetadata schema

  • lexical checks

  • wrapping reasoning

  • structural graph checks

Specific implementations may choose to implement efficient methods for this. For example, a SQL implementation can quickly determine all terms missing definitions with a query over an indexed table.

Currently the main implementation for this is the SqlDatabase implementation, this implements a generic property check using the OntologyMetadata datamodel

See:

term_curies_without_definitions() → Iterable[str][source]

Yields all entities that do not have a definition.

Returns:

validate(configuration: ValidationConfiguration | None = None) → Iterable[ValidationResult][source]

Validate entire ontology or wrapped ontologies.

Validation results might be implementation specific

  • reasoners will yield logical problems

  • shape checkers or schema checkers like SHACL or LinkML will return closed-world structural violations

  • specialized implementations may yield lexical or other kinds of problems

Returns:

validate_mappings(entities: Iterable[str] | None = None, adapters: Dict[str, BasicOntologyInterface] | None = None, configuration: ValidationConfiguration | None = None) → Iterable[MappingValidationResult][source]

Validate mappings for a set of entities.

Different adapters may implement different aspects of mapping validation.

It includes:

  • checking cardinality of mappings (e.g. skos:exactMatch should be 1:1)

  • looking up mapped entities to check they are not obsolete

  • using AI to validate the content of mappings

Parameters:

  • entities – entities to validate mappings for (None=all)

  • adapters – adapter mapping to look up external entities

  • configuration – validation configuration

Returns:

validate_synonyms(entities: Iterable[str] | None = None, adapters: Dict[str, BasicOntologyInterface] | None = None, configuration: ValidationConfiguration | None = None, synonymizer_rules: RuleSet | None = None) → Iterable[ValidationResult][source]

Validate synonyms for a set of entities.

Different adapters may implement different aspects of synonym validation.

It includes:

  • checking for duplicates

  • looking up mapped entities to check they are not obsolete

  • ensuring that a referenced synonym is still supported

  • using AI to validate the content of mappings

Parameters:

  • entities – entities to validate mappings for (None=all)

  • adapters – adapter mapping to look up external entities

  • configuration – validation configuration

Returns:

validate_definitions(entities: Iterable[str] | None = None, adapters: Dict[str, BasicOntologyInterface] | None = None, configuration: ValidationConfiguration | None = None, skip_text_annotation=False, **kwargs) → Iterable[ValidationResult][source]

Validate text definitions for a set of entities.

Different adapters may implement different aspects of mapping validation.

It includes:

  • testing definitions are present

  • validating the text definition against the SRS rubric

  • using AI to align the definition against any references

Parameters:

  • entities –

  • configuration –

  • kwargs –

Returns:

repair(configuration: RepairConfiguration | None = None, dry_run=False) → Iterable[RepairOperation][source]

Finds problems and fixes them.

Parameters:

  • configuration –

  • dry_run –

Returns:

is_coherent() → bool[source]

True if the ontology is logically coherent, as determined by deductive reasoning (e.g. an OWL reasoner)

Returns:

true if coherent

unsatisfiable_classes(exclude_nothing=True) → Iterable[str][source]

Yields all classes that are unsatisfiable, as determined by deductive reasoning (e.g. an OWL reasoner)

Parameters:

exclude_nothing – if True (default) do not include the tautological owl:Nothing

Returns:

class curie iterator

lookup_references(references: List[str], adapters: Dict[str, BasicOntologyInterface] | None = None, **kwargs) → Dict[str, bool | Dict[str, Any]][source]

Lookup references in external ontologies.

Parameters:

  • references – list of CURIEs to look up

  • adapters – adapter mapping to look up external entities

  • kwargs – additional arguments

Returns:

mapping between CURIEs and pub objects

autosave: bool

For adapters that wrap a transactional source (e.g sqlite), this controls whether results should be auto-committed after each operation

exclude_owl_top_and_bottom: bool

Do not include owl:Thing or owl:Nothing