genschema.postprocessing.schema_references¶

Attributes¶

`PathSegment`
`SchemaPath`
`ComparatorFactory`
`SimilarityMetric`
`MergeStrategy`
`NameFactory`
`DEFAULT_COMPARATOR_FACTORIES`
`DEFINITION_SECTION_KEYS`
`STRUCTURAL_CONTAINER_KEYS`
`STRUCTURAL_VARIANT_KEYS`
`MEANINGFUL_PATH_PARTS_BLACKLIST`

Classes¶

`SchemaReferenceExtractionConfig`
`SchemaCandidate`
`CandidateGroup`
`SchemaReferencePostprocessor`	Standalone JSON Schema postprocessor that extracts repeated or highly similar

Module Contents¶

type PathSegment = str | int[source]¶

type SchemaPath = tuple[PathSegment, ...][source]¶

type ComparatorFactory = Callable[[], Comparator][source]¶

type SimilarityMetric = Callable[[frozenset[str], frozenset[str]], float][source]¶

type MergeStrategy = Callable[[list[dict], 'SchemaReferenceExtractionConfig'], dict][source]¶

type NameFactory = Callable[[int, 'CandidateGroup', 'SchemaReferenceExtractionConfig'], str][source]¶

DEFAULT_COMPARATOR_FACTORIES: tuple[ComparatorFactory, Ellipsis][source]¶

DEFINITION_SECTION_KEYS[source]¶

STRUCTURAL_CONTAINER_KEYS = ('items', 'additionalProperties', 'contains', 'if', 'then', 'else', 'not', 'propertyNames',...[source]¶

STRUCTURAL_VARIANT_KEYS = ('anyOf', 'oneOf', 'allOf', 'prefixItems')[source]¶

MEANINGFUL_PATH_PARTS_BLACKLIST[source]¶

class SchemaReferenceExtractionConfig[source]¶

similarity_threshold: float = 0.85[source]¶

min_total_keys: int = 3[source]¶

min_occurrences: int = 2[source]¶

defs_key: str = '$defs'[source]¶

ref_prefix: str | None = None[source]¶

merge_base_of: Literal['anyOf', 'oneOf', 'allOf'] = 'anyOf'[source]¶

merge_pseudo_handler: PseudoArrayHandlerBase | None = None[source]¶

merge_comparator_factories: tuple[ComparatorFactory, Ellipsis][source]¶

similarity_metric: SimilarityMetric[source]¶

merge_strategy: MergeStrategy | None = None[source]¶

name_factory: NameFactory | None = None[source]¶

allowed_root_types: tuple[str, Ellipsis] = ('object', 'array')[source]¶

preserve_common_keywords: bool = True[source]¶

include_root: bool = False[source]¶

skip_existing_definitions: bool = True[source]¶

__post_init__() → None[source]¶

property normalized_ref_prefix: str[source]¶

class SchemaCandidate[source]¶

path: SchemaPath[source]¶

schema: dict[source]¶

type_signature: tuple[str, Ellipsis][source]¶

tokens: frozenset[str][source]¶

total_keys: int[source]¶

class CandidateGroup[source]¶

members: list[SchemaCandidate][source]¶

merged_schema: dict[source]¶

total_keys: int[source]¶

benefit: int[source]¶

definition_name: str = ''[source]¶

class SchemaReferencePostprocessor[source]¶

Standalone JSON Schema postprocessor that extracts repeated or highly similar structures into shared definitions and replaces occurrences with $ref.

The postprocessor is intentionally independent from Converter itself: it can be run on any already-built schema. Candidate groups are merged through a fresh internal Converter run so the resulting definition stays aligned with the project’s normal schema-combination pipeline.

classmethod process( schema: dict, config: SchemaReferenceExtractionConfig | None = None, ) → dict[source]¶

classmethod extract( schema: dict, config: SchemaReferenceExtractionConfig | None = None, ) → dict[source]¶