ontocast.onto

Modules:

Name	Description
constants
content_unit
docling_helpers	Helpers for constructing and normalizing DoclingDocument instances.
embedding_policy
enum
facts
graph_prune	Seed-free graph pruning shared by induced-subgraph retrieval and prompt condensing.
iri_policy
llm_graph_payload	Strict LLM wire coercion for RDF graph fields on canonical Pydantic models.
model
namespace_merge	Helpers for merging RDF namespace prefix bindings without silent loss.
null
ontology
ontology_access	Read-only accessors for ontology context on document vs unit workflow state.
ontology_apply	Apply ontology update deltas onto catalog ontologies by namespace ownership.
ontology_condense	Best-effort condensing of an ontology graph before it becomes prompt text.
ontology_header	Lightweight ontology catalog metadata read without materializing graphs.
ontology_operations
ontology_snapshot	Prompt-facing ontology snapshot: graph view with provenance, no catalog identity.
rdfgraph
retrieval_capabilities	Vector retrieval prerequisites for selected_vector_search_ontology context mode.
run_manifest	Per-document record of what a batch run cost and how it was configured.
section_models	Pydantic models for document outlines and section spans.
sparql_models	Pydantic models for graph mutations.
sparse	Backend-neutral sparse vector representation.
state
tenancy	Tenant/project naming helpers for triple-store datasets and vector collections.
token_usage	Provider-reported token usage for a single LLM call.
triple_index	Stable per-triple ids for the graph a critic is asked to review.
unit_states	Dedicated state models for parallel unit loops.
util