OntoCore workspace engine

OntoCore discovers ontology files under a workspace root, parses them, and builds a queryable catalog.

Workspace API (experimental)

use ontocore::workspace::Workspace;

let ws = Workspace::open("/path/to/ontologies")?;

// Catalog stats
let stats = ws.catalog().data().stats();
println!("{} classes", stats.class_count);

// SQL query
let rows = ws.query("SELECT short_name, labels FROM classes")?;

// Entity search (IRI, short name, label)
let hits = ws.search("Person")?;

// Diagnostics from indexing
for d in ws.diagnostics() {
    println!("{:?}: {}", d.code, d.message);
}

Stable since v0.10: Workspace is the recommended high-level API. Lower-level IndexBuilder remains available for custom pipelines.

Lower-level indexing

use ontocore::catalog::IndexBuilder;
use ontocore::query::query_catalog;

let catalog = IndexBuilder::new()
    .workspace("fixtures")
    .build()?;

let result = query_catalog(&catalog, "SELECT * FROM classes")?;

LSP uses document_overrides on IndexBuilder for unsaved editor buffers.

Indexing pipeline

1. Scan — WorkspaceScanner (ontocore-core) walks the workspace with ignore rules.
2. Parse — RDF via Oxigraph; Turtle also bridged to Horned-OWL for axioms.
3. Catalog — entities, axioms, annotations, imports, namespaces assembled in memory.
4. Diagnostics — parse errors + lint rules collected during build.
5. Store — Oxigraph in-memory store for SPARQL (not exposed directly; use sparql_catalog).

Resource limits

Indexing enforces caps on file size, entity count, triple count, and query result rows. See workspace limits.

CLI and LSP entry points

Entry	Command / method
CLI	ontocore query . "SELECT …"
CLI	ontocore validate .
LSP	ontocore/indexWorkspace

The CLI binary is ontocore (ontocore-cli crate). Install with cargo install ontocore-cli --locked.