Vadalingo is deterministic: each endpoint is a structural translation of a
formal input. There is no natural-language endpoint — free-text descriptions are
turned into Vadalog by the AI agent, which
authors the concept inline and checks it with
validate_concept.SQL to Vadalog
Translate a SQL query into an equivalent Vadalog program.Example
RDF to Vadalog
Translate RDF triples into Vadalog rules.Example
OWL to Vadalog
Translate an OWL ontology into Vadalog rules. Accepts RDF/XML, Turtle, N3, and N-Triples serializations. Whendata_base_path is provided, Parquet datasets are written under disk/ and @bind annotations are injected into the output program. When add_concepts is true, the resulting program is saved as a new concept named owl_to_vadalog in the project (overwriting any existing concept with that name).
Example
concept_id is null unless add_concepts is true, in which case it contains the ID of the saved or updated concept.
