Skip to content

Locy

Locy is Uni's logic programming layer: Logic + Cypher. It extends OpenCypher with recursive rules, hypothetical reasoning, graph derivation, and abductive remediation workflows.

Who This Is For

  • Application engineers who need graph reasoning beyond one-shot queries.
  • Data and policy teams modeling compliance, access control, and propagation rules.
  • Advanced users who want compiler/runtime internals and TCK-level semantics.

What You Get

  • Declarative rules with recursion (CREATE RULE).
  • Goal-directed evaluation (QUERY).
  • Hypothetical analysis (ASSUME ... THEN).
  • Remediation search (ABDUCE).
  • Derivation/proof explainability (EXPLAIN RULE).
  • Graph materialization (DERIVE).

Start Here

  1. Read Foundations.
  2. Run the Quickstart.
  3. Learn syntax in Language Guide.
  4. Move to Advanced Features.

Locy vs Plain Cypher

Use plain Cypher when you need direct reads/writes on existing graph state. Use Locy when you need:

  • Recursive closure and fixed-point reasoning.
  • Explainable inferred facts.
  • What-if analysis with rollback boundaries.
  • Suggested minimal changes to satisfy or prevent outcomes.

API Entry Points

  • Rust: session.locy(program).await? and the LocyBuilder fluent API via session.locy_with(program).param(...).timeout(...).with_config(...).run().await?.
  • Python (sync): session.locy(program) and the fluent API via session.locy_with(program).param(...).with_config(...).run().
  • Python (async): await session.locy(program) and await session.locy_with(program).run().

Sessions are created from a Uni instance: session = db.session().

See Rust API Integration and Python API Integration.