tempus-engine 0.1.0

Deterministic rule execution for decision systems
Documentation

Tempus Engine

Deterministic rule execution for decision systems.

Built on jsonlogic-fast for portable, high-performance JSON-Logic evaluation.

What Tempus adds over raw JSON-Logic

jsonlogic-fast Tempus Engine
Evaluate a rule against context Execute rules with metadata and validation
Raw serde_json::Value output Structured decision results
Performance-first Decision-systems-first
Stateless evaluation Foundation for audit, explain, governance

Use cases

  • Fraud scoring — block, review, or allow transactions based on risk signals
  • Credit eligibility — approve/deny with deterministic, auditable logic
  • Dynamic pricing — compute fees and discounts from rule definitions
  • Policy enforcement — evaluate compliance rules across regulated workflows
  • Feature flags — gate features with structured rule logic

Quick start

Prerequisites

  • Rust 1.75+
  • Python 3.10+
  • uv

Build and test

make setup        # Build Python bindings
make test         # Run Rust core tests
make test-python  # Run Python e2e tests (29 tests)
make test-wasm    # Run WASM runtime tests (15 tests)
make ci-local     # Full quality gate

Architecture

┌──────────────────────────────────────────────────┐
│              Tempus Engine (this repo)            │
│  Rule metadata · Rich validation · Domain demos  │
│  Foundation for: explain · audit · time layers   │
└────────────────────┬─────────────────────────────┘
                     │ depends on
┌────────────────────▼─────────────────────────────┐
│           jsonlogic-fast (separate repo)          │
│  Fast, embeddable, cross-runtime JSON-Logic eval  │
│  Rust core · Python bindings · WASM bindings      │
└──────────────────────────────────────────────────┘

Repository structure

Tempus-Engine/
├── engine/               # Rust crate: tempus-engine
│   └── src/
│       ├── lib.rs        # Re-exports jsonlogic-fast + Tempus extensions
│       └── metadata.rs   # Rule metadata (name, version, tags)
├── examples/             # Domain-specific usage examples
│   └── python/
│       ├── fraud_scoring.py
│       └── eligibility.py
├── docs/                 # Product documentation
├── tests/
│   └── python/           # Python e2e tests
└── Makefile

Roadmap

Phase What Status
1 Rule metadata, rich validation, domain demos In progress
2 Explain mode (decision trace / audit trail) Planned
3 Time layer (effective-date rule selection) Planned
4 Decision database (receipts, replay, hashing) Planned

License

AGPL-3.0-or-later — commercial licensing available.

The underlying evaluation engine (jsonlogic-fast) is MIT/Apache-2.0.