Documentation
Everything you need to build streaming temporal event processing pipelines with tflo: model your domain events, then layer temporal analysis on top.
The engine
Event-time
Processing-time vs event-time, lateness, replay.
Mental model
tflo vs React state, Signals, RxJS, XState.
Provable temporal logic
Deterministic, testable temporal logic.
Production superpowers
What-if, roll-back, replay, audit.
Core Concepts
Computation graphs, properties, timestamps, keyed execution, and windowing explained.
Crate Architecture
How the workspace crates fit together: engine, ops catalogs, state, scripting, connectors, sinks.
Patterns & operators
Event patterns
Closure-based pattern matching with tflo-cep — same six-method API in Rust and TypeScript, side-by-side examples.
Operations & Indicators
Core ops — SMA, EMA, RSI, statistics, derivatives — plus the tflo-fintech TA plugin.
Signals
Cross, glitch, runt, pulse, window detectors — composable primitives for lifecycle events.
Advanced
Checkpointing, async contracts, Deduplicator, Metrics, multi-language scripting, performance.
Demos
Live demonstrations of tflo patterns and operator compositions.
Bindings
Examples
Guides & reference
Quick Start
Get up and running in 5 minutes. Add tflo to your project and run your first streaming computation.
When to use
RxJS / XState / Signals / tflo decision guide.
Contracts
The four pluggable traits: AsyncStateStore, Cursor, ShardRouter, Operator — plus connector auxiliaries.
Deployment shapes
Six concrete shapes — embedded, edge, cluster, WASM, batch / replay — with caveats per shape.
Reference deployment
End-to-end iot-portal example: MQTT → tflo → Kafka → tflo → Influx + Parquet.
Non-goals
What tflo deliberately doesn't do (streaming SQL, JobManager, 2PC exactly-once, cross-shard watermark) and why.
Interop backlog
Designed-but-deferred integrations: tflo-flink, tflo-beam, tflo-kstreams.