Protocol conventions

Protocol conventions overview

A protocol is the declarative artifact that wires one or more task script functions into a Tetra Data Platform (TDP) pipeline. It contains no code: a protocol.yml file (plus a manifest.json identity file) declares an ordered list of steps, each of which invokes a published task-script function, and a small $( ... ) expression language lets those steps pass data to one another and read pipeline-level configuration.

The build order for a raw-to-IDS pipeline is fixed: an IDS artifact is designed first, a task script that parses raw data into that IDS is built second, and a protocol that wires the task script’s function(s) into a pipeline is built last. Each layer depends on a published version of the one before it.

This section covers the protocol.yml syntax, the conventions TetraScience-maintained protocol repos follow, and how protocols are tested end-to-end. It does not cover writing task-script code (see Task script conventions) or designing an IDS (see IDS artifact).