Processing pipeline#

pycsamt.pipeline chains the individual processing tools into one reproducible, auditable workflow — from raw Sites to publication-ready EDIs, QC plots, and a run report — driven by short step codes, presets, or a config file.

These examples are information-first: the point is not the pictures but the workflow itself — what steps exist, how a run is assembled, what it produces, and how to reproduce it. Each script prints the meaningful output (step catalogues, per-step timings, run summaries, the reproducible config) and ends with one small summary figure.

Step catalogue — the 47 processing steps, by category, and how to inspect one;
Build and run — assemble a pipeline of steps, run it on a real survey line, and read the PipelineResult and its output package;
Presets — ready-made recipes (basic_qc, full_processing …) you can run in one line;
Config and reproducibility — serialise a pipeline to YAML and rebuild it exactly with Pipeline.from_yaml();
Stratagem presets — instrument-specific end-to-end workflows.