Introduction

Audience

This guide assumes a competent Python developer who is new to this codebase. It favors architecture and contracts over restating what the code already says. If you only want to install the package and run it against an existing collection, read the User Guide instead; this guide is about working on the code.

What the package does

rms-metadata-tools (importable as metadata_tools) generates PDS3 index, geometry, and cumulative metadata tables, and their PDS3 labels, for planetary science data collections at the PDS Ring-Moon Systems Node. Each row of a table holds the metadata for one data product. The Overview describes the three table kinds and the pipeline that produces them.

Runtime and key dependencies

The package targets Python 3.11+. Its principal dependencies are:

• rms-oops — the geometry/SPICE engine that computes backplanes; the geometry stage is built on it.

• rms-filecache — the FCPath path abstraction used for every file access, so local and remote (gs://, s3://) storage are interchangeable.

• rms-pdstemplate — renders PDS3 labels from templates.

• rms-pdsparser and rms-pdstable — read PDS3 labels and tables.

• rms-pdslogger — structured logging through the package’s global logger.

• rms-julian, rms-vicar, cspyce, and fortranformat — time conversion, VICAR labels, SPICE clock conversion, and Fortran-style numeric formatting.

• rms-cloud-tasks (optional) — distributes per-volume work across GCP.

Design in one sentence

A host-agnostic engine under src/metadata_tools/ does the work; each supported collection contributes a small host configuration package under src/metadata_tools/hosts/<HOST>/ that plugs into the engine. The split is the central idea of the codebase, and is the subject of Architecture and Extending the system.