Repository layout

The repository holds reusable software and the public documentation. It does not hold datasets, calculation output, or the tooling that drives a particular run.

goldilocks-data/
├── src/
│   └── goldilocks_data/  reusable submission and analysis code
├── tests/                scientific and workflow regression tests
└── docs/                 this documentation site

Ownership

Location	Responsibility
src/	Mechanics shared across every run: submission, de-duplication, cleanup, k-mesh schedules, convergence labelling, publishing
tests/	Regression tests that depend on no private data
docs/	User-facing documentation built by MkDocs

What stays outside

A concrete run — its AiiDA group and profile, its controller scripts, its analysis notebook, the dated tables it exports — is operational, not reusable, and lives outside this repository along with the private CSV and CIF paths it reads. What is published from it goes to PSDI as a citable record instead.

A private path must never reach the package API.

The package directory src/goldilocks_data/codes/ contains Python types for identifying DFT codes; it is unrelated to the DFT codes themselves.