# DATA LICENSE

[English](DATA_LICENSE) | [中文](DATA_LICENSE.zh.md)

The *code* of prokname is distributed under the MIT License (see `LICENSE`).

The *data assets* shipped in `src/prokname/data/` follow the licenses of their
sources and are distributed under the corresponding terms:

## `genus_gender.json`, `corpus` LPSN-derived parts

Derived from LPSN (List of Prokaryotic names with Standing in Nomenclature,
https://lpsn.dsmz.de/). LPSN data (including downloads and API data) is
distributed under **CC BY-SA 4.0**. Derivative works must therefore:

1. Cite the LPSN reference designated on the LPSN site and state the access
   date of the data.
2. Link back to the LPSN source pages when redistributing electronically.
3. Inherit the CC BY-SA 4.0 license (ShareAlike).

Human-curated entries added by the prokname authors that are not derived from
LPSN text (plain facts: genus name -> grammatical gender) are released under
CC0 1.0; the file as a whole is distributed under CC BY-SA 4.0 to comply with
the ShareAlike obligation of the LPSN-derived portion.

## `rules.json`, `gender_endings.json`, `person_genitive.json`, `stems.json`

Rules distilled from the International Code of Nomenclature of Prokaryotes
(ICNP) orthography appendix and general Latin grammar. Nomenclatural codes and
rules as such are not subject to copyright; the JSON expression here is
released under **CC0 1.0** (dedicated to the public domain by the authors),
except where a cell is explicitly marked as transcribed from a copyrighted
source, in which case the source's terms apply.

## Status notice

`person_genitive.json` models the latinisation *declension paradigms* behind
person-derived genitive epithets (ICNP Rule 20 orthography), not a
person-gender × stem-letter grid. Only the paradigms anchored to published LPSN
epithets and the attested surnames built on them are flagged `verified: true`;
the `-iae`, `-is` and `-ae`-by-default paradigms ship as `verified: false`
proposals pending sign-off against the ICNP orthography appendix, and the
orthographic variant tolerance list is likewise an M0 deliverable.

Commercial users should additionally note potential CBD benefit-sharing
obligations (Cali Fund) flagged by DSMZ for LPSN-derived digital sequence
information use.

## SeqCode Registry data — resolved 2026-09-25

The upstream terms were read, and they are quoted here rather than paraphrased.
From the footer of https://registry.seqco.de/page/api, retrieved 2026-09-25:

> © 2022-2026 The SeqCode Initiative
> All information contributed to the SeqCode Registry is released under the
> terms of the Creative Commons Attribution (CC BY) 4.0 license

So the "SeqCode data: CC-BY 4.0" attribution that `storage/store.py`,
`tests/fixtures/manifest.json` and `corpus_seed.json` emit is **correct**, and
it is now backed by the source instead of carried over from an earlier draft.

What that means concretely for this repository:

* `src/prokname/data/seqcode_registered.json` — a dated snapshot of the names
  the Registry reports as validly published under SeqCode, built by
  `scripts/build_seqcode_snapshot.py` — is redistributed under CC BY 4.0. The
  obligation is attribution: keep the `_meta.licence` / `_meta.attribution`
  fields, and cite the SeqCode Initiative plus the per-name
  `https://seqco.de/i:<id>` URIs when republishing.
* CC BY 4.0 carries **no ShareAlike**, unlike the CC BY-SA 4.0 that governs the
  LPSN-derived assets above. Where a single file mixes both kinds of content,
  the stricter (BY-SA) obligation governs the whole file — see item 2 below.
* This addresses copyright and licence. It does **not** settle whether the
  European sui generis database right is engaged by extracting the full name
  list; the Registry publishes no statement on that. The snapshot's own
  `_meta` records the retrieval date and source endpoint so the extent of the
  extraction is at least transparent. If a release artefact redistributes the
  snapshot separately rather than shipping it inside the package, ask the
  SeqCode Initiative to confirm.

**Still unresolved, recorded rather than assumed away:**

1. **What CC BY-SA 4.0 ShareAlike implies for our own derived assets.**
   ShareAlike travels with the *derivative work*, not with the fact that a
   figure was re-typed by hand. Concretely: an asset that reproduces or is
   based upon LPSN text (e.g. `genus_gender.json`, whose gender values are
   transcribed from LPSN) must be re-published under CC BY-SA 4.0 — which is
   why that file is distributed under BY-SA even though its CC0-worthy cells
   are author-curated. A rule table whose *wording and selection* is
   independent of LPSN text (nomenclatural rules are not themselves subject to
   copyright) can stay CC0 1.0 — but the moment such a file's contents are
   copied from, or made by bulk reference to, an LPSN-derived source, the
   BY-SA obligation attaches to the whole file and the file-level licence must
   change with it. Mixing both kinds of content in one asset therefore forces
   the stricter licence, and the per-asset derivations below have not been
   audited cell by cell.
2. **Consequence for re-publishing derived rule assets.** Before any asset is
   mirrored into a public dataset, a release archive or any redistributed
   supplement, its derivation must be recorded (source, access date, which cells
   are author-curated). That audit has not been performed; it is an M0 item.
3. **Two shipped assets are not enumerated by the sections above.**
   `type_genus_stems.json` (type-genus declension stems plus a list of attested
   suprageneric names) and `lpsn_status.json` (a map of LPSN's
   `lpsn_taxonomic_status` labels) are both keyed on vocabulary published by
   LPSN, so the CC BY-SA 4.0 section is plausibly the governing one — but that
   has not been audited entry by entry, and neither file carries a licence
   field. Until the audit is done, treat both as governed by the LPSN terms and
   do not re-license them as CC0.

Resolving items 1–3 (and the database-right question raised above) is a
prerequisite for any public redistribution of these assets, including a data
availability statement.
