python

wh40kdc

The 40kdc Warhammer 40K dataset behind a linked, typed API — find units, follow them to their weapons, abilities, phases, and factions. The Python counterpart to @alpaca-software/40kdc-data (npm) and wh40kdc (crates.io), held in behavioral lockstep with both by the shared conformance corpus in the 40kdc-data repository.

pip install wh40kdc

from wh40kdc import Dataset

ds = Dataset.embedded()
unit = ds.units.find("Khârn the Betrayer")
for weapon in unit.weapons:
    print(weapon.name)

What ships:

• Linked dataset API — Dataset with id/name lookups (diacritic- and punctuation-insensitive), reverse indexes, and join queries across units, weapons, abilities, phases, and factions.
• Importers — ListForge, NewRecruit (JSON / WTC-compact / WTC-full / simple), GW app export, and Rosterizer, all resolving to one Roster shape.
• Exporters — the same roster formats back out, byte-identical to the TypeScript and Rust implementations.
• Cruncher — the damage-projection engine (attacks → hits → wounds → unsaved → damage → after-FNP → models-killed) with buff resolution and per-buff attribution.
• Scoring, terrain, translation — secondary-mission scoring, terrain layout resolution and keystone measurements, and the ability-DSL / scoring-card describers.
• Validator — validate(target, value) against the canonical JSON Schemas, emitting the closed error-code enum shared by all implementations.

The conformance runner (used by the cross-implementation parity differ) is invoked as python -m wh40kdc.runner — see conformance/RUNNER_PROTOCOL.md in the repository.

Development

cd python
uv venv && uv pip install -e ".[dev]"
pytest          # includes the conformance suite when run inside the repo
ruff check .
mypy src

Generated artifacts (_bundle.json, _spec.py, _types.py, src/wh40kdc/schemas/) are produced by the scripts in codegen/ and checked for drift in CI — regenerate with:

python3 codegen/sync_bundle.py && python3 codegen/sync_spec.py && python3 codegen/gen_typeddicts.py

License

MIT. The schema content the dataset describes is CC0; enrichment data is CC BY 4.0 — see the repository's licensing table.