5e-database

JSON game data for the D&D 5th Edition SRD, used by the Draconian Dungeons game. This is a fork of 5e-bits/5e-database. The data is edited as plain JSON files and checked by Zod schemas and Vitest tests; there is no database or build step.

src/
  2014/
    en/                              English source data, one 5e-SRD-{Collection}.json per collection
    {locale}/                        Translation overlays (fr-FR, pt-BR, ru); not published
    schemas/                         One Zod schema per collection
    tests/                           Schema, duplicate-index, table and translation tests
    5e-SRD-Monsters-Normalized.json  Generated from en/5e-SRD-Monsters.json
    normalize-monsters.ts            Generator for the file above (`pnpm normalize:monsters`)
    monster-engine-annotations.json  Monster annotations merged in by the generator
  2024/                              Same layout for the 2024 SRD (no normalized monsters)
  i18n/                              localize(), overlay validator and coverage report (see CONTRIBUTING.md)
  provenance/                        SRD baseline and source markers for data not taken from the SRD (see NOTICE.md)
  schemas/common.ts                  Schema pieces shared by both years
  tests/translationTestUtils.ts      Shared translation test suite
docs/
  game-engine-changes.md             What this fork changes relative to upstream, and why
  monster-environments.md            The environment_tags list of the monsters
  icon-keys.md                       The icon keys of the spells, class features and classes
  rule-reference.md                  The markdown of the rules text, and the order, summary, keywords and related links
  i18n-overlay-migration.md          What the move to text-only translation overlays (#31) changed
  upstream-changelog.md              Upstream's release history, kept for reference and no longer generated
schema/                              Generated JSON Schema, one per collection (see CONTRIBUTING.md)
ids/                                 Generated id lists, published ids and removed-id tombstones (see CONTRIBUTING.md)
types/                               Generated .d.ts types, one per collection (see CONTRIBUTING.md)

Every collection is kept. Record counts come from the data files. Status means:

• game data: core data for the game.
• extensible: reference content that is kept and open to more entries and fields.
• reference: kept as SRD reference data.
• dormant: the 2024 data. The 2024 (SRD 5.2) set is frozen: it is kept as the in-repo source of more backgrounds and feats (three backgrounds and sixteen feats of it are now in the 2014 collections), is not reshaped, and is excluded from the published package. Its schema tests stay.

The npm package is named @markpol/dd-srd; the repository keeps the name 5e-database. Every 2014 collection has a typed entry:

import monsters from '@markpol/dd-srd/monsters'; // Monster[], types from types/5e-SRD-Monsters.d.ts
import spells from '@markpol/dd-srd/spells';

The older deep paths (@markpol/dd-srd/src/2014/en/5e-SRD-Spells.json, .../schema/..., .../ids/...) keep resolving for one major version. The package contains the English data only; the translation overlays are in the repository but not published, pending a provenance check. See CONTRIBUTING.md.

You need Node 22 or 24 (see .nvmrc) and pnpm.

pnpm install
pnpm test       # schema, uniqueness, table and translation checks
pnpm lint       # ESLint, including JSON and Prettier rules

CI (.github/workflows/ci.yml) runs pnpm lint, pnpm test, tsc --noEmit, a check that schema/, types/ and ids/ are up to date with the Zod schemas and data, and a smoke test that packs the package and imports every entry from the tarball, on Node 22 and 24 on every pull request and on pushes to main. It also prints a per-locale translation coverage report (pnpm run i18n:coverage), which never fails the build.

A published id is never renamed or removed by default, because stored data may refer to ids directly. When a removal is unavoidable it gets a tombstone in ids/removed.json (and an alias where one fits), and a test fails on any published id that is gone without one. Published ids are tracked from upstream release v5.12.0 in ids/published.json; see CONTRIBUTING.md.

See CONTRIBUTING.md for how to change data, add a field, add a translation and cut a release. Game data problems can be reported with the "Game data issue" issue template.

This project is licensed under the terms of the MIT licence. See LICENSE.md. The game content comes from the SRD 5.1 (2014 data) and SRD 5.2 (2024 data) under CC-BY-4.0; NOTICE.md has the attribution and lists the data written for the game rather than taken from the SRD.