@graphmd/dataset

The @graphmd/dataset package is the canonical home for GraphMD dataset semantics. It contains the parsing, validation, hashing, and canonicalization logic that defines what a GraphMD dataset means. Other packages (including the legacy packages/core wrapper) should treat this as the authoritative implementation.

Install

npm install @graphmd/dataset

Usage

Validate a snapshot

import { validateDatasetSnapshot } from '@graphmd/dataset';

const snapshot = {
  files: new Map([
    ['types/book.md', new TextEncoder().encode('---\ntypeId: book\nfields: {}\n---\n')],
  ]),
};

const result = validateDatasetSnapshot(snapshot);
if (!result.ok) {
  console.error(result.errors);
}

Compute a dataset hash

import { computeGdHashV1 } from '@graphmd/dataset';

const result = computeGdHashV1(snapshot, 'snapshot');
if (result.ok) {
  console.log(result.cid);
}

Canonicalize a snapshot

import { canonicalizeDatasetSnapshot } from '@graphmd/dataset';

const canonical = canonicalizeDatasetSnapshot(snapshot);

API surface

The public API is defined by the barrel export in src/index.ts. Use that file as the canonical map of exported functions and types.

Development

# Build ESM + CJS outputs
npm --workspace packages/dataset run build

# Run tests
npm --workspace packages/dataset test

# Typecheck
npm --workspace packages/dataset run typecheck