kinship

Deterministic pedigree graph engine for TypeScript.

kinship turns raw pedigree records into a validated, queryable kinship graph for data-driven breeding workflows.

Install

npm install @naviary-sanctuary/kinship

Also available with other package managers:

pnpm add @naviary-sanctuary/kinship
yarn add @naviary-sanctuary/kinship
bun add @naviary-sanctuary/kinship

Core Architecture: The Separated Partner Model

Genealogical DAGs often break when horizontal mating (partner) relationships create cycles. kinship prevents this by maintaining two strict internal topologies:

1. Descent Graph: A pure, strict parent-to-child DAG for lightning-fast biological lineage traversal and exact generation depth tracking.
2. Partner Map: An auxiliary bidirectional map for mating relationships.

These topologies remain mathematically isolated during processing and are only unified into KinshipLink arrays during subgraph extraction (extractFamilyNetwork).

Quick Start

import { build, getAncestors, getDescendants, extractFamilyNetwork } from '@naviary-sanctuary/kinship';

const records = [
  { id: 'A', sireId: 'B', damId: 'C' },
  { id: 'B', sireId: 'D', damId: 'E' },
  { id: 'C' },
  { id: 'D' },
  { id: 'E' },
];

const result = build(records);

if (!result.ok) {
  console.error(result.fatal.code, result.fatal.message);
  process.exit(1);
}

const ancestorsOfA = [...getAncestors(result.graph, 'A')];
const descendantsOfB = [...getDescendants(result.graph, 'B')];
const familyOfA = extractFamilyNetwork(result.graph, 'A');

console.log({
  ancestorsOfA,
  descendantsOfB,
  familyOfA,
  issues: result.issues, // non-fatal diagnostics
});

Data Model

type PedigreeId = string;

interface PedigreeRecordInput {
  id: PedigreeId;
  sireId?: PedigreeId | null;
  damId?: PedigreeId | null;
}

Normalization rules:

• id is trimmed and must be non-empty.
• sireId / damId allow null or undefined for unknown parent.
• Blank parent IDs are invalid.
• Invalid or missing parent links are removed from normalized records and reported as issues.

Diagnostics

The build pipeline returns explainable diagnostics with level, code, message, and optional id/field.

Code	Level	Meaning
EMPTY_ID	error	Individual ID is empty
EMPTY_PARENT_ID	error	Parent ID is empty
SELF_PARENT	error	Parent ID equals child ID
SAME_PARENT	error	sireId and damId are identical
MISSING_PARENT	error	Parent reference does not exist in input set
DUPLICATE_ID	error	Duplicate ID exists (first record is kept)
CYCLE_DETECTED	fatal	Cycle detected in child-to-parent graph

If a fatal issue exists, build returns { ok: false, fatal, issues }.

Development

bun install
bun run typecheck
bun run lint
bun run fmt
bun test

Community

• Contributing
• Code of Conduct
• AI Policy