test-data-generator

Seeded deterministic fake-data generator with schema-driven generation. Same seed produces identical output every time — critical for reproducible test suites.

Quickstart

import { TestDataGenerator } from 'test-data-generator';

const gen = new TestDataGenerator({
  id: 'sequence()',
  name: 'firstName',
  email: 'email',
  company: 'company',
  salary: { type: 'int', min: 40000, max: 150000 },
}, 42); // seed=42 for reproducibility

const users = gen.generate(5);
// Every run with seed 42 produces identical output

API

Constructor

new TestDataGenerator(schema, seed?)

• schema: Record<string, GeneratorSpec> — field definitions
• seed: number (default: 42) — PRNG seed for reproducibility

Generator Types

Built-in generators (string names or objects with type):

• 'firstName' — Random first name from pool
• 'lastName' — Random last name from pool
• 'email' — Random email address
• 'uuid' — UUID v4-style string
• 'int' / { type: 'int', min?: number, max?: number } — Integer (default 0–100)
• 'float' / { type: 'float', min?: number, max?: number, precision?: number } — Float (default 0–1)
• 'bool' — Random boolean
• 'date' / { type: 'date', daysAgo?: number } — Date within last N days (default 30)
• 'pick' / { type: 'pick', values: any[] } — Random item from array
• 'company' — Random company name
• 'street' — Random street address

Special generators:

• 'sequence()' — Auto-incrementing counter (1, 2, 3, …)
• 'ref(fieldName)' — Reference previous record's field
• Custom function: (state) => any — Full control

Scope & Limits

• Zero runtime dependencies — crypto and perf_hooks not used
• Deterministic only: Same seed, same data. Reseeding restarts the sequence.
• No persistence — generates in-memory; no database or file I/O
• Word pools fixed — first/last names and companies are built-in and small
• No relationships beyond simple ref() — no foreign-key constraints or cascading

Example

const gen = new TestDataGenerator({
  id: 'sequence()',
  firstName: 'firstName',
  lastName: 'lastName',
  email: 'email',
  department: { type: 'pick', values: ['Sales', 'Engineering', 'HR'] },
  startDate: { type: 'date', daysAgo: 365 },
  managerId: (state) => state.random() > 0.5 ? null : Math.floor(state.random() * 10),
}, 123);

const staff = gen.generate(100);
console.log(staff[0]); // Reproducible across runs

License

MIT

Sponsored by Ferrow

Part of the ferrow-toolkit collection · Sponsored by Ferrow