fabrikate 0.2.1

Generate contextually coherent mock data — people, companies, and transactions that actually make sense together.

fabrikate 🏭

Contextually coherent mock data generation for Python.

Unlike random fake data generators, fabrikate ensures everything makes sense together. A person generated in Tokyo gets a Japanese name, address, phone format, and bank. A company's employees, products, and transactions are internally consistent.

Installation

pip install fabrikate

Quick Start

from fabrikate import World

# Create a world set in Japan
world = World(locale="ja_JP", seed=42)

# Generate a company with employees
company = world.company(industry="tech")
employees = company.hire(10)

# Generate products for that company
products = world.products(5, company=company)

# Generate transactions
transactions = world.transactions(100, receiver=company)

# Export everything as JSON
print(world.to_json())

Why fabrikate?

Faker generates random data — a person might have an American name, a Japanese phone number, and a German address. fabrikate keeps everything coherent within a cultural and business context.

Feature	Faker	fabrikate
Random names	✓	✓
Locale-aware names	✓	✓
Coherent address + phone + bank	✗	✓
Company → employee inheritance	✗	✓
Industry-appropriate products	✗	✓
Linked transactions	✗	✓
Full dataset export	✗	✓
Seed reproducibility	✓	✓

Supported Locales

• en_US — United States (English)
• ja_JP — Japan (Japanese)
• de_DE — Germany (German)
• pt_BR — Brazil (Portuguese)

More coming soon. PRs welcome!

Core Concepts

World

A World is your mock data universe. Everything generated within a world shares context and can reference each other.

world = World(locale="de_DE", seed=123, industry="finance")

Context

The Context controls what locale, industry, and time period data is generated for. You usually don't create these directly — World handles it.

# Child contexts inherit but can override
child_ctx = world.ctx.child(locale="en_US")

Generators

Each entity type has a generator that respects the context:

person = world.person()              # Coherent person
company = world.company()            # Coherent company
product = world.product(company=co)  # Product tied to company
tx = world.transaction(sender=p)     # Transaction from person

Company → Employee Inheritance

When you hire employees through a company, they inherit its context:

company = world.company(industry="tech")  # German tech company
team = company.hire(5)                     # 5 German employees

for person in team:
    print(person.full_name)    # German names
    print(person.phone)        # +49 format
    print(person.bank)         # German bank
    print(person.company_id)   # Links to company

Full Dataset Export

Dump everything to JSON for test fixtures:

data = world.dataset()
# {
#   "metadata": {"locale": "de_DE", "seed": 123, ...},
#   "people": [...],
#   "companies": [...],
#   "products": [...],
#   "transactions": [...]
# }

API Reference

World(locale, seed, year, industry)

• locale (str): Locale code, default "en_US"
• seed (int|None): Random seed for reproducibility
• year (int): Reference year, default 2025
• industry (str|None): Default industry context

world.person(**overrides) → Person

world.people(count, **overrides) → list[Person]

world.company(**overrides) → Company

world.companies(count, **overrides) → list[Company]

world.product(company?, **overrides) → Product

world.products(count, company?, **overrides) → list[Product]

world.transaction(sender?, receiver?, product?, **overrides) → Transaction

world.transactions(count, sender?, receiver?, **overrides) → list[Transaction]

world.dataset() → dict

world.to_json(indent=2) → str

License

MIT