A pure, composable Haskell faker library for generating fake data in tests. Built with the State monad for deterministic, reproducible test data generation.

Features

• Pure Functions: No global state or I/O effects in generators
• Deterministic: Seeded generation for reproducible tests
• Composable: Monadic combinators for building complex generators
• Type Safe: Concrete return types, no stringly-typed APIs
• Minimal Dependencies: Only standard Haskell libraries
• Easy to Use: Simple, intuitive API

Installation

cabal build
cabal test

Usage

Basic Generation

import Fake
import qualified Fake.Person as Person
import qualified Fake.Internet as Internet
-- Deterministic generation
let name = runFakerSeed 42 Person.fullName
let email = runFakerSeed 42 Internet.email
-- Non-deterministic generation
randomName <- runFaker Person.fullName

Complex Objects

data User = User
  { name :: String
  , email :: String
  , age :: Int
  } deriving (Show)
generateUser :: Fake User
generateUser = do
  n <- Person.fullName
  e <- Internet.email
  a <- integerRange 18 80
  return $ User n e a
user <- runFaker generateUser

Using Combinators

-- Generate 5-10 emails
emails <- runFaker $ vectorOf (5, 10) Internet.email
-- Optional value (50% chance)
username <- runFaker $ maybeGen Internet.username
-- Pick one from list
email <- runFaker $ oneof [Internet.safeEmail, Internet.freeEmail]

Architecture

The library uses a State monad wrapping StdGen for pure, composable random generation:

newtype Fake a = Fake (State StdGen a)
  deriving (Functor, Applicative, Monad)

Three execution modes:

• runFaker :: Fake a -> IO a - System random
• runFakerSeed :: Int -> Fake a -> a - Deterministic
• runFakerWith :: StdGen -> Fake a -> a - Custom RNG

Available Generators

Primitives

• Numbers: integer, integerRange, natural, float, double, bool
• Characters: char, alpha, numeric, alphanumeric
• Strings: string, stringBounded
• Selection: elements, shuffle

Combinators

• Repetition: vector, vectorOf, sequence
• Optionality: maybeGen, oneof
• Functional: filterGen

Person

• Names: firstName, lastName, fullName, name
• Titles: title, jobTitle, prefix, suffix
• Bio: bio

Internet

• Email: email, safeEmail, freeEmail
• Usernames: username, domain, domainSuffix
• URLs: url, slug

Address

• Geography: city, country, countryCode
• Postal: postalCode, zipCode
• Coordinates: latitude, longitude

Testing

import Test.Hspec
import Fake
import qualified Fake.Person as Person
spec :: Spec
spec = describe "User generation" $ do
  it "generates consistent names with same seed" $ do
    let name1 = runFakerSeed 123 Person.fullName
    let name2 = runFakerSeed 123 Person.fullName
    name1 `shouldBe` name2

Run tests:

cabal test