GENERATORS

Fake generator: choose a useful test asset

A fake generator might mean sample names, form data, images or documents. FakeKit serves a specific job: fictional data for software testing. Start with the behavior you need to check, then choose an asset that exposes a clear expected result.

Which asset fits your test?

Match the asset to the behavior
TaskUseWhat it cannot establish
Check required fields and datesFictional records with missing-name, expired and date-order casesReal document authenticity or a person's identity
Check a parser or exportMarked JSON or common-field CSVCompatibility with a government schema
Check text-region orderThree original abstract OCR sheets and their manifestRecognition accuracy on passports or IDs
Check weighted arithmeticSingle-field check-digit calculator and test vectorsA full MRZ pass or document validity

Run a bounded test

  1. Write the rule and expected result before creating an input.
  2. Use an ordinary fictional case and a case designed to fail that rule.
  3. Keep the non-valid labels attached when importing or displaying an asset.
  4. Record the observed result and the software version; passing one rule says nothing about untested rules.

The output boundary

Every exported record is marked FICTIONAL TEST DATA — NOT VALID. Test document numbers start with NOT-VALID. The issuer is fictional, and validForIdentification is always false. The record utility exports plain data; the OCR assets are abstract sheets with known text and controlled image conditions. Neither provides government-document artwork, realistic credentials or a general image generator.