Functions

Functions let you move custom generation logic into Python when YAML becomes too limited.

Where they live

Put modules under functions/ in your faux-db project.

functions/
├── __init__.py
└── text.py

How schemas reference functions

Use a func column with an import path:

slug:
  type: func
  func: functions.text.make_slug
  params:
    source: '{{ col("title") }}'

Callable behavior

faux-db imports the callable once when planning generation. At runtime it passes:

• declared params
• context if the function signature accepts context
• row if the function signature accepts row

Example

def make_slug(source: str) -> str:
    return source.lower().replace(" ", "-")

When to use functions instead of providers

Use functions when:

• logic is too specific for a reusable provider
• you need branching or more complex transformations
• you need direct access to the current row or generation context

Use providers when:

• the logic should be named and reused across many schemas
• the config should stay fully declarative

Best practices

• Keep functions small and deterministic where possible.
• Prefer explicit parameter names over reading unrelated global state.
• Use functions to complement configuration, not replace it entirely.