Skip to content

Tutorial 1: Interactive safety

You are collecting arcade sensor readings that must stay valid integers while you experiment.

Omnipy models continuously validate data and type-mimic the wrapped Python type.

A Pydantic model is a use-once-then-discard validator; an Omnipy model is a continuous static-type repairman in a dynamic environment.

Setup

>>> import omnipy as om
>>> om.runtime.config.data.model.interactive = True

Parse and inspect

>>> readings = om.Model[list[int]]((120, '135', 142.0))
>>> readings

╭──────────────────╮
│ Model[list[int]] │
│                  │
│ [120, 135, 142]  │
╰──────────────────╯

Error + rollback demonstration

>>> before_error = readings.copy()
>>> try:
...     readings.append('bonus-ticket')
... except Exception as err:
...     print(type(err).__name__)
>>> after_error = readings
>>> print(before_error == after_error)
>>> readings
ValidationError
True

╭──────────────────╮
│ Model[list[int]] │
│                  │
│ [120, 135, 142]  │
╰──────────────────╯

Interactive mode framing

Choose based on trade-offs, not environment:

  • runtime.config.data.model.interactive = True: snapshot + rollback robustness.
  • runtime.config.data.model.interactive = False: lower memory overhead and raw write speed, but no rollback after validation errors.
>>> om.runtime.config.data.model.interactive = False
>>> fast_mode = om.Model[list[int]]([10, 20, 30])
>>> try:
...     fast_mode.append('VIP')
... except Exception as err:
...     print(type(err).__name__)
>>> fast_mode
ValidationError

With interactive=False, Omnipy still raises a validation exception immediately, but the invalid value remains in the model state ([10, 20, 30, 'VIP']).

╭─────────────────────╮
│  Model[list[int]]   │
│                     │
│ [10, 20, 30, 'VIP'] │
╰─────────────────────╯

What you learned

  • Continuous validation + type mimicking keep model edits safe.
  • Rollback behavior is controlled via runtime.config.data.model.interactive.

Common pitfalls

  • Assuming non-interactive mode is "strictly safer" because it still raises errors. It raises, but it does not roll back invalid mutations.

Next steps