Add __repr__ to Invalid so it surfaces the path and error type (#358) - #546

Fixes #358.

Problem

Invalid.__str__ produces an informative message including the path and error type:

expected int for dictionary value @ data['k']

But Invalid has no __repr__, so repr() falls back to the default exception repr and drops that context:

>>> repr(err)
TypeInvalid('expected int')     # path and error_type lost

This matters in practice: MultipleInvalid.__repr__ renders its .errors list, so inspecting or logging a MultipleInvalid shows a list of uninformative sub-error reprs.

Fix

Give Invalid a __repr__ that surfaces the same context __str__ already does — matching the __repr__ that MultipleInvalid already defines:

def __repr__(self) -> str:
    return '%s(%r)' % (self.__class__.__name__, str(self))

Result:

>>> repr(err)
TypeInvalid("expected int for dictionary value @ data['k']")
>>> repr(multiple)
MultipleInvalid([TypeInvalid("expected int for dictionary value @ data['k']")])

No validation behavior changes — only the repr of the error objects.

Testing

• Added test_invalid_repr_includes_path_and_type, asserting the repr includes the path/type and that it surfaces inside the MultipleInvalid repr. It fails on master (TypeInvalid('expected int')) and passes with this change.
• Full suite passes (170 passed). black and flake8 are clean on the changed files.
• One unrelated doctest (README.md:93) fails on this machine under Python 3.14; it fails identically without this change (the tox matrix is 3.9–3.12) and is untouched here.

Disclosure: this change was prepared with the assistance of an AI tool (Claude Code). I verified the behavior, added and ran the test, ran the suite and linters, and take responsibility for the contribution and will respond to review feedback personally.