Give a concrete example: "My PagedResults class implements __len__ for total count, __iter__ to yield items across pages, and __getitem__ for index access. Dataclasses handle __init__, __repr__, and __eq__ automatically."
Python's data model lets custom classes integrate with built-in operations by implementing special methods: __len__ for len(), __iter__/__next__ for iteration, __eq__/__lt__/__hash__ for comparison and hashing, __repr__/__str__ for string representation, __getitem__ for indexing, and __enter__/__exit__ for context managers. Strong candidates discuss the total_ordering decorator, the difference between __repr__ and __str__, and dataclasses as a shortcut for common dunder methods.
Tests Python depth beyond surface usage. Candidates who cannot implement basic dunder methods will struggle with Pythonic design. Ask them to sketch a class that works as a context manager — this tests __enter__/__exit__ understanding.