A generator lazily produces values using yield, keeping only its current state in memory rather than materialising all values at once.

Generators are useful for streams, large files, and data pipelines where building a full list would be expensive or impossible. A generator function suspends at each yield and resumes on the next next() call. Generators implement the iterator protocol automatically — they have both __iter__ and __next__. yield from delegates to a sub-iterator, flattening it and enabling composable generator pipelines without additional nesting.