examples.md

Iterator — where it lives outside this repo

Cited, real implementations to study (or point an agent at) when designing or reviewing traversal code.

Python standard library

• Generators and itertools — the pattern as language feature plus a toolbox of composable iterators (count, islice, chain, tee). docs.python.org/3/library/itertools.html
• os.walk / pathlib.Path.iterdir — lazy filesystem traversal: a directory tree of any size, constant memory. docs.python.org/3/library/os.html#os.walk
• csv.reader — file rows as an iterator; the file object underneath is itself an iterator of lines. docs.python.org/3/library/csv.html

Major ecosystems

• Django QuerySet — lazily evaluated; .iterator() streams rows over a server-side cursor instead of caching the whole result: the page-hiding move at ORM scale. docs.djangoproject.com/en/stable/ref/models/querysets/#iterator
• boto3 paginators — AWS list APIs return truncated pages; a paginator wraps the continuation-token dance into one iterable, exactly this unit's iterate_pages shape. boto3.amazonaws.com/v1/documentation/api/latest/guide/paginators.html (unverified)

What to notice across all of them

Every one hides a cursor protocol (continuation tokens, DB cursors, file offsets) behind the one protocol Python already speaks. And every one documents its laziness as a feature with consequences — Django warns that .iterator() skips caching; file iterators exhaust. When reviewing, ask: does the signature promise Iterator, and does anything downstream silently materialize it?