Add chunk_size to collection iterate() methods #1149

JS client parity: DatasetClient.iterate_items(), KeyValueStoreClient.iterate_keys() and RequestQueueClient.iterate_requests() take chunk_size, but the iterate() methods of the collection clients (Actors, builds, runs, schedules, tasks, webhooks, webhook dispatches, store, datasets, key-value stores, request queues) don't, for example build_collection.py#L93 and dataset_collection.py#L78-L113. The shared pagination helper already supports a page size (_pagination.py#L32-L56), so callers just can't reach it.

The JS client accepts chunkSize on every paginated list (utils.ts#L368-L372, used in api_client.ts#L102-L139).

Add chunk_size to those iterate() methods. It caps the limit of each page request; without it the API keeps choosing the page size.

✍️ Drafted by Claude Code