Elixir Streams

The post explains that Enum performs eager collection operations, while Stream describes lazy computations that are only realized when an Enum function forces them. It uses ranges, file and stdin streams, and stream constructors such as repeatedly, iterate, unfold, and resource to show how lazy composition works on large or infinite data.

Reading notes

#

• Enum transforms collections immediately, while Stream stores future computations.
• Streams still work with Enum because they implement Enumerable.
• Enum.to_list and Enum.take force a stream to run.
• A large range can be mapped lazily, and only the requested items are computed.
• Streams can be composed with filter and map without realizing the whole collection.
• File.stream! returns file lines or bytes as a stream.
• IO.stream gives the same kind of interface for standard input.
• GenEvent.stream creates a stream of events from a GenEvent manager.
• Stream.repeatedly builds an infinite stream by calling a function each time an item is requested.
• Stream.iterate starts from an initial value and generates each next value from the previous one.
• Stream.unfold uses an accumulator and a generator function that returns {next_element, new_accumulator}.
• Stream.resource is suited to streams that wrap external resources.
• The GitHub example uses Stream.resource to fetch paginated API results lazily.
• process_page handles the end of pagination, the need to fetch another page, and the normal case that yields items.
• The resulting API lets users query repositories with a stream-like interface and combine organizations with Stream.flat_map.