ConcurrentStream

The ConcurrentStream aims to combine the functionality of execution in parallel and AsyncSequence.

let stream = (1...100).stream.map(heavyWork)

while let next = try await stream.next() {
    ...
}

The heavyWorks are executed in parallel, while the completed ones are reported in order.

Getting Started

ConcurrentStream uses Swift Package Manager as its build tool. If you want to import in your own project, it's as simple as adding a dependencies clause to your Package.swift:

dependencies: [
    .package(url: "https://github.com/Vaida12345/ConcurrentStream.git", branch: "main")
]

and then adding the appropriate module to your target dependencies.

Using Xcode Package support

You can add this framework as a dependency to your Xcode project by clicking File -> Swift Packages -> Add Package Dependency. The package is located at:

https://github.com/Vaida12345/ConcurrentStream

Documentation

Full documentation available as DocC (View raw ConcurrentStream documentation).

Implementation Details

A stream, not a sequence

As the name suggests, the package provides a stream, not a sequence. Which means that you cannot iterate using for-loop. However, you could

use stream.next()
use stream.sequence to convert it into a sequence. The seqeunce method would wait for all elements to present before returning.
use stream.async to convert it into an async sequence. This method returns immediately, and new elements can be obtained using for await.

Important: A stream is fragile, elements are discarded during traversal. Hence do never reuse a stream. (Similar to TaskGroup)

Eazy conversion

This framework also offers eazy ways to convert between stream and sequence. For example,

for i in try await (1...10).stream.sequence {
    print(i)
}

Would provide exactly the same as

for i in (1...10) {
    print(i)
}

With arguably similar performance.

To read more about choice and implementation details, see Principle