@silyze/async-pipe

Async Pipe

An asynchronous, single-producer single-consumer channel implementation built on top of @mojsoski/async-stream.

AsyncPipe<T> allows you to write data asynchronously into the pipe and asynchronously read it out in order. Internally, it manages flow control using promises and resolver locks, providing a backpressure-aware, event-driven mechanism to stream data between asynchronous producers and consumers.

This makes it ideal for scenarios where you want to:

• Connect asynchronous data producers and consumers with minimal overhead
• Stream data in an ordered, sequential fashion
• Use async iteration (for await) to process streamed items
• Integrate smoothly with @mojsoski/async-stream transforms and utilities

---

Install

npm install @silyze/async-pipe

Usage

import { AsyncPipe } from "@silyze/async-pipe";

const pipe = new AsyncPipe<string>();

pipe.transform().forEach((item) => console.log(item));

pipe.write("Hello, World!");

Notes on Readers

AsyncPipe<T> implements the interface AsyncStream<T>, combining:

• write(input: T): Promise<void> to push data into the pipe
• read(signal?: AbortSignal): AsyncIterable<T> to consume data asynchronously
• transform(): AsyncTransform<T> to get a transform stream wrapper over the pipe

Single Reader Limitation

AsyncPipe is designed for a single active reader. Only one consumer should call .read() or consume .transform() at a time.

Multiple simultaneous readers will cause internal resolver conflicts and potentially lost or duplicated data.

Supporting Multiple Readers

To safely share the data stream with multiple readers, use .buffer() on the transform stream:

const buffered = pipe.transform().buffer({
  clear: "all-read", // clears buffer only after all readers have consumed
});

buffered.forEach((item) => {
  console.log("Reader 1:", item);
});

buffered.forEach((item) => {
  console.log("Reader 2:", item);
});

buffered.forEach((item) => {
  console.log("Reader 3:", item);
});

• .buffer() creates a buffered stream that caches data internally for multiple consumers.
• You can configure buffering and clearing policies with AsyncBufferConfig.

Using .buffer() enables multiple consumers to independently and concurrently read from the pipe without conflicts, at the cost of extra memory to hold buffered data.