@effectionx/stream-helpers

中文文档
0.8.3发布于 2026/06/17MITTypeScriptES Module构建来源已验证近 30 天下载 4.8万

Type-safe stream operators like filter, map, reduce, and forEach

原始使用文档
文档目录18 节
  1. Included Helpers
  2. Filter
  3. Map
  4. Reduce
  5. Batch
  6. Valve
  7. Drain
  8. First
  9. Last
  10. Take
  11. TakeWhile
  12. TakeUntil
  13. ForEach
  14. Subject
  15. Passthrough Tracker
  16. Composing stream helpers
  17. Testing Streams
  18. Faucet

Stream Helpers

A collection of type-safe stream helpers built on top of Effection for efficient and controlled stream processing.

Included Helpers

Filter

The filter helper narrows a stream by only passing through items that match a predicate.

import { filter } from "@effectionx/stream-helpers";
import { each } from "effection";

// Example: Synchronous filtering
function* syncExample(source: Stream<number, unknown>) {	

  const gt5 = filter<number>(function* (x) { return x > 5 });

  for (const value of yield* each(gt5(stream))) {
    console.log(value); // Only values > 5
    yield* each.next();
  }
};

// Example: Asynchronous filtering
function* asyncExample(source: Stream<number, unknown>) {

  const evensOf = filter<number>(function* (x) {
    yield* sleep(100); // Simulate async operation
    return x % 2 === 0; // Keep only even numbers
  });

  for (const value of yield* each(evensOf(stream))) {
    console.log(value); // Only even numbers
    yield* each.next();
  }
});

Map

The map helper transforms each item in a stream using a provided function. This is useful for data transformation operations where you need to process each item individually.

import { map } from "@effectionx/stream-helpers";
import { each } from "effection";

function* example(stream: Stream<number, unknown>) {
  const double = map<number>(function* (x) {
    return x * 2;
  });

  for (const value of yield* each(double(stream))) {
    console.log(value); // Each value is doubled
    yield* each.next();
  }
}

Reduce

The reduce helper transforms each item in a stream by applying it to an accumulated value.

import { reduce } from "@effectionx/stream-helpers";
import { each } from "effection";

function* example(stream: Stream<number, unknown>) {
  let sum = reduce(function* (total: number, current: number) {
    return total + current;
  }, 0);

  for (let value of yield* each(sum(streamOf([1,2,3]))) {
    console.log(value) // logs 1 -> 3 -> 6
    yield* each.next();
  }
}

Batch

The batch helper is useful when you want to convert individual items passing through the stream into arrays of items. The batches can be created either by specifying a maximum time or a maximum size. If both are specified, the batch will be created when either condition is met.

import { batch } from "@effectionx/stream-helpers";
import { each } from "effection";

// Example: Batch by size
function* exampleBySize(stream: Stream<number, unknown>) {
  const byThree = batch({ maxSize: 3});

  for (const items of yield* each(byThree(stream))) {
    console.log(batch); // [1, 2, 3], [4, 5, 6], ...
    yield* each.next();
  }
};

// Example: Batch by time
function* exampleByTime(stream: Stream<number, unknown>) {
  const stream = batch({ maxTime: 1000 })(sourceStream);

  for (const batch of yield* each(stream)) {
    console.log(batch); // Items received within 1 second
    yield* each.next();
  }
});

// Example: Combined batching
function* exampleCombined(stream: Stream<number, unknown>) {

  const batched = batch({
    maxSize: 5,
    maxTime: 1000,
  });

  for (const batch of yield* each(batched(stream))) {
    console.log(batch); // Up to 5 items within 1 second
    yield* each.next();
  }
});

Valve

Allows to apply backpressure to the source stream to prevent overwhelming the downstream consumer. This is useful with any stream that generates items faster than the consumer can consume them. It was originally designed for use with Kafka where the producer can cause the service to run out of memory when the producer produces many faster than the consumer to process the messages. It can be used as a buffer for any infinite stream.

import { valve } from "@effectionx/stream-helpers";
import { each } from "effection";

function* example() {
  const regulated = valve({
    // buffer size threshold when close operation will invoked
    closeAt: 1000,
    *close() {
      // pause the source stream
    },

    // buffer size threshold when open operation will be invoked
    openAt: 100,
    *open() {
      // resume the source stream
    },
  })(stream);

  for (const value of yield* each(regulated)) {
    console.log(value);
    yield* each.next();
  }
}

Drain

The drain helper exhausts a stream, discarding all yielded values, and returns the close value. This is useful when you only care about the final result of a stream, not the intermediate values.

import { drain } from "@effectionx/stream-helpers";

function* example() {
  // Get the response from a request channel (ignoring any progress)
  const channel = yield* transport.send(request);
  const response = yield* drain(channel);
  console.log(response); // The close value
}

First

The first helper returns the first value yielded by a stream, or undefined if the stream closes without yielding any values. Use first.expect() to throw an error instead of returning undefined.

import { first } from "@effectionx/stream-helpers";
import { streamOf } from "@effectionx/stream-helpers";

function* example() {
  const stream = streamOf([1, 2, 3]);
  const value = yield* first(stream);
  console.log(value); // 1

  const empty = streamOf([]);
  const none = yield* first(empty);
  console.log(none); // undefined

  // Use first.expect() to throw if stream is empty
  const required = yield* first.expect(stream); // throws if empty
}

Last

The last helper returns the last value yielded by a stream, or undefined if the stream closes without yielding any values. It exhausts the entire stream to find the last value. Use last.expect() to throw an error instead of returning undefined.

import { last } from "@effectionx/stream-helpers";
import { streamOf } from "@effectionx/stream-helpers";

function* example() {
  const stream = streamOf([1, 2, 3]);
  const value = yield* last(stream);
  console.log(value); // 3

  const empty = streamOf([]);
  const none = yield* last(empty);
  console.log(none); // undefined

  // Use last.expect() to throw if stream is empty
  const required = yield* last.expect(stream); // throws if empty
}

Take

The take helper creates a stream transformer that yields the first n values from the source stream, then closes with the last taken value.

import { take, forEach, streamOf } from "@effectionx/stream-helpers";
import { pipe } from "remeda";

function* example() {
  const stream = streamOf([1, 2, 3, 4, 5]);
  // yields 1, 2, then closes with 3
  const closeValue = yield* forEach(function* (value) {
    console.log(value); // 1, then 2
  }, take(3)(stream));
  console.log(closeValue); // 3
}

// Works with pipe
const limited = pipe(source, take(3));

TakeWhile

The takeWhile helper creates a stream transformer that yields values while the predicate returns true. When the predicate returns false, the stream closes immediately (the failing value is not included).

import { takeWhile, forEach, streamOf } from "@effectionx/stream-helpers";
import { pipe } from "remeda";

function* example() {
  const stream = streamOf([1, 2, 3, 4, 5]);
  // yields 1, 2 (stops when value >= 3)
  yield* forEach(function* (value) {
    console.log(value); // 1, then 2
  }, takeWhile((x) => x < 3)(stream));
}

// Works with pipe
const filtered = pipe(source, takeWhile((x) => x.isValid));

TakeUntil

The takeUntil helper creates a stream transformer that yields values until the predicate returns true. When the predicate matches, the stream closes with the matching value. This is useful for "iterate until a condition is met" patterns.

import { takeUntil, forEach } from "@effectionx/stream-helpers";
import { pipe } from "remeda";

function* example() {
  // Iterate validation progress until we get a terminal status
  const result = yield* forEach(
    function* (progress) {
      showSpinner(progress.status);
    },
    takeUntil((p) => p.status === "valid" || p.status === "invalid")(
      validationStream,
    ),
  );

  // result is the validation object with terminal status
  if (result.status === "valid") {
    console.log("Validation passed!");
  }
}

// Works with pipe
const untilDone = pipe(source, takeUntil((x) => x.done));

Key difference between takeWhile and takeUntil:

HelperYieldsCloses WithUse Case
takeWhileWhile predicate is trueundefined"Keep going while good"
takeUntilUntil predicate is trueMatching value"Stop when you find it"

ForEach

The forEach helper invokes a function for each item passing through a stream. This is useful when you need to perform side effects or operations on each item without transforming the stream itself. Unlike other stream helpers that return transformed streams, forEach consumes the entire stream and returns the stream's close value.

import { forEach } from "@effectionx/stream-helpers";
import { createSignal, spawn } from "effection";

function* example() {
  const stream = createSignal<number, void>();

  // Process each item in the stream
  yield* spawn(() =>
    forEach(function* (item) {
      console.log(`Processing: ${item}`);
      // Perform any side effects here
    }, stream)
  );

  stream.send(1);
  stream.send(2);
  stream.send(3);
  stream.close();
}

// Example: Handling stream close value
function* exampleWithCloseValue() {
  const stream = createSignal<string, number>();

  const result = yield* spawn(() =>
    forEach(function* (item) {
      console.log(`Item: ${item}`);
    }, stream)
  );

  stream.send("hello");
  stream.send("world");
  stream.close(42); // Close with value 42

  const closeValue = yield* result;
  console.log(`Stream closed with: ${closeValue}`); // 42
}

Subject

Subject helper converts any stream into a multicast stream that replays the latest value to new subscribers. It's analogous to RxJS BehaviorSubject.

import { createSubject } from "@effectionx/stream-helpers";
import { createChannel, spawn } from "effection";

function* example() {
  const subject = createSubject<number>();
  const channel = createChannel<number, void>();
  const downstream = subject(channel);

  // First subscriber
  const sub1 = yield* downstream;

  yield* channel.send(1);
  yield* channel.send(2);

  console.log(yield* sub1.next()); // { done: false, value: 1 }
  console.log(yield* sub1.next()); // { done: false, value: 2 }

  // Late subscriber gets the latest value immediately
  const sub2 = yield* downstream;
  console.log(yield* sub2.next()); // { done: false, value: 2 }
}

Use it with a pipe operator to convert any stream into a behavior subject:

import { createSubject, map } from "@effectionx/stream-helpers";
import { pipe } from "remeda";

const subject = createSubject<string>();

const stream = pipe(
  source,
  map(function* (x) {
    return x.toString();
  }),
  subject,
);

Passthrough Tracker

Passthrough Tracker stream helper provides a way to know if all items that passed through the stream have been handled. This is especially helpful when you want to ensure that all items were processed before completing an operation.

It's different from other stream helpers because you must first call createTracker function which retuns an object. The actual helper is on the passthrough method which you can call and chain as you would with other helpers.

import { each, signal } from "effection";
import { createTracker } from "@ffectionx/stream-helpers"

const source = signal(0);

// create the tracker
const tracker = yield* createTracker();

// create  passthrough stream helper
const track = tracker.passthrough();

for (const value of yield* each(track(source))) {
  // mark items 
  tracker.markOne(value);
  yield* each.next();
}

// will resolve when all items that passed through the stream were seen
yield* tracker;

Composing stream helpers

You can use a simple pipe() to compose a series of stream helpers together. In this example, we use one from remeda,

import { batch, filter, forEach, map, valve } from "@effectionx/stream-helpers";
// any standard pipe function should work
import { pipe } from "remeda";

function* example(source: Stream<number, unknown>) {
  // Compose stream helpers using pipe
  const stream = pipe(
    source,
    valve({ open, close, openAt: 100, closeAt: 100 }),
    filter(function* (x) {
      return x > 0;
    }),
    map(function* (x) {
      return x * 20;
    }),
    batch({ maxSize: 50 }),
  );

  // Use forEach to process each value in the composed stream
  yield* forEach(function* (value) {
    console.log(value);
  }, stream);
}

Testing Streams

The library includes testing utilities to help you test your stream processing code. These are available in @effectionx/stream-helpers/test-helpers export.

Faucet

The useFaucet function creates a stream that can be used to test the behavior of streams that use backpressure. It's particularly useful in tests where you need a controllable source stream.

import { useFaucet } from "@effectionx/stream-helpers/test-helpers";
import { each, run, spawn } from "effection";

await run(function* () {
  const faucet = yield* useFaucet<number>({ open: true });

  // Remember to spawn the stream subscription before sending items to the stream
  yield* spawn(function* () {
    for (let i of yield* each(faucet)) {
      console.log(i);
      yield* each.next();
    }
  });

  // Pass an array of items to send items to the stream one at a time synchronously
  yield* faucet.pour([1, 2, 3]);

  // Pass an operation to control the rate at which items are sent to the stream
  yield* faucet.pour(function* (send) {
    yield* sleep(10);
    yield* send(5);
    yield* sleep(30);
    yield* send(6);
    yield* sleep(10);
    yield* send(7);
  });

  // You can close the faucet to stop items from being sent
  faucet.close();

  // And open it again when needed
  faucet.open();
});

When passing a function to faucet.pour, the send function will return an operation. This operation will block when the faucet is closed. This is particularly helpful when testing backpressure, because you can send many event without worrying about the open/close state.

版本与发布记录

@effectionx/stream-helpers 共发布 16 个版本,首次发布于 2025/05/19;过去 12 个月发布了 12 个版本。最新版本附带 npm provenance 构建来源证明,可追溯到公开的 CI 构建记录。

版本发布日期距今
0.8.3latest2026/06/174 个月前
0.8.22026/03/237 个月前
0.8.12026/02/248 个月前
0.8.02026/02/208 个月前
0.7.32026/02/158 个月前
0.7.22026/02/129 个月前

关于 @effectionx/stream-helpers 的常见问题

@effectionx/stream-helpers 是什么?

Type-safe stream operators like filter, map, reduce, and forEach

如何安装 @effectionx/stream-helpers?

在项目目录运行 npm install @effectionx/stream-helpers,或使用 pnpm add / yarn add / bun add。它还需要宿主项目提供 peer 依赖:effection(^3 || ^4)。npm 7 及以上与 pnpm 8 及以上默认会自动安装必需的 peer 依赖,Yarn 需要手动安装。

@effectionx/stream-helpers 支持 TypeScript 吗?

支持。包内自带 TypeScript 类型声明(package.json 声明了 types 入口),安装后即可获得类型提示,无需再安装 @types 包。

@effectionx/stream-helpers 支持 ESM 还是 CommonJS?

以 ES Module 形式发布,推荐使用 import 引入。在 CommonJS 项目中,Node.js 20.19+ / 22.12+ 可以直接 require() 不含顶层 await 的 ES Module,更早的版本需要改用动态 import()。

@effectionx/stream-helpers 可以商用吗?

@effectionx/stream-helpers 使用 MIT 许可。宽松许可:允许商业使用、修改和再分发,通常只需在分发时保留原版权与许可声明。以上为概括,不构成法律意见,请以许可原文为准。

@effectionx/stream-helpers 的下载量有多少?

根据 npm 官方下载统计,@effectionx/stream-helpers 在 2026/10/13 – 2026/11/11 这 30 天内被下载 47,680 次,日均约 1,589 次。下载量会把 CI 构建、镜像同步和作为其他包依赖的安装一并计入,适合横向比较热度,不能直接等同于使用人数。

@effectionx/stream-helpers 还在维护吗?

最新版本 0.8.3 发布于 2026/06/17(4 个月前);至今共发布 16 个版本,其中过去 12 个月发布了 12 个;首次发布于 2025/05/19。发布频率只是参考信号:功能稳定的工具可能很少需要更新,建议结合源码仓库的 issue 和提交记录判断。