92green-rxjs 中文文档教程
92green-rxjs
yarn add 92green-rxjs
Rxjs 算法。
- multiCache
- memoryCache
- zipDiff
- operators/bufferDistinct
- operators/complete
- dynamodb/batchGetWithRetry
- dynamodb/batchWriteWithRetry
- dynamodb/queryAll
- dynamodb/scanAll
multiCache
接受一组源(旨在成为内存缓存、数据库缓存和实际数据源之类的东西),并针对每个项目尝试按顺序从每个源检索项目,直到找到为止。 一旦找到,它还会将项目保存在缓存中。 它不保证商品顺序。
import {multiCache} from '92green-rxjs';
let myMemoryCacheOperator = {
name: 'memory-cache',
load, // operator that loads from cache
save, // operator that saves to cache (not needed on data source)
clear, // operator that clears item from cache (not needed on data source)
};
// ^ all of the above operators must accept a payload of {id, item} and return {id, item}
// item may be undefined if not known or not found
let cache = multiCache([
myMemoryCacheOperator, // first cache
myDatabaseCacheOperator, // deeper cache
myDataSource // data source
])
from([
{id: 'a'},
{id: 'b'},
{id: 'c'}
])
.pipe(cache.load);
// Example output observable:
// {id: 'a', item: {id: 'a', name: 'Hi'}, source: 'data-source'}
// {id: 'b', item: {id: 'b', name: 'Hello'}, source: 'memory-cache'}
// {id: 'c', item: undefined, source: undefined}
memoryCache
一个 rx 缓存,其工作方式很像 graphql/dataloader。
与 multiCache
一起使用:
import {from} from 'rxjs'
import {flatMap} from 'rxjs/operators'
import {multiCache} from '92green-rxjs';
import {memoryCache} from '92green-rxjs';
let dataSource = {
name: 'data-source',
load: flatMap(async (payload) => ({
...payload,
item: await fetch(payload.id)
}))
};
let cache = multiCache([
memoryCache(),
dataSource
]);
from([
{id: 'a'},
{id: 'b'},
{id: 'a'}
])
.pipe(cache.load);
// Example output observable:
// {id: 'a', item: {id: 'a', name: 'Hi'}, source: 'data-source'}
// {id: 'b', item: {id: 'b', name: 'Hello'}, source: 'data-source'}
// {id: 'a', item: {id: 'a', name: 'Hi'}, source: 'memory-cache'}
// a is only requested once
zipDiff
获取两个可观察对象,并根据键查找其中一个或两个对象中存在哪些项目。 它发出的项目与在一个对象上压缩在一起的匹配项目:{a: itemA, b: itemB}
, 并在具有单个键的对象上发出不匹配的项,如 {a: itemA}
、{b: itemB}
。
keyBy
函数在每个项目上被调用。 返回的数据用作标识符 当两个项目匹配时计算出来。
如果您需要为第二个可观察对象使用不同的 keyBy 函数,则可以添加第二个 keyBy
函数。
项目会尽早发出,因此内部缓冲区可以尽可能小。 成对的匹配项在匹配时立即发出,不匹配的项在知道不可能匹配时立即发出。 (例如,如果缓冲区包含来自 A 的项目,并且输入 observable B 完成,则已知永远不会匹配任何缓冲的 A,因此它们都被发射)
import {zipDiff} from '92green-rxjs';
zipDiff(
obsA: Observable,
obsB: Observable,
keyBy: (item) => any,
keyByB?: (item) => any
): Observable
zipDiff(obsA, obsB, itemA => itemA.id)
// obsA adds {id: "1", name: "One"}
// obsA adds {id: "2", name: "Two"}
// obsB adds {id: "3", name: "Three"}
// obsB adds {id: "2", name: "Too"}
// output:
{
a: {id: "2", name: "Two"},
b: {id: "2", name: "Too"}
}
{
a: {id: "1", name: "One"}
}
{
b: {id: "3", name: "Three"}
}
operators/bufferDistinct
缓冲项目,同时给定的谓词函数返回相同的东西,并发射当给定的谓词函数返回其他内容或流关闭时缓冲区内容。 使用 Object.is()
执行值比较。
如果传递了 flushObs
,则每次 flushObs
发出时都会刷新缓冲区。
import {bufferDistinct} from '92green-rxjs/operators';
bufferDistinct(item => comparisonValue, flushObs: ?Observable)
obs.pipe(
bufferDistinct(item => item.id)
);
// Obs adds {id: 'a', value: 1}
// Obs adds {id: 'a', value: 2}
// Obs adds {id: 'b', value: 3}
// Obs adds {id: 'c', value: 4}
// Obs adds {id: 'c', value: 5}
// output:
[
{id: 'a', value: 1},
{id: 'a', value: 2}
]
[
{id: 'b', value: 3}
]
[
{id: 'c', value: 4},
{id: 'c', value: 5}
]
operators/complete
在可观察对象完成时发出单个项目。
import {complete} from '92green-rxjs/operators';
complete()
obs.pipe(
complete(),
tap(() => {
console.log("I'm done mate");
})
);
dynamodb/batchGetWithRetry
将 AWS DocClient.batchGet()
变成一个管道可观察对象,它接受 id 的可观察对象并调用 batchGet()
,一次将项目批处理到 100 个并重试 dynamo 的 UnprocessedKeys
自动。
import {batchGetWithRetry} from '92green-rxjs/dynamodb';
batchGetWithRetry({
docClient: DocClient,
tableName: string,
returnItems?: boolean
}): Observable => Observable
let keys = [
{id: 1},
{id: 2},
{id: 3}
];
from(keys)
.pipe(batchGetWithRetry(...))
.toPromise();
dynamodb/batchWriteWithRetry
将 AWS DocClient.batchWrite()
变成一个管道可观察对象,它接受参数的可观察对象并调用 batchWrite()
,一次将项目批处理到 25 个并重试 dynamo 的 UnprocessedItems
自动。
import {batchWriteWithRetry} from '92green-rxjs/dynamodb';
batchWriteWithRetry({
docClient: DocClient,
tableName: string,
returnItems?: boolean
}): Observable => Observable
from([{
{
PutRequest: {
Item: {
foo: 200
}
}
}
...
}])
.pipe(batchWriteWithRetry(...))
.toPromise();
dynamodb/queryAll
将 AWS DocClient.query()
变成一个可观察对象,默认情况下,只要有更多数据需要分页,它就会继续请求。
import {queryAll} from '92green-rxjs/dynamodb';
queryAll(
docClient: DocClient,
params: Params,
feedbackObservable: ?Observable
): Observable
dynamodb/scanAll
将 AWS DocClient.scan()
变成一个可观察对象,默认情况下,只要有更多数据需要分页,它就会继续请求。
import {scanAll} from '92green-rxjs/dynamodb';
scanAll(
docClient: DocClient,
params: Params,
feedbackObservable: ?Observable
): Observable
92green-rxjs
yarn add 92green-rxjs
Rxjs algorithms.
- multiCache
- memoryCache
- zipDiff
- operators/bufferDistinct
- operators/complete
- dynamodb/batchGetWithRetry
- dynamodb/batchWriteWithRetry
- dynamodb/queryAll
- dynamodb/scanAll
multiCache
Accepts an array of sources (intended to be things like in-memory caches, database caches and actual data sources) and for each item attempts to retrieve the item from each source sequentially until it is found. It also saves items in caches once found. it does not guarantee item order.
import {multiCache} from '92green-rxjs';
let myMemoryCacheOperator = {
name: 'memory-cache',
load, // operator that loads from cache
save, // operator that saves to cache (not needed on data source)
clear, // operator that clears item from cache (not needed on data source)
};
// ^ all of the above operators must accept a payload of {id, item} and return {id, item}
// item may be undefined if not known or not found
let cache = multiCache([
myMemoryCacheOperator, // first cache
myDatabaseCacheOperator, // deeper cache
myDataSource // data source
])
from([
{id: 'a'},
{id: 'b'},
{id: 'c'}
])
.pipe(cache.load);
// Example output observable:
// {id: 'a', item: {id: 'a', name: 'Hi'}, source: 'data-source'}
// {id: 'b', item: {id: 'b', name: 'Hello'}, source: 'memory-cache'}
// {id: 'c', item: undefined, source: undefined}
memoryCache
An rx cache that works a lot like graphql/dataloader.
Usage with multiCache
:
import {from} from 'rxjs'
import {flatMap} from 'rxjs/operators'
import {multiCache} from '92green-rxjs';
import {memoryCache} from '92green-rxjs';
let dataSource = {
name: 'data-source',
load: flatMap(async (payload) => ({
...payload,
item: await fetch(payload.id)
}))
};
let cache = multiCache([
memoryCache(),
dataSource
]);
from([
{id: 'a'},
{id: 'b'},
{id: 'a'}
])
.pipe(cache.load);
// Example output observable:
// {id: 'a', item: {id: 'a', name: 'Hi'}, source: 'data-source'}
// {id: 'b', item: {id: 'b', name: 'Hello'}, source: 'data-source'}
// {id: 'a', item: {id: 'a', name: 'Hi'}, source: 'memory-cache'}
// a is only requested once
zipDiff
Takes two observables and finds which items exist in one or both, according to a key. It emits items with matching items zipped together on an object: {a: itemA, b: itemB}
, and emits unmatched items on objects with single keys like {a: itemA}
, {b: itemB}
.
The keyBy
function is called on each item. The data returned is used as an identifier to work out when two items are matching.
A second keyBy
function can be added if you need a different keyBy function for the second observable.
Items are emitted as early as they can be, so the internal buffer can try and be as small as possible. Pairs of matching items are emitted as soon as they match, and non-matching items are emitted as soon as it's known that it's impossible that there will be a match. (e.g. if the buffer contains items from A, and input observable B completes, then it's known there will never be matches for any buffered A's, so they are all emitted)
import {zipDiff} from '92green-rxjs';
zipDiff(
obsA: Observable,
obsB: Observable,
keyBy: (item) => any,
keyByB?: (item) => any
): Observable
zipDiff(obsA, obsB, itemA => itemA.id)
// obsA adds {id: "1", name: "One"}
// obsA adds {id: "2", name: "Two"}
// obsB adds {id: "3", name: "Three"}
// obsB adds {id: "2", name: "Too"}
// output:
{
a: {id: "2", name: "Two"},
b: {id: "2", name: "Too"}
}
{
a: {id: "1", name: "One"}
}
{
b: {id: "3", name: "Three"}
}
operators/bufferDistinct
Buffers items while the given predicate function returns the same thing, and emits the buffer contents when the given predicate function returns something else or the stream closes. Value comparison is performed using Object.is()
.
If flushObs
is passed, the buffer will be flushed any time flushObs
emits.
import {bufferDistinct} from '92green-rxjs/operators';
bufferDistinct(item => comparisonValue, flushObs: ?Observable)
obs.pipe(
bufferDistinct(item => item.id)
);
// Obs adds {id: 'a', value: 1}
// Obs adds {id: 'a', value: 2}
// Obs adds {id: 'b', value: 3}
// Obs adds {id: 'c', value: 4}
// Obs adds {id: 'c', value: 5}
// output:
[
{id: 'a', value: 1},
{id: 'a', value: 2}
]
[
{id: 'b', value: 3}
]
[
{id: 'c', value: 4},
{id: 'c', value: 5}
]
operators/complete
Emits a single item upon completion of the observable.
import {complete} from '92green-rxjs/operators';
complete()
obs.pipe(
complete(),
tap(() => {
console.log("I'm done mate");
})
);
dynamodb/batchGetWithRetry
Turns AWS DocClient.batchGet()
into a pipeable observable which accepts an observable of ids and calls batchGet()
, batching items to 100 at a time and retrying dynamo's UnprocessedKeys
automatically.
import {batchGetWithRetry} from '92green-rxjs/dynamodb';
batchGetWithRetry({
docClient: DocClient,
tableName: string,
returnItems?: boolean
}): Observable => Observable
let keys = [
{id: 1},
{id: 2},
{id: 3}
];
from(keys)
.pipe(batchGetWithRetry(...))
.toPromise();
dynamodb/batchWriteWithRetry
Turns AWS DocClient.batchWrite()
into a pipeable observable which accepts an observable of params and calls batchWrite()
, batching items to 25 at a time and retrying dynamo's UnprocessedItems
automatically.
import {batchWriteWithRetry} from '92green-rxjs/dynamodb';
batchWriteWithRetry({
docClient: DocClient,
tableName: string,
returnItems?: boolean
}): Observable => Observable
from([{
{
PutRequest: {
Item: {
foo: 200
}
}
}
...
}])
.pipe(batchWriteWithRetry(...))
.toPromise();
dynamodb/queryAll
Turns AWS DocClient.query()
into an observable which will by default keep requesting whenever there is more data to be paginated.
import {queryAll} from '92green-rxjs/dynamodb';
queryAll(
docClient: DocClient,
params: Params,
feedbackObservable: ?Observable
): Observable
dynamodb/scanAll
Turns AWS DocClient.scan()
into an observable which will by default keep requesting whenever there is more data to be paginated.
import {scanAll} from '92green-rxjs/dynamodb';
scanAll(
docClient: DocClient,
params: Params,
feedbackObservable: ?Observable
): Observable