English | 简体中文
Redis database components for coajs, including data cache, message queue, timing task, distributed lock, etc.
- Functional: Basic data connection based on ioredis. Pay attention to performance, full-featured
- Lightweight: Only hundreds of lines of code, do not rely on other third-party libraries
- TypeScript: All written in TypeScript, type constraint, IDE friendship
- Data Cache RedisCache: Data cache for Key-Value type
- Message Queue RedisQueue RedisQueueWorker: Lightweight and efficient message queue based on Redis implementation
- Timing Task RedisCron: Perform tasks on time through the cron expression and message queue
- Distributed Lock RedisLock: Basic distributed lock mechanism by Redis
yarn add coa-redis
import { RedisBin } from 'coa-redis'
const redisConfig = {
// host address
host: '127.0.0.1',
// port
port: 6379,
// password, if not write an empty string
password: '123456',
// database, default is 0
db: 0,
// Key prefix, distinguish between different project
prefix: 'pre_',
// Whether to display a query statement
trace: false,
}
// Create a configuration instance, follow-up all components depending on this configuration instance
// Generally a database only needs to use an instance, the internal management connection pool
const redisBin = new RedisBin(redisConfig)
Basic usage
// Create a cache instance
const redisCache = new RedisCache(redisBin)
// Note: In order to constrain the isolation of module data, all operations of the cache instance need to pass nsp parameters and cannot be omitted.
// Generally, the module name can be used as nsp, and the data between each nsp is isolated from each other.
// Set cache data
await redisCache.set(
'module1',
'id001',
'value001',
5 * 60 * 1000 /*5 minutes*/
) // 1
// Read cache data
await redisCache.get('module1', 'id001') // value001
// Delete cache data (support to delete multiple data under the same nsp)
await redisCache.delete('module1', ['id001', 'id002']) // 2
Batch operate
// Batch set cache data
await redisCache.mSet(
'module1',
{ id101: 'value101' },
5 * 60 * 1000 /*5 minutes*/
) // 1
await redisCache.mSet(
'module2',
{ id201: 'value201', id202: { name: 'A2', title: 'a2' } },
5 * 60 * 1000 /*5 minutes*/
) // 2
// Batch read cache data
await redisCache.mGet('module1', ['id101']) // 'value101'
await redisCache.mGet('module2', ['id201', 'id202']) // { id201: 'value201', id202: { name: 'A2', title: 'a2' }
// Batch delete cache data (support to delete multiple data under different nsp)
await redisCache.mDelete([
['module1', ['id101']],
['module2', ['id201', 'id202']],
]) // 3
Syntactic sugar
// Get cache data, if there is no existence, follow the method to read and save
const resultWarp1 = await redisCache.warp(
'module1',
'id301',
() => {
// Do something here
return Math.random() // return result
},
10 * 60 * 1000 /*10 minutes*/
)
resultWarp1 // return 0.3745813097015189 with in 10 minutes
// This Syntactic sugar is equivalent to the following
async function getAndSetCache(nsp: string, id: string) {
let result = await redisCache.get(nsp, id)
if (result === undefined) {
// Do something here
result = Math.random() // get result
await redisCache.set(nsp, id, result, 10 * 60 * 1000 /*10 minutes*/)
}
return result
}
// Batch get cache data
const resultWarp2 = await redisCache.mWarp(
'module1',
['id301', 'id302'],
(ids) => {
const result = {} as { [id: string]: number }
for (const id of ids) {
// Do something here
result[id] = Math.random()
}
return result // Return the result, be sure to ensure the key value
},
10 * 60 * 1000 /*10 minutes*/
)
resultWarp2 // { id301: 0.32430600236596074, id302: 0.29829421673682566 }
// Define a queue name and message type name
const QUEUE_NAME_1 = 'CHANNEL-1',
MESSAGE_NAME_1 = 'NORMAL-MESSAGE-1'
// Define a message queue
const queue = new RedisQueue(redisBin, QUEUE_NAME_1)
// Define the consumer of the queue
const worker = new RedisQueueWorker(queue)
worker.on(MESSAGE_NAME_1, async (id, data) => {
console.log(`message id is ${id}, message included with ${data}`)
})
// Produce a message
await queue.push(MESSAGE_NAME_1, 'message-id-001', { value: '001' }) // 1
// Timing task relies on the consumer of the message queue, and at the same time, you must specify the version number to avoid version upgrade instant task conflicts
const cron = new RedisCron(quque.worker, env.version)
// Execute at 10:00 and 16:00 each day
cron.on('0 0 10,16 * * *', () => {
/**Do something**/
})
// Execute at 0:30 each day
cron.on('0 30 0 * * *', () => {
/**Do something**/
})
// Execute every 10 minutes
cron.on('0 */10 * * * *', () => {
/**Do something**/
})
// Execute at 0:00 on every Monday, Wednesday, Friday
cron.on('0 0 0 * * 1,3,5', () => {
/**Do something**/
})
// Execute at 0:00 on every month of 1 day and 16
cron.on('0 0 0 1,16 * *', () => {
/**Do something**/
})
// Create a lock method instance
const redisLock = new RedisLock(redisBin)
// Blocking execution, if the call is repeated, will wait for the last call to complete
await redisLock.start('lock-for-user-register', () => {
// Do something, this thing will not be executed concurrently
})
// Try to execute, if it is repeated call, it will throw the RedisLock.Running error
await redisLock.try('lock-for-user-register', () => {
// Do something, this thing will not be executed concurrently
})
// The throttle is executed, only 1 second is allowed once, and the repeat call will be queued to wait for the next 1 second.
await redisLock.throttle(
'lock-for-user-register',
() => {
// Do something, this thing will not be executed concurrently
},
1000 /*1 second*/
)