5m 中文文档教程
5m
是一个接近实时的实用程序,5mb 或 5 分钟
Installation
npm i 5m
Use case
然后有一个触发器将在 S3 上创建的每个文件提供给 AWS Lambda 它将文件内容发送到使用 socket.io 广播它的服务器。 是的,我知道还有其他很棒的工具,如 CloudWatch、Kinesis 等。 (但是 5m 在演讲中是免费的,在啤酒中也是免费的;)
很酷的部分是每次都可以触发 AWS Lambda 创建一个新文件,例如将其加载到数据库中。
下面是一个示例代码,用于在 S3 上实现日志记录和上传。
// File: log.js
//
// Prints to STDOUT if DEBUG environment variable is set properly.
// Uploads logs on some S3 bucket.
//
// const log = require('./log')
// log('Hello world')
//
const AWS = require('aws-sdk')
const debug = require('debug')
const fiveM = require('5m')
const s3 = new AWS.S3()
// Use package name as namespace for logging, just as an example.
const pkg = require('./package.json')
const namespace = pkg.name
const debug5m = debug(`${namespace}:5m`)
const debugPkg = debug(namespace)
const writeOnS3 = (data) => {
const Bucket = 'my-bucket'
const tstamp = new Date().getTime()
const Key = `my/path/${tstamp}.log`
s3.upload({ Bucket, Key }, error => {
if (error) debug5m(error)
else debug5m()
})
}
function log (message) {
const tstamp = new Date().getTime()
const record = `${tstamp} ${namespace} ${message}`
fiveM(namespace, writeOnS3)(record)
debugPkg(message)
}
module.exports = log
Annotated source
将数据存储在命名空间的桶中,以及其刷新数据的功能。
var bucket = {}
var flush = {}
考虑到 1 个字符应该是 1 个字节,并且日期以毫秒表示, 以下常量表示 5 MB 和 5 分钟。 近似是可以的,因为我们想要实现近实时。 可以设置用于刷新数据的超时时间,用于测试或其他目的 通过 FIVEM_TIMEOUT_MILLISECONDS
环境变量。
const fiveMb = 1024 * 1024 * 5
const fiveMin = 300 * 1000
const flushTimeout = process.env.FIVEM_TIMEOUT_MILLISECONDS || fiveMin
确保退出时没有数据丢失。
process.on('exit', () => {
for (var namespace in bucket) flush[namespace]()
})
创建 5m 函数。
因为 5m 可以作为 npm 包名,但不能作为 JavaScript 标识符,使用像 Vm 这样的罗马数字可能会造成混淆,所以也许 将函数命名为 fiveM 是个好主意。
/**
* **@param** `{String}` namespace
* **@param** `{Function}` write
*
* **@returns** `{Function}` logger
*/
function fiveM (namespace, write) {
创建命名空间 flush 函数:写入数据并清理。
flush[namespace] = () => {
if (bucket[namespace]) {
write(bucket[namespace])
delete bucket[namespace]
}
}
创建记录器 函数。
/**
* @param {*} data
*/
return function logger (data) {
如有必要,初始化数据桶并设置超时以稍后刷新它。
if (typeof bucket[namespace] === 'undefined') {
bucket[namespace] = ''
setTimeout(flush[namespace], flushTimeout)
}
将数据附加到命名桶。
bucket[namespace] += data
检查数据是否大于 5 MB。
const exceededSpace = bucket[namespace] && (bucket[namespace].length > 0) && (bucket[namespace].length > fiveMb)
如果是,请冲洗它!
if (exceededSpace) flush[namespace]()
}
}
导出 5m 函数。
module.exports = fiveM
License
5m
is a near real time utility, 5mb or 5min
Installation | Use case | Annotated source | License
Installation
npm i 5m
Use case
I use this package with debug and aws-sdk to upload logs on S3.
Then there is a trigger that feeds every file created on S3 to an AWS Lambda which sends file content to a server that broadcast it using socket.io. Yes I know there are other cool tools like CloudWatch, Kinesis, etc. (but 5m is free as in speach and free as in beer too ;)
The cool part is that it is possible to trigger an AWS Lambda every time a new file is created, and for example load it on a database.
Here is an example code to achieve logging and upload on S3.
// File: log.js
//
// Prints to STDOUT if DEBUG environment variable is set properly.
// Uploads logs on some S3 bucket.
//
// const log = require('./log')
// log('Hello world')
//
const AWS = require('aws-sdk')
const debug = require('debug')
const fiveM = require('5m')
const s3 = new AWS.S3()
// Use package name as namespace for logging, just as an example.
const pkg = require('./package.json')
const namespace = pkg.name
const debug5m = debug(`${namespace}:5m`)
const debugPkg = debug(namespace)
const writeOnS3 = (data) => {
const Bucket = 'my-bucket'
const tstamp = new Date().getTime()
const Key = `my/path/${tstamp}.log`
s3.upload({ Bucket, Key }, error => {
if (error) debug5m(error)
else debug5m()
})
}
function log (message) {
const tstamp = new Date().getTime()
const record = `${tstamp} ${namespace} ${message}`
fiveM(namespace, writeOnS3)(record)
debugPkg(message)
}
module.exports = log
Annotated source
Store data in a namespaced bucket, as well as its function to flush data.
var bucket = {}
var flush = {}
Considering 1 char should be 1 byte, and dates are expressed in milliseconds, the following constants express 5 MB and 5 minutes. Aproximation is ok, since we want to achieve a near real time. The timeout used for flushing data can be set, for testing or other purpouse via the FIVEM_TIMEOUT_MILLISECONDS
environment variable.
const fiveMb = 1024 * 1024 * 5
const fiveMin = 300 * 1000
const flushTimeout = process.env.FIVEM_TIMEOUT_MILLISECONDS || fiveMin
Make sure no data is lost on exit.
process.on('exit', () => {
for (var namespace in bucket) flush[namespace]()
})
Create the 5m function.
Since 5m is allowed as npm package name, but not as JavaScript identifier, using a roman number like Vm can be confusing, so maybe naming the function as fiveM is a good idea.
/**
* **@param** `{String}` namespace
* **@param** `{Function}` write
*
* **@returns** `{Function}` logger
*/
function fiveM (namespace, write) {
Create the namespaced flush function: write data and clean up.
flush[namespace] = () => {
if (bucket[namespace]) {
write(bucket[namespace])
delete bucket[namespace]
}
}
Create the logger function.
/**
* @param {*} data
*/
return function logger (data) {
If necessary, initialize data bucket and set timeout to flush it later.
if (typeof bucket[namespace] === 'undefined') {
bucket[namespace] = ''
setTimeout(flush[namespace], flushTimeout)
}
Append data to named bucket.
bucket[namespace] += data
Check if data is bigger than 5 MB.
const exceededSpace = bucket[namespace] && (bucket[namespace].length > 0) && (bucket[namespace].length > fiveMb)
If yes, flush it!
if (exceededSpace) flush[namespace]()
}
}
Export 5m function.
module.exports = fiveM