hyperfilter

Distributed map and filter. Useful for filtering distributed streams of data, similar to Apache Spark, but implemented in ~16 lines of code.

Installation

$ npm install hyperfilter

Usage

var hypercore = require('hypercore')
var filter = require('hyperfilter')
var memdb = require('memdb')

var core = hypercore(memdb())
var inp = core.createFeed()
var out = core.createFeed()

// uppercase all input, filter out 'hello world 1'
filter(inp, out, function (data, cb) {
  if (data.toString() === 'hello world 1') return cb()
  cb(null, data.toString().toUpperCase())
})

inp.append(new Buffer('hello world 0'))
inp.append(new Buffer('hello world 1'))
inp.append(new Buffer('hello world 2'))

out.createReadStream({live: true}).pipe(process.stdout)
// => HELLO WORLD 0
// => HELLO WORLD 2

Why

We believe ops doesn't need to be complicated. If hypercore is distributed streams, hyperfilter is a distributed filter for streams. We needed this to to turn our feed of server logs into a feed of server errors.

API

filter(inputFeed, outputFeed, mapFunction(data, next))

Create a new filter function that reads data from an input hypercore feed to and writes it to an output hypercore feed. Each chunk of data is passed throug the map function. The next function has a signature of err, data. If neither an error or data is written, it skips to the next chunk, acting like a filter function.

See Also

• https://github.com/mafintosh/hypercore
• https://github.com/mafintosh/hyperpipe
• https://github.com/mafintosh/hypername
• https://github.com/yoshuawuyts/hypertail

License

MIT