peek-stream

What is peek-stream?

The peek-stream npm package is a Node.js module that allows you to peek at the beginning of a stream to determine its content before deciding how to process it. This can be useful when you need to inspect the start of a stream to make decisions based on its initial data without consuming the stream entirely.

What are peek-stream's main functionalities?

Peeking at stream content

This code sample demonstrates how to use peek-stream to inspect the first 10 bytes of a file stream. If the file starts with 'PNG', it swaps the stream with a new PngDecoderStream; otherwise, it continues with the original stream.

const peek = require('peek-stream');
const fs = require('fs');

const stream = fs.createReadStream('example.txt');
const peekStream = peek({ maxBuffer: 10 }, function(data, swap) {
  // data is the first 10 bytes of the stream
  if (data.toString().startsWith('PNG')) {
    swap(null, new PngDecoderStream());
  } else {
    swap();
  }
});

stream.pipe(peekStream).pipe(process.stdout);

Other packages similar to peek-stream

through2

Through2 is a tiny wrapper around Node.js streams.Transform that makes it easier to create transform streams. It is similar to peek-stream in that it allows you to manipulate stream data, but it does not provide the same functionality to peek at the stream before deciding how to process it.

stream-spigot

Stream-spigot is a module for creating readable streams that emit data from a given array or function. While it can be used to create streams with predetermined data, unlike peek-stream, it does not offer the ability to peek into an existing stream to make processing decisions.

pumpify

Pumpify combines an array of streams into a single duplex stream using pump and duplexify. It is useful for streamlining the process of piping streams together but does not have the peeking capability that peek-stream provides.

README

peek-stream

Transform stream that lets you peek the first line before deciding how to parse it

npm install peek-stream

Usage

var peek = require('peek-stream')
var ldjson = require('ldjson-stream')
var csv = require('csv-parser')

var isCSV = function(data) {
  return data.toString().indexOf(',') > -1
}

var isJSON = function(data) {
  try {
    JSON.parse(data)
    return true
  } catch (err) {
    return false
  }
}

// call parser to create a new parser
var parser = function() {
  return peek(function(data, swap) {
    // maybe it is JSON?
    if (isJSON(data)) return swap(null, ldjson())

    // maybe it is CSV?
    if (isCSV(data)) return swap(null, csv())

    // we do not know - bail
    swap(new Error('No parser available'))
  })
}

The above parser will be able to parse both line delimited JSON and CSV

var parse = parser()

parse.write('{"hello":"world"}\n{"hello":"another"}\n')
parse.on('data', function(data) {
  console.log(data) // prints {hello:'world'} and {hello:'another'}
})

Or

var parse = parser()

parse.write('test,header\nvalue-1,value-2\n')
parse.on('data', function(data) {
  console.log(data) // prints {test:'value-1', header:'value-2'}
})

Per default data is the first line (or the first 65535 bytes if no newline is found). To change the max buffer pass an options map to the constructor

var parse = peek({
  maxBuffer: 10000
}, function(data, swap) {
  ...
})

If you want to emit an error if no newline is found set strict: true as well.

License

MIT