How to compress a string?

I would like to have a reversible compression for a type of string so that i can include it in URLs without keeping track of what it refers to. The string i would like to compress is SVG path string, here is a short primer: http://apike.ca/prog_svg_paths.html

Basically, the s开发者_开发知识库tring contains a character, followed by arbitrary number of integers, then another character followed by arbitrary number of integers and so on.

If anyone knows of a good resource for this, it would be much appreciated!

Jason

Many compression algorithms are well documented, a couple even have js implementations:

• GZip A common (reasonably) good compression algorithm, I know there's a JS impl, i'm just hunting the URL

• LZW Another question points to an LZW implementation in JS

• Arithmetic coding (i did this, but the model it uses is stupid so doesn't achieve the best compression rates it could)

The following solution returns a compressed Base64 encoded string.

Create a file called zip.js with the code below and then see usage below that.

// Apply LZW-compression to a string and return base64 compressed string.
export function zip (s) {
  try {
    var dict = {}
    var data = (s + '').split('')
    var out = []
    var currChar
    var phrase = data[0]
    var code = 256
    for (var i = 1; i < data.length; i++) {
      currChar = data[i]
      if (dict[phrase + currChar] != null) {
        phrase += currChar
      } else {
        out.push(phrase.length > 1 ? dict[phrase] : phrase.charCodeAt(0))
        dict[phrase + currChar] = code
        code++
        phrase = currChar
      }
    }
    out.push(phrase.length > 1 ? dict[phrase] : phrase.charCodeAt(0))
    for (var j = 0; j < out.length; j++) {
      out[j] = String.fromCharCode(out[j])
    }
    return utoa(out.join(''))
  } catch (e) {
    console.log('Failed to zip string return empty string', e)
    return ''
  }
}

// Decompress an LZW-encoded base64 string
export function unzip (base64ZippedString) {
  try {
    var s = atou(base64ZippedString)
    var dict = {}
    var data = (s + '').split('')
    var currChar = data[0]
    var oldPhrase = currChar
    var out = [currChar]
    var code = 256
    var phrase
    for (var i = 1; i < data.length; i++) {
      var currCode = data[i].charCodeAt(0)
      if (currCode < 256) {
        phrase = data[i]
      } else {
        phrase = dict[currCode] ? dict[currCode] : oldPhrase + currChar
      }
      out.push(phrase)
      currChar = phrase.charAt(0)
      dict[code] = oldPhrase + currChar
      code++
      oldPhrase = phrase
    }
    return out.join('')
  } catch (e) {
    console.log('Failed to unzip string return empty string', e)
    return ''
  }
}

// ucs-2 string to base64 encoded ascii
function utoa (str) {
  return window.btoa(unescape(encodeURIComponent(str)))
}
// base64 encoded ascii to ucs-2 string
function atou (str) {
  return decodeURIComponent(escape(window.atob(str)))
}

Usage:

import { zip, unzip } from './zip'

// Zip a string
const str = 'zip it'
const base64CompressedString = zip(str)

// Zip an object
const obj = { a: 123, b: 'zipit' }
const base64CompressedString = zip(JSON.stringify(obj))

// Unzip the base64 compressed string back to an object.
const originalObject = JSON.parse(unzip(base64CompressedString))

BTW... if you're concerned about escape/unescape being depreciated consider a polyfill

LZW algorithm from here and base64 encoding from here

Sounds like you might benefit from single and double RLE compression.

A primer on this can be seen here:

http://pp19dd.com/2011/10/query-string-limits-encoding-hundreds-of-checkboxes-with-rle/#demo

The library should be flexible enough to modify your compression pattern to something more preferable. The writeup explains how this works; might be a good start to optimize your SVG case.

You could try Huffman compression. Number of different chars is 20-30, and if the string is long, compression should be effective.