node package manager



Build Status

written provides a set of utilities for manipulating text, with a focus on providing typographic tools rather than pure string manipulation. It can be added as a set of mixins to Underscore or used in it's own right, both in front and back end contexts.

npm install written
bower install written

This readme is also the source code for this module. Each function shows examples and the implementation.


written can be used as a module in Node and AMD contexts, and will otherwise be made available as a global variable (window.written).

((root, factory) ->
  if typeof define is "function" and define.amd
    define [], factory
  else if typeof exports is "object"
    module.exports = factory()
    root.written = factory()
) this, ->

Some style guides prefer the numbers 12 and under to be written, so we'll include those in here. If more or fewer numbers need to be added, or those from another language, see Language Support.

Following the APA style guide (for ease and practicality) conjunctions, articles, and short prepositions of less than four letters will be left in lowercase when calling capitalizeAll().

A rule is needed to determine the correct ordinal for any number. For English, we use match in such a way that the first value in the matching array is returned, unless it is 11, 12 or 13. We use this number to determine the correct ordinal form.

  dictionary =
      noncaps: ///^(

        written: ["one", "two", "three", "four",
                  "five", "six", "seven", "eight",
                  "nine", "ten", "eleven", "twelve"]

        written: ["first", "second", "third", "fourth",
                  "fifth", "sixth", "seventh", "eighth",
                  "ninth", "tenth", "eleventh", "twelfth"]

        rule:    /((1{0,1}[123])|(\d))\b/

          "1": "st"
          "2": "nd"
          "3": "rd"
          "n": "th"


Capitalize the first letter of a string.


w.capitalize("obviously")                        # Obviously 
  capitalize = (str) ->
    str.charAt(0).toUpperCase() + str.slice(1)

Capitalize all words in a string apart from some common lower case words. This can be tested with the internal noncaps regular expression, which are stored by language code, or by passing a regular expression of your own.


w.capitalizeAll("this and that")                 # This and That 
w.capitalizeAll("the cat in the hat")            # The Cat in the Hat 
  capitalizeAll = (str, regEx = dictionary["EN"].noncaps) ->
    unless is "[object RegExp]"
      regEx = dictionary[regEx].noncaps

    (for s, i in str.split(/\s/g)
      if i > 0 and regEx.test(s) then s else capitalize(s)
    ).join(" ")


enclose wraps a string within two other strings, repeating the first if needs be. cleanJoin joins an array of words with falsy, non-string values removed with some glue. Both are used internally but are offered in case of their external value.


w.enclose("'""string")                         # 'string' 
w.cleanJoin(["this"null"that"]" and ")     # this and that 
  enclose = (a, b, c) ->
    "#{a}#{b}#{c or a}"

  cleanJoin = (arr, glue = "") ->
    (a for a in arr when a and typeof a is "string").join(glue)


Replace all white-space in a string with a single space character


w.collapse("this   \t\t and \n    that")         # this and that 
  collapse = (str) ->
    str.replace(/\s+/g, " ")


Transform strings between common code cases.


w.camelCase("some-thing")                        # someThing 
w.hyphenCase("some_thing")                       # some-thing 
w.snakeCase("someThing")                         # some_thing 
w.humanCase("fromA_to-Z")                        # from A to Z 

This helps to split "cased" words into their constituent parts...

  caseSplitter = /[-_\s]+|(!?[A-Z][a-z]*)/g

  camelCase = (str) ->
      for s, i in str.split(caseSplitter) when s
        if i is 0 then s else capitalize(s)

    str.replace(/[\s_-]+(\w)/g, (a, w) -> w.toUpperCase())

  hyphenCase = (str, leading) ->
    (if leading then "-" else "") +
    cleanJoin(str.split(caseSplitter), "-").toLowerCase()

  snakeCase = (str) ->
    cleanJoin(str.split(caseSplitter), "_").toLowerCase()

  humanCase = (str) ->
    cleanJoin(str.split(caseSplitter), " ")


Enclose a string inside an HTML tag.


w.wrapInTag("Hello world!")                      # <span>Hello world!</span> 
w.wrapInTag("Hello world!""em")                # <em>Hello world!</em> 
w.wrapInTag(                                     # <a href="/url" class="b" disabled="disabled">Link</a> 
    href: "/url",
    class: ["b"],
    disabled: true
  wrapInTag = (str, tag = "span", attributes = {}) ->
    attrs =
      for key, val of attributes when val
        val = key if typeof val is "boolean"
        val = val.join(" ") if typeof val.join is "function"
        " #{key}=\"#{val}\""

    enclose "<#{tag}#{attrs.join("")}>", str, "</#{tag}>"


Group strings into a grammatically correct list with an arbitrary limit. The final example shows all the possible options available.


w.prettyList(["Ben""Bob"])                     # Ben and Bob 
w.prettyList(["Ben""Bob""Bill"])             # Ben, Bob and Bill 
w.prettyList(["Ben""Bob""Bill""Max"]2)   # Ben, Bob and 2 more 
w.prettyList(["Ben""Bob"]1{more: "other"}) # Ben and 1 other 
w.prettyList([                                   # Document 1 & two other files 
  {file: "Document 1"},
  {file: "Document 2"},
  {file: "Document 3"}
  amp: "&"
  written: true,
  more: "other file",
  quantify: true,
  key: "file"
  prettyList = (arr, max, opts = {}) ->
    more   = opts.more or "more"
    amp    = opts.amp or "and"

    if opts.key
      arr = (obj[opts.key] for obj in arr when typeof obj is "object")

    if opts.wrap
      arr = (wrapInTag(s, opts.wrap) for s in arr)

    if max < (len = arr.length)
      diff = len - max
      more = quantify(more, diff, numberless: true) if opts.quantify
      diff = writtenNumber(diff, opts.lang) if opts.written
      arr  = arr.slice(0, max)
      arr  = arr.concat("#{diff} #{more}")

      .slice(0, -1)
      .join(", ")
      .concat((if arr.length is 1 then "" else " #{amp} "), arr.slice(-1))


Add soft hyphens every n characters so that the CSS attribute hyphens: manual will allow for nice breaks in long strings of text. This is especially useful on mobile devices, where long strings can break the layout.


w.hyphenate("antidisestablishmentarianism")      # antidisest%C2%ADablishm... 
  hyphenate = (str = "", n = 10, softHyphen = "\u00AD") ->
    str.replace /(^|[^>])+(?=$|\<)/g, (sub) ->
      sub.replace new RegExp("(\\w{#{n - 1}})(\\w)", "g"), (w, a, b) ->
        a + softHyphen + b


Add an "s" to a string when an amount is non-singular, disregarding the order of the arguments passsed. If an array or collection is passed, it’s length will be used as the numerical input.


w.quantify("monkey"1)                          # 1 monkey 
w.quantify(1"monkey")                          # 1 monkey 
w.quantify("monkey"9{written: true})         # nine monkeys 
w.quantify("person"9{plural: "people"})      # 9 people 
w.quantify([123]"number")                  # 3 numbers 
  quantify = (str, n, {numberless, written, lang, plural} = {}) ->
    [n, str] = [str, n] unless typeof str is "string"

    n = n.length ? n
    s = if n is 1 then str else (plural or "#{str}s")
    n = writtenNumber(n, lang) if written
    n = if numberless then "" else "#{n} "

    n + s

Written Numbers

Convert numbers between one and twelve into their written counter-parts.


w.writtenNumber(1)                               # one 
w.writtenNumber(2"DE")                         # zwei 
  writtenNumber = (n, lang = "EN", gender = "m") ->
    if num = dictionary[lang]?.cardinals.written[n - 1]
      num[gender] and num[gender] or num


Wrap a string in single or double quotes or guillemets (angle quotes).


w.quote("pastry chef""s")                      # ‘pastry chef’ 
w.quote("cats cradle")                           # “cats cradle” 
w.quote("tres chic""a")                        # «tres chic» 
w.quote("Gol""!")                              # ¡Gol! 
w.quote("Cómo estás""?")                       # ¿Cómo estás? 
  quote = (str, type) ->
    [a, z] =
      switch type
        when "s", "single"
          ["‘", "’"]
        when "a", "angle", "g", "guillemets"
          ["«", "»"]
        when "!"
          ["¡", "!"]
        when "?"
          ["¿", "?"]
          ["“", "”"]

    enclose a, str, z


Convert a number from it's cardinal to ordinal equivalent.


w.ordinal(1)                                     # 1st 
w.ordinal(2{written: true})                    # second 
w.ordinal(3{wrap: true})                       # 3<sup>rd</sup> 
w.ordinal(4{wrap: "em"})                       # 4<em>th</em> 
  ordinal = (n, opts = {}, gender = "m") ->
    { suffixes, rule, written } = dictionary[opts.lang or "EN"].ordinals
    base   = n.toString().match(rule)[0]

    if opts.written? and w = written[+n - 1]
      w[gender] or w

      suffix = suffixes[base] or suffixes.n
      suffix = suffix[gender] or suffix

      if opts.wrap and not w
        opts.wrap = "sup" if typeof opts.wrap is "boolean"
        suffix = wrapInTag(suffix, opts.wrap)

      n + suffix


Format a number in various ways and parse one from a string.


w.prettyNumber(1000)                             # 1,000 
w.prettyNumber(10.52)                          # 10.50 
w.prettyNumber(9999" "2",")                # 9 999,00 
w.prettyPrice(4)                                 # $4.00 
w.prettyPrice(1200"£")                         # £1,200.00 
w.prettyPrice(                                   # €4<sup>00</sup> 
      currency: "",
      wrap: "sup"
w.prettyPercent(0.5)                             # 50% 
w.prettyPercent(14)                            # 25% 
w.prettyPercent(132)                         # 33.33% 
w.parseNumber(1000)                              # 1000 
w.parseNumber("1,000.00")                        # 1000 
w.parseNumber("99%")                             # 0.99 
w.parseNumber("some 44,000 participants")        # 44000 
  prettyNumber = (n, delimiter = ",", decimals = 0, dot = ".") ->
    decimals = delimiter if typeof delimiter is "number"
    n = parseNumber(n)
    n = n.toFixed(decimals) if decimals > 0
    n = n.toString().replace(".", dot) if dot
    [ int, frac ] = n.toString().split(dot)
    [ int.replace(/\B(?=(\d{3})+(?!\d))/g, delimiter), frac ]
      .filter((e) -> e?)

  prettyPrice = (n, currency) ->
    if typeof currency is "object"
      {currency, wrap, decimals, delimiter, dot} = currency

    currency ?= "$"
    decimals ?= 2
    dot ?= "."

    n = prettyNumber(n, delimiter, decimals, dot)

    if wrap
      [integer, fraction] = n.split(dot)
      n = integer + wrapInTag(fraction, wrap)

    currency + n

  prettyPercent = (numerator = 0, denominator = 1, decimals = 0) ->
    percent = (numerator / denominator * 100) or 0

  parseNumber = (n) ->
    if typeof n is "string"
      n = n.replace(/[^\d\.]+/g, "") / if n.slice(-1) is "%" then 100 else 1

    if n <= Infinity then n else -1


Provide quick access to different typographic glyphs without the need commit them to memory or look at a reference table.


w.glyphs()                                       # Create map of ASCII glyphs 
w.glyph("!")                                     # &#33; 
  fromTo = (x, y) ->
    Array.apply(0, length: y - x + 1).map((e, i) -> i + x)

  glyphs = (glyphs = {}) ->
    for code in fromTo(161, 255)
      .concat(fromTo(338, 402))
      .concat(fromTo(8211, 8230))
      .concat([8240, 8364, 8482])
        glyphs[code] = String.fromCharCode(code)


  glyph = (c) ->
    enclose "&#", c.charCodeAt(0), ";"

Language Support

Set cardinal and ordinal numbers and non-caps words for different languages as appropriate. Please note that only partial support for French, German, Italian, Spanish and Swedish is currently implemented. If using in the browser, ensure that the document's charset is set to UTF-8. Pull requests which extend language support are encouraged.

  setLanguage = (object, lang) ->
    dictionary[lang] = object


Pack up the written object (with some aliases...)

  return written =
    camelCase: camelCase
    capitalize: capitalize
    capitalizeAll: capitalizeAll
    cleanJoin: cleanJoin
    collapse: collapse
    count: quantify
    dasherize: hyphenCase
    dashify: hyphenCase
    enclose: enclose
    glyph: glyph
    glyphs: glyphs
    humanCase: humanCase
    hyphenate: hyphenate
    hyphenCase: hyphenCase
    numerate: quantify
    ordinal: ordinal
    parseNumber: parseNumber
    prettyList: prettyList
    prettyNumber: prettyNumber
    prettyPrice: prettyPrice
    prettyPercent: prettyPercent
    quantify: quantify
    quote: quote
    setLanguage: setLanguage
    slugify: snakeCase
    snakeCase: snakeCase
    titleCase: capitalizeAll
    underscore: snakeCase
    wrapInTag: wrapInTag
    writtenNumber: writtenNumber