# coding: utf-8

[0] Fix | Delete

[1] Fix | Delete

# Copyright Ayumu Nojima (野島 歩) and Martin J. Dürst (duerst@it.aoyama.ac.jp)

[2] Fix | Delete

[3] Fix | Delete

# additions to class String for Unicode normalization

[4] Fix | Delete

class String

[5] Fix | Delete

# === Unicode Normalization

[6] Fix | Delete

#

[7] Fix | Delete

# :call-seq:

[8] Fix | Delete

# str.unicode_normalize(form=:nfc)

[9] Fix | Delete

#

[10] Fix | Delete

# Returns a normalized form of +str+, using Unicode normalizations

[11] Fix | Delete

# NFC, NFD, NFKC, or NFKD. The normalization form used is determined

[12] Fix | Delete

# by +form+, which is any of the four values :nfc, :nfd, :nfkc, or :nfkd.

[13] Fix | Delete

# The default is :nfc.

[14] Fix | Delete

#

[15] Fix | Delete

# If the string is not in a Unicode Encoding, then an Exception is raised.

[16] Fix | Delete

# In this context, 'Unicode Encoding' means any of UTF-8, UTF-16BE/LE,

[17] Fix | Delete

# and UTF-32BE/LE, as well as GB18030, UCS_2BE, and UCS_4BE. Anything

[18] Fix | Delete

# else than UTF-8 is implemented by converting to UTF-8,

[19] Fix | Delete

# which makes it slower than UTF-8.

[20] Fix | Delete

#

[21] Fix | Delete

# _Examples_

[22] Fix | Delete

#

[23] Fix | Delete

# "a\u0300".unicode_normalize #=> 'à' (same as "\u00E0")

[24] Fix | Delete

# "a\u0300".unicode_normalize(:nfc) #=> 'à' (same as "\u00E0")

[25] Fix | Delete

# "\u00E0".unicode_normalize(:nfd) #=> 'à' (same as "a\u0300")

[26] Fix | Delete

# "\xE0".force_encoding('ISO-8859-1').unicode_normalize(:nfd)

[27] Fix | Delete

# #=> Encoding::CompatibilityError raised

[28] Fix | Delete

#

[29] Fix | Delete

def unicode_normalize(form = :nfc)

[30] Fix | Delete

require 'unicode_normalize/normalize.rb' unless defined? UnicodeNormalize

[31] Fix | Delete

## The following line can be uncommented to avoid repeated checking for

[32] Fix | Delete

## UnicodeNormalize. However, tests didn't show any noticeable speedup

[33] Fix | Delete

## when doing this. This comment also applies to the commented out lines

[34] Fix | Delete

## in String#unicode_normalize! and String#unicode_normalized?.

[35] Fix | Delete

# String.send(:define_method, :unicode_normalize, ->(form = :nfc) { UnicodeNormalize.normalize(self, form) } )

[36] Fix | Delete

UnicodeNormalize.normalize(self, form)

[37] Fix | Delete

end

[38] Fix | Delete

[39] Fix | Delete

# :call-seq:

[40] Fix | Delete

# str.unicode_normalize!(form=:nfc)

[41] Fix | Delete

#

[42] Fix | Delete

# Destructive version of String#unicode_normalize, doing Unicode

[43] Fix | Delete

# normalization in place.

[44] Fix | Delete

#

[45] Fix | Delete

def unicode_normalize!(form = :nfc)

[46] Fix | Delete

require 'unicode_normalize/normalize.rb' unless defined? UnicodeNormalize

[47] Fix | Delete

# String.send(:define_method, :unicode_normalize!, ->(form = :nfc) { replace(unicode_normalize(form)) } )

[48] Fix | Delete

replace(unicode_normalize(form))

[49] Fix | Delete

end

[50] Fix | Delete

[51] Fix | Delete

# :call-seq:

[52] Fix | Delete

# str.unicode_normalized?(form=:nfc)

[53] Fix | Delete

#

[54] Fix | Delete

# Checks whether +str+ is in Unicode normalization form +form+,

[55] Fix | Delete

# which is any of the four values :nfc, :nfd, :nfkc, or :nfkd.

[56] Fix | Delete

# The default is :nfc.

[57] Fix | Delete

#

[58] Fix | Delete

# If the string is not in a Unicode Encoding, then an Exception is raised.

[59] Fix | Delete

# For details, see String#unicode_normalize.

[60] Fix | Delete

#

[61] Fix | Delete

# _Examples_

[62] Fix | Delete

#

[63] Fix | Delete

# "a\u0300".unicode_normalized? #=> false

[64] Fix | Delete

# "a\u0300".unicode_normalized?(:nfd) #=> true

[65] Fix | Delete

# "\u00E0".unicode_normalized? #=> true

[66] Fix | Delete

# "\u00E0".unicode_normalized?(:nfd) #=> false

[67] Fix | Delete

# "\xE0".force_encoding('ISO-8859-1').unicode_normalized?

[68] Fix | Delete

# #=> Encoding::CompatibilityError raised

[69] Fix | Delete

#

[70] Fix | Delete

def unicode_normalized?(form = :nfc)

[71] Fix | Delete

require 'unicode_normalize/normalize.rb' unless defined? UnicodeNormalize

[72] Fix | Delete

# String.send(:define_method, :unicode_normalized?, ->(form = :nfc) { UnicodeNormalize.normalized?(self, form) } )

[73] Fix | Delete

UnicodeNormalize.normalized?(self, form)

[74] Fix | Delete

end

[75] Fix | Delete

end

[76] Fix | Delete

[77] Fix | Delete

[78] Fix | Delete