unicode-display_width 0.1.0

data/lib/unicode/display_size.rb

@@ -0,0 +1 @@
+require File.dirname(__FILE__) + "/display_width"

data/lib/unicode/display_width.rb

@@ -0,0 +1,82 @@
+# encoding: utf-8
+module Unicode; end
+
+module Unicode::DisplayWidth
+  VERSION = '0.1.0'
+
+  DATA_DIR = File.join(File.dirname(__FILE__), '../../data/')
+  FIELDS = [:codepoint, :width, :name]
+  INDEX_FILE = DATA_DIR + 'EastAsianWidth.index'
+  DATA_FILE  = DATA_DIR + 'EastAsianWidth.txt'
+
+  def self.data
+    @@data ||= File.open DATA_FILE
+  end
+
+  def self.offsets
+    @@offsets ||= Marshal.load File.respond_to?(:binread) ? File.binread(INDEX_FILE) : File.read(INDEX_FILE)
+  end
+
+  class Codepoint < Struct.new(*FIELDS)
+    def initialize(*args)
+      super
+      self.codepoint = self.codepoint.to_i(16) if self.codepoint && self.codepoint !~ /\.\./
+                                                  # TODO cleaner    # FIXME ranges
+    end
+
+    def self.from_line(line)
+      line =~ /(.*);(.*) # (.*)$/
+        raise 'BUG' unless line
+      new $1,$2,$3
+    end
+  end
+
+  def self.line(n)
+    data.rewind
+    offset = offsets[n] or raise ArgumentError
+    data.seek offset
+    data.readline.chomp
+  end
+
+  def self.codepoint(n)
+    Codepoint.from_line line(n)
+  end
+
+  def self.valid_index?
+    !!offsets rescue false
+  end
+
+  def self.build_index
+    data.rewind
+    offsets = {}
+    dir = File.dirname INDEX_FILE
+    Dir.mkdir(dir) unless Dir.exists?(dir)
+    data.lines.map do |line|
+      offsets[Codepoint.from_line(line).codepoint] = data.pos - line.size
+    end
+    File.open(INDEX_FILE, 'wb') { |f| Marshal.dump(offsets, f) }
+  end
+end
+
+class String
+  def display_width(ambiguous=1)
+    #codepoints.inject(0){ |a,c|
+    unpack('U*').inject(0){ |a,c|
+      width = case Unicode::DisplayWidth.codepoint(c).width
+              when *%w[F W]
+                2
+              when *%w[N Na H]
+                1
+              when *%w[A] # TODO
+                ambiguous
+              else
+                1
+              end
+      a + width
+    }
+  end
+  alias display_size   display_width
+  alias display_length display_width
+end
+
+# J-_-L

metadata

@@ -0,0 +1,77 @@
+--- !ruby/object:Gem::Specification 
+name: unicode-display_width
+version: !ruby/object:Gem::Version 
+  hash: 27
+  prerelease: false
+  segments: 
+  - 0
+  - 1
+  - 0
+  version: 0.1.0
+platform: ruby
+authors: 
+- Jan Lelis
+autorequire: 
+bindir: bin
+cert_chain: []
+
+date: 2011-01-03 00:00:00 +01:00
+default_executable: 
+dependencies: []
+
+description: This gem adds String#display_size to get the display size of a string using EastAsianWidth.txt.
+email: mail@janlelis.de
+executables: []
+
+extensions: []
+
+extra_rdoc_files: 
+- README.rdoc
+- LICENSE.txt
+files: 
+- lib/unicode/display_size.rb
+- lib/unicode/display_width.rb
+- LICENSE.txt
+- README.rdoc
+- data/EastAsianWidth.txt
+- data/EastAsianWidth.index
+- Rakefile
+- .gemspec
+has_rdoc: true
+homepage: http://github.com/janlelis/unicode-display_size
+licenses: 
+- MIT
+post_install_message: 
+rdoc_options: []
+
+require_paths: 
+- lib
+required_ruby_version: !ruby/object:Gem::Requirement 
+  none: false
+  requirements: 
+  - - ">="
+    - !ruby/object:Gem::Version 
+      hash: 3
+      segments: 
+      - 0
+      version: "0"
+required_rubygems_version: !ruby/object:Gem::Requirement 
+  none: false
+  requirements: 
+  - - ">="
+    - !ruby/object:Gem::Version 
+      hash: 23
+      segments: 
+      - 1
+      - 3
+      - 6
+      version: 1.3.6
+requirements: []
+
+rubyforge_project: 
+rubygems_version: 1.3.7
+signing_key: 
+specification_version: 3
+summary: Support for east_asian_width String sizes.
+test_files: []
+