normalize on char level

1 files changed, 35 insertions, 0 deletions

diff --git a/normchr b/normchr
new file mode 100755
index 0000000..f8e5798
--- /dev/null
+++ b/normchr
@@ -0,0 +1,35 @@
+#!/usr/bin/env ruby
+
+# http://www.utf8-chartable.de/unicode-utf8-table.pl?start=8192&number=128&utf8=string-literal
+# https://www.cs.tut.fi/~jkorpela/chars/spaces.html
+
+require 'htmlentities'
+
+STDIN.set_encoding 'utf-8'
+STDOUT.set_encoding 'utf-8'
+
+coder = HTMLEntities.new
+
+while line = STDIN.gets
+  line.sub! "\xef\xbb\xbf", ""          # BOM
+  line.strip!                           # superfluous
+  line.lstrip!                          #   whitespace
+  line.gsub! /[[:cntrl:]]+/, " "        # control characters
+  line.gsub! /\u{00a0}/, " "            # misc whitespace
+  line.gsub! /\u{1680}/, " "            # ^
+  line.gsub! /\u{180e}/, " "            # ^
+  line.gsub! /\u{3000}/, " "            # ^
+  line.gsub! /\u{feff}/, " "            # ^
+  line = line.scan(/[[:print:]]/).join  # only printable characters
+  line.gsub! /[\u{e000}-\u{f8ff}]/, " " # UTF-8 PUA
+  line.gsub! /[\u{f0000}-\u{ffffd}]/, " "
+  line.gsub! /[\u{100000}-\u{10fffd}]/, " "
+  line.gsub! "\r", " "                  # carriage return  
+  line.gsub! /[\u{2000}-\u{200f}]/, " " #                   EN QUAD -- RIGHT-TO-LEFT MARK
+  line.gsub! /[\u{2028}-\u{202f}]/, " " #            LINE SEPARATOR -- NARROW NO-BREAK SPACE
+  line.gsub! /[\u{205f}-\u{206f}]/, " " # MEDIUM MATHEMATICAL SPACE -- NOMINAL DIGIT SHAPES
+  line.gsub! /\s*\xc2\xad\s*/, ""       # remove hyphens
+  line.gsub! /[[:space:]]+/, " "        # collapse space
+  puts coder.decode(line)
+end
+