3 use Unicode::Normalize qw( NFKD );
4 use Text::Unidecode qw( unidecode );
7 my %g; # present group classes
15 !defined $v || $_ eq $v ? 1 # identical
16 : $v =~ /\A\p{Mn}+\z/ ? 9 # combining accent
17 : $v =~ /\A[\p{Sk}\p{Lm}]+\z/ ? 8 # modifier symbol
18 : $v =~ /\A[\pM\pP]+\z/ ? 7 # mark
19 : NFKD($v) =~ /\Q$_/ ? 2 # decomposed equivalent
20 : unidecode($v) =~ /\Q$_\E+/i ? 4 # transliterated
21 : $v =~ /^\p{Latin}/ ? 5 # latin script
31 g1 => ['unaltered', "same results as without modifier"],
32 g2 => ['accented', "decomposes to the original letter with a combining accent"],
33 g4 => ['similar', "transliterates (mostly) into the unmodified letter"],
34 g5 => ['latin', "a different (accented) latin letter"],
35 g6 => ['symbol', "other character not directly deducible from key"],
36 g7 => ['punctuation', "(punctuation) mark"],
38 g8 => ['mark', "modifier letter or mark (spacing diacritic)"],
39 g9 => ['combining', "diacritical mark to be combined with a following character"],