1 |
#!/usr/bin/perl -w |
2 |
|
3 |
#-------------------------------------------------------------------- |
4 |
|
5 |
my @affix_regexp; |
6 |
my @affix_add; |
7 |
my @affix_sub; |
8 |
|
9 |
sub load_affix { |
10 |
my ($filename) = @_; |
11 |
|
12 |
my $suffixes=0; |
13 |
|
14 |
my ($regexp,$add,$sub); |
15 |
|
16 |
open (A,$filename) || die "$filename: $!"; |
17 |
while(<A>) { |
18 |
chomp; |
19 |
next if (/^#|^[\s\t\n\r]*$/); |
20 |
|
21 |
if (/^suffixes/i) { |
22 |
$suffixes++; |
23 |
next; |
24 |
} |
25 |
|
26 |
next if (! $suffixes); |
27 |
|
28 |
if (/^flag[\s\t]+\*{0,1}(.):/i) { |
29 |
undef $regexp; |
30 |
undef $add; |
31 |
undef $sub; |
32 |
next; |
33 |
} |
34 |
|
35 |
if (/^[\s\t]*([^>#]+)>[\s\t]+-([^\,\s\t]+),([^\s\t]+)/) { |
36 |
$regexp = $1; |
37 |
$add = $2; |
38 |
$sub = $3 if ($3 ne "-"); |
39 |
} elsif (/^[\s\t]*([^>#]+)>[\s\t]+([^\s\t\#]+)/) { |
40 |
$regexp = $1; |
41 |
$sub = $2; |
42 |
} |
43 |
|
44 |
sub nuke_s { |
45 |
my $tmp = $_[0]; |
46 |
return if (!$tmp); |
47 |
$tmp=~s/^ *//g; |
48 |
$tmp=~s/ *$//g; |
49 |
$tmp=~s/ *//g; |
50 |
return $tmp; |
51 |
} |
52 |
|
53 |
push @affix_regexp,nuke_s($regexp); |
54 |
push @affix_add,nuke_s($add); |
55 |
push @affix_sub,nuke_s($sub); |
56 |
} |
57 |
} |
58 |
|
59 |
sub normalize_word { |
60 |
my @out; |
61 |
foreach my $word (@_) { |
62 |
push @out,$word; # save original word |
63 |
next if (length($word) < 3); # cludge: preskoci kratke |
64 |
for(my $i=0; $i<=$#affix_regexp; $i++) { |
65 |
my $regexp = $affix_regexp[$i]; |
66 |
my $add = $affix_add[$i]; |
67 |
my $sub = $affix_sub[$i]; |
68 |
next if length($word) < length($sub); |
69 |
my $tmp_word = $word; |
70 |
if ($sub) { |
71 |
next if ($word !~ m/$sub$/i); |
72 |
$tmp_word =~ s/$sub$/$add/i if ($add); |
73 |
} else { |
74 |
$tmp_word = $word.$add; |
75 |
} |
76 |
if ($tmp_word =~ m/$regexp/ix) { |
77 |
# print "$word -> $tmp_word\t-$sub, +$add, regexp: $regexp\n"; |
78 |
push @out,lc($tmp_word); |
79 |
} |
80 |
} |
81 |
} |
82 |
return @out; |
83 |
} |
84 |
|
85 |
|
86 |
|
87 |
#-------------------------------------------------------------------- |
88 |
|
89 |
load_affix("/tmp/dpavlin/docs/nn/search/croatian.aff"); |
90 |
|
91 |
print "elements: ",$#affix_regexp,"\n"; |
92 |
|
93 |
print join(", ",normalize_word( ("zakoni","oprostu") )),"\n"; |
94 |
|
95 |
print join(", ",normalize_word( "preirez" )),"\n"; |
96 |
print join(", ",normalize_word( "preireza" )),"\n"; |
97 |
|