/[nn]/search/nn-swish.cgi
This is repository of my old source code which isn't updated any more. Go to git.rot13.org for current projects!
ViewVC logotype

Annotation of /search/nn-swish.cgi

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1.14 - (hide annotations)
Sun Sep 28 02:19:55 2003 UTC (20 years, 7 months ago) by dpavlin
Branch: MAIN
Changes since 1.13: +81 -36 lines
full text search, move Lingua::Spelling::Alternative generation to cgi script
(to decrease size of index file), search by years

1 dpavlin 1.1 #!/usr/bin/perl -w
2    
3     use strict;
4     use CGI qw/:standard -no_xhtml/;
5     use CGI::Carp qw(fatalsToBrowser);
6     use SWISH;
7 dpavlin 1.5 use Unicode::String qw(utf8 utf16);
8 dpavlin 1.14 use Lingua::Spelling::Alternative;
9 dpavlin 1.3 require Unicode::Map8;
10 dpavlin 1.13 use GDBM_File;
11 dpavlin 1.1
12 dpavlin 1.14 my $dir='/home/dpavlin/nn';
13 dpavlin 1.12 my $prog='/usr/bin/swish-e';
14 dpavlin 1.14 my $url='http://www.nn.hr/clanci/sluzbeno/';
15 dpavlin 1.1
16     my $hits=0;
17 dpavlin 1.7 my $max_hits=100;
18    
19     my %labels = (100=>' 100', 200=>' 200', 500=>' 500', 0=>'neogranièeno');
20 dpavlin 1.14 my %index = ('title'=>'naslovu', 'text'=>'tekstu');
21 dpavlin 1.1
22 dpavlin 1.13 my %brs_labels;
23    
24 dpavlin 1.14 my %brzakona;
25     tie %brzakona, 'GDBM_File', "$dir/swish/brzakona.gdbm", &GDBM_READER, 0640 || die "tie: $!";
26     foreach (sort keys %brzakona) {
27     $brs_labels{$_} = sprintf("%-8s (%d zakona)",$_,$brzakona{$_});
28 dpavlin 1.13 }
29 dpavlin 1.14 untie %brzakona;
30    
31     my %file2title;
32     tie %file2title, 'GDBM_File', "$dir/swish/file2title.gdbm", &GDBM_READER, 0640 || die "tie: $!";
33    
34     my $hr = new Lingua::Spelling::Alternative( DEBUG => 0 );
35     $hr->load_findaffix("$dir/prvih_50.txt");
36 dpavlin 1.13
37 dpavlin 1.1 print header(-charset=>'iso-8859-2'),start_html(-title=>'NN pretrazivanje',-lang=>'hr'),start_form;
38 dpavlin 1.7 print "Potra¾i ",popup_menu(-name=>'max_hits',-values=>[ sort keys %labels ],-labels=>\%labels,-default=>$max_hits)," zakona sa rijeèima: ",textfield('search');
39 dpavlin 1.14 print " u ",popup_menu(-name=>'index',-values=>[ sort keys %index ],-labels=>\%index,-default=>'title')," zakona ";
40 dpavlin 1.1 print submit(-value=>'prika¾i');
41 dpavlin 1.14 print br,checkbox(-name=>'br_limit', -checked=>0, -label=>"ogranièi pretra¾ivanje samo na godinu "),popup_menu(-name=>'br',-values=>[sort keys %brs_labels],-labels=>\%brs_labels);
42 dpavlin 1.1 print end_form,hr;
43    
44     if (param('search')) {
45    
46 dpavlin 1.6 my $s;
47     # re-write query from +/- to and/and not
48     foreach (split(/\s+/,param('search'))) {
49     if (m/^([+-])(\S+)/) {
50     $s.= ($s) ? "and " : "";
51     $s.="not " if ($1 eq "-");
52 dpavlin 1.14 $s.="(".join("* or ",$hr->alternatives($2)).") ";
53 dpavlin 1.6 } else {
54 dpavlin 1.14 $s .= "(".join("* or ",$hr->alternatives($_)).") ";
55 dpavlin 1.6 }
56     }
57 dpavlin 1.1 $s=~tr/šðžèæŠÐŽÈÆ/¹ð¾èæ©Ð®ÈÆ/; # 1250 -> iso8859-2
58     $s=~tr/¹©ðÐèÈæƾ®/sSdDcCcCzZ/;
59    
60 dpavlin 1.3 my $l2_map = Unicode::Map8->new("ISO-8859-2") || die;
61 dpavlin 1.5 my $us = Unicode::String->new();
62 dpavlin 1.3
63 dpavlin 1.14 my $sw_q;
64     my $sh;
65    
66     if (param('index') eq 'title') {
67    
68     $sh = SWISH->connect('Fork',
69     prog => $prog,
70     indexes => "$dir/swish/nn.index",
71     properties => [qw/god br nr/],
72     results => sub {
73     my ($sh,$hit) = @_;
74    
75     $us->utf8($hit->swishtitle);
76    
77     print "<tt><a href=\"$url",$hit->swishdocpath,"\">NN",$hit->god,"/",$hit->br,"</a> ",$hit->nr," </tt>",$l2_map->to8($us->utf16)," [",$hit->swishrank,"]<br>\n";
78    
79     # print $_[1]->as_string,"<br>\n";
80     # my @fields = $hit->field_names;
81     # print "Field '$_' = '", $hit->$_, "'<br>\n" for sort @fields;
82     },
83     maxhits => param('max_hits') || $max_hits,
84     );
85    
86     die $SWISH::errstr unless $sh;
87    
88     $sw_q = "naslov_czs=($s)";
89     if (param('br_limit')) {
90     $sw_q .= " and god=".int(param('br')) if (param('br'));
91     }
92    
93     } else {
94     # search in full text
95    
96     $sh = SWISH->connect('Fork',
97     prog => $prog,
98     indexes => "$dir/swish/sluzbeno.index",
99     results => sub {
100     my ($sh,$hit) = @_;
101    
102     my $path = $hit->swishdocpath;
103     if ($file2title{$path}) {
104     my ($god,$br,$nr,undef,$naslov) = split(/ /,$file2title{$path},5);
105     print "<tt><a href=\"$url",$path,"\">NN$god/$br</a> $nr</tt> $naslov [",$hit->swishrank,"]<br>\n";
106    
107     } else {
108     print "<!-- error! can't find $path -->\n";
109     }
110    
111     },
112     maxhits => param('max_hits') || $max_hits,
113     );
114    
115     die $SWISH::errstr unless $sh;
116    
117     $sw_q = $s;
118     if (param('br_limit')) {
119     $sw_q .= " and swishdocpath=".int(param('br')) if (param('br'));
120     }
121    
122 dpavlin 1.13 }
123    
124     print "<!-- swish query: $sw_q -->";
125    
126     $hits = $sh->query($sw_q);
127 dpavlin 1.1
128     if ($hits > 0) {
129 dpavlin 1.13 print p,hr,"Prikazujem $hits zakona";
130     print " iz broja ",param('br') if (param('br_limit'));
131     print " (maks. ",param('max_hits') || $max_hits,")... <small>($s)</small>";
132 dpavlin 1.1 } else {
133 dpavlin 1.6 print p,"Nije naðen niti jedan zakon... <small>($s, ",$sh->errstr,")</small>";
134 dpavlin 1.1 }
135     } else {
136 dpavlin 1.4 print p('Kod pretra¾ivanja pretra¾ivaè pronalazi sve zakone u kojima se pojavljuju <b>sve upisanje rijeèi</b>.',br,'Ako ispred rijeèi upi¹ete minus (-) neæe se prikazivati zakoni koji imaju takvu rijeè. Npr. <tt>+kava +zakon -dopunama</tt>');
137 dpavlin 1.14 print p("Mo¾ete proèitati i <a href=\"http://www.rot13.org/~dpavlin/nn.html\">èlanak</a> o tome kako je ovaj pretra¾ivaè napravljen i za¹to."),p("Ovo je <i>nova i br¾a verzija pretra¾ivaèa</i> o kojoj <a href=\"http://www.rot13.org/~dpavlin/nn.html#buducnost\">vi¹e mo¾ete saznati</a> iz èlanka."),p("Novosti od rujna 2003.: pretra¾ivanje po godinama i po punom tekstu zakona!");
138 dpavlin 1.1 }
139 dpavlin 1.14
140     untie %file2title;

  ViewVC Help
Powered by ViewVC 1.1.26