/[webpac]/trunk2/all2all.pl
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 /trunk2/all2all.pl

Parent Directory Parent Directory | Revision Log Revision Log


Revision 423 - (hide annotations)
Sat Sep 11 20:21:28 2004 UTC (19 years, 6 months ago) by dpavlin
File MIME type: text/plain
File size: 5595 byte(s)
added show/hide to tree view

1 dpavlin 1 #!/usr/bin/perl -w
2    
3 dpavlin 348 =head1 NAME
4    
5 dpavlin 354 all2all.pl - basic script for all WebPAC needs
6 dpavlin 348
7     =cut
8    
9 dpavlin 1 use strict;
10     use Data::Dumper;
11 dpavlin 352 use Carp;
12 dpavlin 1
13 dpavlin 352 use lib './lib';
14 dpavlin 354 use WebPAC;
15 dpavlin 390 use WebPAC::jsFind;
16 dpavlin 410 use WebPAC::Index;
17 dpavlin 348
18 dpavlin 354 my $webpac = new WebPAC(
19 dpavlin 352 code_page => 'ISO-8859-2',
20 dpavlin 423 limit_mfn => 500,
21 dpavlin 372 # debug => 1,
22 dpavlin 423 # low_mem => 1,
23 dpavlin 352 ) || die;
24    
25 dpavlin 421 my $log = $webpac->_get_logger() || die "can't get logger";
26 dpavlin 373
27 dpavlin 421 $log->debug("creating WebPAC::jsFind object");
28    
29 dpavlin 390 my $index = new WebPAC::jsFind(
30     index_path => './out/index',
31 dpavlin 413 keys => 10,
32 dpavlin 390 ) || die;
33    
34 dpavlin 410 my $thes;
35    
36 dpavlin 10 $|=1;
37 dpavlin 9
38 dpavlin 352 my $maxmfn = $webpac->open_isis(
39 dpavlin 357 filename => shift @ARGV || '/data/hidra/THS/THS',
40 dpavlin 352 lookup => [
41 dpavlin 348 { 'key' => 'd:v900', 'val' => 'v250^a' },
42 dpavlin 352 # { 'eval' => '"v901^a" eq "Podruèje"', 'key' => 'pa:v561^4:v562^4:v461^1', 'val' => 'v900' },
43     # { 'eval '=> '"v901^a" eq "Mikrotezaurus"', 'key' => 'a:v561^4:v562^4:v461^1', 'val' => 'v900' },
44     # { 'eval' => '"v901^a" eq "Deskriptor"', 'key' => 'a:v561^4:v562^4:v461^1', 'val' => 'v900' },
45     { 'key' => 'a:v561^4:v562^4:v461^1', 'val' => 'v900' },
46 dpavlin 374 { 'key' => '900_mfn:v900', 'val' => 'v000' },
47 dpavlin 419 # tree structure
48     { 'eval' => 'length("v251") == 2', 'key' => 'root:v251', 'val' => 'v900' },
49     { 'eval' => '"v251"', 'key' => 'code:v900', 'val' => 'v561^4:v251' },
50 dpavlin 352 ],
51 dpavlin 54 );
52 dpavlin 3
53 dpavlin 421 $log->debug("isis file ",$webpac->{'isis_filename'}," opened");
54    
55 dpavlin 372 $log->info("rows: $maxmfn");
56 dpavlin 170
57 dpavlin 363 $webpac->open_import_xml(type => 'isis_hidra_ths');
58    
59 dpavlin 362 while (my $rec = $webpac->fetch_rec) {
60 dpavlin 358
61 dpavlin 370 my @ds = $webpac->data_structure($rec);
62 dpavlin 366
63 dpavlin 411 if (0 && $log->is_debug) {
64 dpavlin 372 $log->debug("rec = ",Dumper($rec));
65     $log->debug("ds = ",Dumper(\@ds));
66     }
67    
68 dpavlin 374 next if (! @ds);
69 dpavlin 366
70 dpavlin 421 my $filename = $webpac->{'current_filename'} || $log->logdie("no current_filename in webpac object");
71 dpavlin 374
72     if ($filename) {
73 dpavlin 411 $webpac->output_file(
74     file => $filename,
75 dpavlin 374 template => 'html.tt',
76     data => \@ds,
77 dpavlin 398 headline => $webpac->{'headline'},
78 dpavlin 374 );
79     } else {
80     print $webpac->output(
81     template => 'text.tt',
82     data => \@ds,
83 dpavlin 398 headline => $webpac->{'headline'},
84 dpavlin 374 );
85     }
86    
87 dpavlin 398 my $headline = $webpac->{'headline'};
88 dpavlin 390
89 dpavlin 392 my $f = $filename;
90     $f =~ s!out/!!;
91    
92 dpavlin 390 # save into index
93     foreach my $ds (@ds) {
94     next if (! $ds->{'swish'});
95    
96     $index->insert(
97     index_name => $ds->{'tag'},
98 dpavlin 392 path => $f,
99 dpavlin 390 headline => $headline,
100     words => join(" ",@{$ds->{'swish'}})
101     );
102     }
103    
104 dpavlin 410 # save into sorted index (thesaurus)
105     foreach my $ds (@ds) {
106     next if (! $ds->{'index'});
107    
108     $thes->{$ds->{'tag'}} ||= new WebPAC::Index;
109    
110 dpavlin 415 foreach my $h (@{$ds->{'index'}}) {
111     $thes->{$ds->{'tag'}}->insert(
112     path => $f,
113     headline => $h,
114     );
115     }
116 dpavlin 410 }
117    
118 dpavlin 390 # print Dumper(\@ds);
119    
120 dpavlin 3 }
121    
122 dpavlin 411 foreach my $t (keys %{$thes}) {
123    
124     my @e = $thes->{$t}->elements;
125     if (! @e) {
126     $log->logwarn("no elements in sorted index $t?");
127     next;
128     }
129    
130 dpavlin 415 my $file = "./out/bfilter/$t.txt";
131     $log->info("saving sorted index $t to '$file' [".scalar(@e)." elements]");
132 dpavlin 411
133     $webpac->output_file(
134 dpavlin 415 file => $file,
135 dpavlin 411 template => 'index.tt',
136     data => \@e,
137     index_name => $t,
138     );
139     }
140    
141     if (0 && $log->is_debug) {
142 dpavlin 372 $log->debug("lookup hash: ",Dumper($webpac->{'lookup'}));
143     $log->debug("data hash: ",Dumper($webpac->{'data'}));
144 dpavlin 410 foreach my $t (keys %{$thes}) {
145     $log->debug("thesaurus $t hash: ",Dumper($thes->{$t}));
146     }
147 dpavlin 372 }
148 dpavlin 390
149 dpavlin 420 $log->info("lookup hash: ",Dumper($webpac->{'lookup'}));
150    
151 dpavlin 419 #
152     # tree dump implementation which shouldn't be here :-)
153     #
154    
155     my $file = 'out/browse.html';
156     open(HTML, "> $file") || $log->logdie("can't open '$file'");
157    
158     print HTML qq{
159     <html>
160     <head>
161     <title>Browse</title>
162     <meta http-equiv="Content-Type" content="text/html; charset=iso-8859-2"/>
163 dpavlin 423 <script type="text/javascript" src="bfilter/hide.js"></script>
164 dpavlin 419 </head>
165     <body>
166     <ul>
167     };
168    
169     foreach my $code (sort keys %{$webpac->{'lookup'}}) {
170    
171     my $l = $webpac->{'lookup'} || $log->logconfess("can't find lookup");
172    
173     if ($code =~ m/^root:/) {
174     my $v900 = shift @{ $l->{$code} } || $log->logconfess("can't lookup '$code'");
175     $code =~ s/^root:// || die;
176    
177     my $term = shift @{ $l->{"d:${v900}"} } || die;
178     my $mfn = shift @{ $l->{"900_mfn:${v900}"} } || die;
179    
180 dpavlin 421 $log->debug("$code -> $v900 : $term [$mfn]");
181 dpavlin 419
182 dpavlin 423 print HTML qq{ <li><a href="#mfn$mfn" onClick="return hide_show('mfn$mfn');">$term</a>&nbsp;<a href="thes/$mfn.html">&raquo;</a></li>\n} if (-e "out/thes/$mfn.html");
183 dpavlin 419
184 dpavlin 420 unless ($l->{"a:${code}::"}) {
185     $log->warn("can't find 'a:${code}::'");
186     next;
187     }
188 dpavlin 419
189 dpavlin 423 print HTML qq{ <a name="mfn$mfn"></a><ul id="mfn$mfn" style="display: none">\n};
190 dpavlin 419
191     foreach my $l2_v900 (@{ $l->{"a:${code}::"} }) {
192    
193     my $l2_code = shift @{ $l->{"code:${l2_v900}"} } || die;
194     my $l2_term = shift @{ $l->{"d:$l2_v900"} } || die;
195     my $l2_mfn = shift @{ $l->{"900_mfn:${l2_v900}"} } || die;
196    
197 dpavlin 421 $log->debug("$l2_code -> $l2_v900 : $l2_term [$l2_mfn]");
198 dpavlin 419
199 dpavlin 423 print HTML qq{ <li><a href="#mfn$l2_mfn" onClick="return hide_show('mfn$l2_mfn');">$l2_term</a>&nbsp;<a href="thes/$mfn.html">&raquo;</a></li>\n} if (-e "out/thes/$l2_mfn.html");
200 dpavlin 420
201     unless ($l->{"a:${l2_code}:"}) {
202     $log->warn("can't find 'a:${l2_code}:'");
203     next;
204     }
205 dpavlin 419
206 dpavlin 423 print HTML qq{ <a name="mfn$l2_mfn"></a><ul id="mfn$l2_mfn" style="display: none">\n};
207 dpavlin 419
208     foreach my $l3_v900 (@{ $l->{"a:${l2_code}:"} }) {
209    
210     my $l3_term = shift @{ $l->{"d:$l3_v900"} } || die;
211     my $l3_mfn = shift @{ $l->{"900_mfn:${l3_v900}"} } || die;
212    
213 dpavlin 421 $log->debug(" -> $l3_v900 : $l3_term [$l3_mfn]");
214 dpavlin 419
215 dpavlin 423 print HTML qq{ <li><a href="#mfn$l3_mfn.html">$l3_term</a>&nbsp;<a href="thes/$mfn.html">&raquo;</a></li>\n} if (-e "out/thes/$l3_mfn.html");
216 dpavlin 419 }
217    
218     print HTML qq{ </ul>\n};
219     }
220     print HTML qq{ </ul>\n};
221     }
222     }
223    
224     print HTML qq{
225     </ul>
226     </body>
227     </html>
228     };
229    
230 dpavlin 390 $index->close;
231    
232 dpavlin 422 $log->info("elapsed time: ",$webpac->fmt_time(time() - $webpac->{'start_t'}));

Properties

Name Value
cvs2svn:cvs-rev 1.64
svn:executable *

  ViewVC Help
Powered by ViewVC 1.1.26