/[webpac2]/trunk/run.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 /trunk/run.pl

Parent Directory Parent Directory | Revision Log Revision Log


Revision 251 - (hide annotations)
Thu Dec 15 14:12:00 2005 UTC (18 years, 4 months ago) by dpavlin
File MIME type: text/plain
File size: 2621 byte(s)
various updates to make lookups work (but they don't still)

1 dpavlin 74 #!/usr/bin/perl -w
2    
3     use strict;
4    
5     use Cwd qw/abs_path/;
6     use File::Temp qw/tempdir/;
7     use Data::Dumper;
8     use lib './lib';
9    
10     use WebPAC::Lookup;
11     use WebPAC::Input::ISIS;
12 dpavlin 209 use WebPAC::Store 0.03;
13 dpavlin 74 use WebPAC::Normalize::XML;
14     use WebPAC::Output::TT;
15 dpavlin 210 use WebPAC::Output::Estraier 0.02;
16 dpavlin 141 use YAML qw/LoadFile/;
17 dpavlin 210 use LWP::Simple;
18 dpavlin 74
19 dpavlin 76 my $limit = shift @ARGV;
20    
21 dpavlin 141 my $config = LoadFile('conf/config.yml');
22    
23     print "config = ",Dumper($config);
24    
25 dpavlin 210 die "no databases in config file!\n" unless ($config->{databases});
26 dpavlin 141
27 dpavlin 213 my $total_rows = 0;
28    
29 dpavlin 210 while (my ($database, $db_config) = each %{ $config->{databases} }) {
30 dpavlin 141
31 dpavlin 210 my $abs_path = abs_path($0);
32     $abs_path =~ s#/[^/]*$#/#;
33 dpavlin 141
34 dpavlin 210 my $db_path = $config->{webpac}->{db_path} . '/' . $database;
35 dpavlin 74
36 dpavlin 210 my $db = new WebPAC::Store(
37     path => $db_path,
38 dpavlin 217 database => $database,
39 dpavlin 221 debug => 1,
40 dpavlin 210 );
41 dpavlin 74
42 dpavlin 233 my $log = $db->_get_logger;
43     $log->info("working on $database in $db_path");
44    
45 dpavlin 210 my $est_config = $config->{hyperestraier} || $log->logdie("can't find 'hyperestraier' part in confguration");
46     $est_config->{database} = $database;
47 dpavlin 113
48 dpavlin 213 $log->info("using HyperEstraier URL $est_config->{masterurl}");
49 dpavlin 74
50 dpavlin 210 my $est = new WebPAC::Output::Estraier(
51     %{ $est_config },
52     );
53 dpavlin 74
54 dpavlin 213 #
55     # now, iterate through input formats
56     #
57 dpavlin 74
58 dpavlin 213 my @inputs;
59     if (ref($db_config->{input}) eq 'ARRAY') {
60     @inputs = @{ $db_config->{input} };
61     } else {
62     push @inputs, $db_config->{input};
63     }
64 dpavlin 74
65 dpavlin 213 foreach my $input (@inputs) {
66 dpavlin 233
67     my $type = lc($input->{type});
68    
69     die "I know only how to handle input type isis, not '$type'!\n" unless ($type eq 'isis');
70    
71     my $lookup = new WebPAC::Lookup(
72     lookup_file => $input->{lookup},
73     );
74    
75 dpavlin 213 $log->info("working on input $input->{path} [$input->{type}]");
76 dpavlin 74
77 dpavlin 213 my $isis = new WebPAC::Input::ISIS(
78     code_page => $config->{webpac}->{webpac_encoding},
79     limit_mfn => $input->{limit},
80 dpavlin 251 lookup => $lookup,
81 dpavlin 213 );
82 dpavlin 113
83 dpavlin 213 my $maxmfn = $isis->open(
84     filename => $input->{path},
85     code_page => $input->{encoding}, # database encoding
86     );
87 dpavlin 113
88 dpavlin 251 $log->info( Dumper($lookup->{_lookup_data}) );
89    
90 dpavlin 213 my $n = new WebPAC::Normalize::XML(
91     # filter => { 'foo' => sub { shift } },
92     db => $db,
93     lookup_regex => $lookup->regex,
94     lookup => $lookup,
95 dpavlin 221 prefix => $input->{name},
96 dpavlin 213 );
97 dpavlin 113
98 dpavlin 213 $n->open(
99     tag => $input->{normalize}->{tag},
100     xml_file => $input->{normalize}->{path},
101 dpavlin 210 );
102    
103 dpavlin 213 for ( 0 ... $isis->size ) {
104 dpavlin 210
105 dpavlin 213 my $row = $isis->fetch || next;
106    
107     my $mfn = $row->{'000'}->[0] || die "can't find MFN";
108    
109     my $ds = $n->data_structure($row);
110    
111     $est->add(
112     id => $input->{name} . "#" . $mfn,
113     ds => $ds,
114     type => $config->{hyperestraier}->{type},
115     );
116    
117     $total_rows++;
118     }
119    
120 dpavlin 210 };
121    
122     $log->info("$total_rows records indexed");
123     }
124    

Properties

Name Value
svn:executable *

  ViewVC Help
Powered by ViewVC 1.1.26