/[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

Contents of /trunk/run.pl

Parent Directory Parent Directory | Revision Log Revision Log


Revision 268 - (show annotations)
Fri Dec 16 21:09:42 2005 UTC (18 years, 3 months ago) by dpavlin
File MIME type: text/plain
File size: 3112 byte(s)
 r11742@llin:  dpavlin | 2005-12-17 01:26:41 +0100
 cleanup

1 #!/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::Common 0.02;
11 use WebPAC::Lookup;
12 use WebPAC::Input::ISIS;
13 use WebPAC::Store 0.03;
14 use WebPAC::Normalize::XML;
15 use WebPAC::Output::TT;
16 use WebPAC::Output::Estraier 0.05;
17 use YAML qw/LoadFile/;
18 use LWP::Simple;
19
20 my $limit = shift @ARGV;
21
22 my $config = LoadFile('conf/config.yml');
23
24 print "config = ",Dumper($config);
25
26 die "no databases in config file!\n" unless ($config->{databases});
27
28 my $total_rows = 0;
29
30 while (my ($database, $db_config) = each %{ $config->{databases} }) {
31
32 my $log = _new WebPAC::Common()->_get_logger();
33
34 #
35 # open Hyper Estraier database
36 #
37
38 my $est_config = $config->{hyperestraier} || $log->logdie("can't find 'hyperestraier' part in confguration");
39 $est_config->{database} = $database;
40
41 $log->info("using HyperEstraier URL $est_config->{masterurl}");
42
43 my $est = new WebPAC::Output::Estraier(
44 %{ $est_config },
45 );
46
47 #
48 # now WebPAC::Store
49 #
50 my $abs_path = abs_path($0);
51 $abs_path =~ s#/[^/]*$#/#;
52
53 my $db_path = $config->{webpac}->{db_path} . '/' . $database;
54
55 $log->info("working on $database in $db_path");
56
57 my $db = new WebPAC::Store(
58 path => $db_path,
59 database => $database,
60 debug => 1,
61 );
62
63
64 #
65 # now, iterate through input formats
66 #
67
68 my @inputs;
69 if (ref($db_config->{input}) eq 'ARRAY') {
70 @inputs = @{ $db_config->{input} };
71 } elsif ($db_config->{input}) {
72 push @inputs, $db_config->{input};
73 } else {
74 $log->info("database $database doesn't have inputs defined");
75 }
76
77 foreach my $input (@inputs) {
78
79 my $type = lc($input->{type});
80
81 die "I know only how to handle input type isis, not '$type'!\n" unless ($type eq 'isis');
82
83 my $lookup = new WebPAC::Lookup(
84 lookup_file => $input->{lookup},
85 );
86
87 $log->info("working on input $input->{path} [$input->{type}]");
88
89 my $isis = new WebPAC::Input::ISIS(
90 code_page => $config->{webpac}->{webpac_encoding},
91 limit_mfn => $input->{limit},
92 lookup => $lookup,
93 );
94
95 my $maxmfn = $isis->open(
96 filename => $input->{path},
97 code_page => $input->{encoding}, # database encoding
98 );
99
100 my $n = new WebPAC::Normalize::XML(
101 # filter => { 'foo' => sub { shift } },
102 db => $db,
103 lookup_regex => $lookup->regex,
104 lookup => $lookup,
105 prefix => $input->{name},
106 );
107
108 $n->open(
109 tag => $input->{normalize}->{tag},
110 xml_file => $input->{normalize}->{path},
111 );
112
113 for ( 0 ... $isis->size ) {
114
115 my $row = $isis->fetch || next;
116
117 my $mfn = $row->{'000'}->[0] || die "can't find MFN";
118
119 my $ds = $n->data_structure($row);
120
121 $est->add(
122 id => $input->{name} . "#" . $mfn,
123 ds => $ds,
124 type => $config->{hyperestraier}->{type},
125 );
126
127 $total_rows++;
128 }
129
130 };
131
132 $log->info("$total_rows records indexed");
133
134 #
135 # add Hyper Estraier links to other databases
136 #
137 if (ref($db_config->{links}) eq 'ARRAY') {
138 foreach my $link (@{ $db_config->{links} }) {
139 $log->info("adding link $database -> $link->{to} [$link->{credit}]");
140 $est->add_link(
141 from => $database,
142 to => $link->{to},
143 credit => $link->{credit},
144 );
145 }
146 }
147
148 }
149

Properties

Name Value
svn:executable *

  ViewVC Help
Powered by ViewVC 1.1.26