/[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 286 - (hide annotations)
Sun Dec 18 21:06:46 2005 UTC (18 years, 4 months ago) by dpavlin
File MIME type: text/plain
File size: 3717 byte(s)
 r11778@llin:  dpavlin | 2005-12-19 03:59:54 +0100
 move work on input

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 dpavlin 255 use WebPAC::Common 0.02;
11 dpavlin 74 use WebPAC::Lookup;
12 dpavlin 285 use WebPAC::Input::ISIS 0.02;
13 dpavlin 209 use WebPAC::Store 0.03;
14 dpavlin 74 use WebPAC::Normalize::XML;
15     use WebPAC::Output::TT;
16 dpavlin 255 use WebPAC::Output::Estraier 0.05;
17 dpavlin 141 use YAML qw/LoadFile/;
18 dpavlin 210 use LWP::Simple;
19 dpavlin 74
20 dpavlin 76 my $limit = shift @ARGV;
21    
22 dpavlin 141 my $config = LoadFile('conf/config.yml');
23    
24     print "config = ",Dumper($config);
25    
26 dpavlin 210 die "no databases in config file!\n" unless ($config->{databases});
27 dpavlin 141
28 dpavlin 213 my $total_rows = 0;
29    
30 dpavlin 210 while (my ($database, $db_config) = each %{ $config->{databases} }) {
31 dpavlin 141
32 dpavlin 255 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 dpavlin 210 my $abs_path = abs_path($0);
51     $abs_path =~ s#/[^/]*$#/#;
52 dpavlin 141
53 dpavlin 210 my $db_path = $config->{webpac}->{db_path} . '/' . $database;
54 dpavlin 74
55 dpavlin 255 $log->info("working on $database in $db_path");
56    
57 dpavlin 210 my $db = new WebPAC::Store(
58     path => $db_path,
59 dpavlin 217 database => $database,
60 dpavlin 221 debug => 1,
61 dpavlin 210 );
62 dpavlin 74
63 dpavlin 233
64 dpavlin 213 #
65     # now, iterate through input formats
66     #
67 dpavlin 74
68 dpavlin 213 my @inputs;
69     if (ref($db_config->{input}) eq 'ARRAY') {
70     @inputs = @{ $db_config->{input} };
71 dpavlin 255 } elsif ($db_config->{input}) {
72     push @inputs, $db_config->{input};
73 dpavlin 213 } else {
74 dpavlin 255 $log->info("database $database doesn't have inputs defined");
75 dpavlin 213 }
76 dpavlin 74
77 dpavlin 286 my @supported_inputs = keys %{ $config->{webpac}->{inputs} };
78    
79 dpavlin 213 foreach my $input (@inputs) {
80 dpavlin 233
81     my $type = lc($input->{type});
82    
83 dpavlin 286 die "I know only how to handle input types ", join(",", @supported_inputs), " not '$type'!\n" unless (grep(/$type/, @supported_inputs));
84 dpavlin 233
85     my $lookup = new WebPAC::Lookup(
86     lookup_file => $input->{lookup},
87     );
88    
89 dpavlin 286 my $input_module = $config->{webpac}->{inputs}->{$type};
90 dpavlin 74
91 dpavlin 286 $log->info("working on input $input->{path} [$input->{type}] using $input_module");
92    
93     sub new_input {
94     my $name = shift;
95     my $args = shift;
96     new $name->($args);
97     }
98    
99     my $input = new_input($input_module,{
100 dpavlin 213 code_page => $config->{webpac}->{webpac_encoding},
101     limit_mfn => $input->{limit},
102 dpavlin 251 lookup => $lookup,
103 dpavlin 286 });
104     $log->logdie("can't create input using $input_module") unless ($input);
105 dpavlin 113
106 dpavlin 286 my $maxmfn = $input->open(
107 dpavlin 285 path => $input->{path},
108 dpavlin 213 code_page => $input->{encoding}, # database encoding
109     );
110 dpavlin 113
111 dpavlin 213 my $n = new WebPAC::Normalize::XML(
112     # filter => { 'foo' => sub { shift } },
113     db => $db,
114     lookup_regex => $lookup->regex,
115     lookup => $lookup,
116 dpavlin 221 prefix => $input->{name},
117 dpavlin 213 );
118 dpavlin 113
119 dpavlin 269 my $normalize_path = $input->{normalize}->{path};
120 dpavlin 210
121 dpavlin 269 if ($normalize_path =~ m/\.xml$/i) {
122     $n->open(
123     tag => $input->{normalize}->{tag},
124     xml_file => $input->{normalize}->{path},
125     );
126     } elsif ($normalize_path =~ m/\.(?:yml|yaml)$/i) {
127     $n->open_yaml(
128     path => $normalize_path,
129     tag => $input->{normalize}->{tag},
130     );
131     }
132    
133 dpavlin 286 for ( 0 ... $input->size ) {
134 dpavlin 210
135 dpavlin 286 my $row = $input->fetch || next;
136 dpavlin 213
137     my $mfn = $row->{'000'}->[0] || die "can't find MFN";
138    
139     my $ds = $n->data_structure($row);
140    
141     $est->add(
142     id => $input->{name} . "#" . $mfn,
143     ds => $ds,
144     type => $config->{hyperestraier}->{type},
145     );
146    
147     $total_rows++;
148     }
149    
150 dpavlin 210 };
151    
152     $log->info("$total_rows records indexed");
153 dpavlin 255
154     #
155     # add Hyper Estraier links to other databases
156     #
157     if (ref($db_config->{links}) eq 'ARRAY') {
158     foreach my $link (@{ $db_config->{links} }) {
159     $log->info("adding link $database -> $link->{to} [$link->{credit}]");
160     $est->add_link(
161     from => $database,
162     to => $link->{to},
163     credit => $link->{credit},
164     );
165     }
166     }
167    
168 dpavlin 210 }
169    

Properties

Name Value
svn:executable *

  ViewVC Help
Powered by ViewVC 1.1.26