/[Search-Estraier]/trunk/scripts/dbi-indexer.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/scripts/dbi-indexer.pl

Parent Directory Parent Directory | Revision Log Revision Log


Revision 147 - (hide annotations)
Wed May 10 21:41:35 2006 UTC (18 years ago) by dpavlin
File MIME type: text/plain
File size: 1901 byte(s)
added pk_col to config hash
1 dpavlin 131 #!/usr/bin/perl -w
2    
3     use strict;
4 dpavlin 144 use Search::Estraier 0.06;
5 dpavlin 131 use DBI;
6     use Data::Dumper;
7     use Encode qw/from_to/;
8 dpavlin 146 use Time::HiRes qw/time/;
9 dpavlin 131
10     =head1 NAME
11    
12     dbi-indexer.pl - example indexer of DBI sources for Search::Estraier
13    
14     =cut
15    
16 dpavlin 144 my $c = {
17     node_url => 'http://localhost:1978/node/dbi',
18     dbi => 'Pg:dbname=azop',
19     sql => qq{
20     select * from history_collection_view_cache
21     },
22 dpavlin 147 pk_col => '_id',
23 dpavlin 145 db_encoding => 'iso-8859-2',
24 dpavlin 144 debug => 0,
25 dpavlin 131 };
26    
27     # create and configure node
28     my $node = new Search::Estraier::Node(
29 dpavlin 144 url => $c->{node_url},
30 dpavlin 131 user => 'admin',
31     passwd => 'admin',
32     croak_on_error => 1,
33 dpavlin 144 create => 1,
34     debug => $c->{debug},
35 dpavlin 131 );
36    
37     # create DBI connection
38 dpavlin 144 my $dbh = DBI->connect("DBI:$c->{dbi}","","") || die $DBI::errstr;
39 dpavlin 131
40 dpavlin 144 my $sth = $dbh->prepare($c->{sql}) || die $dbh->errstr();
41 dpavlin 131 $sth->execute() || die $sth->errstr();
42    
43 dpavlin 144 warn "# columns: ",join(",",@{ $sth->{NAME} }),"\n" if ($c->{debug});
44 dpavlin 131
45     my $total = $sth->rows;
46     my $i = 1;
47    
48 dpavlin 146 my $t = time();
49 dpavlin 147 my $pk_col = $c->{pk_col} || 'id';
50 dpavlin 146
51 dpavlin 131 while (my $row = $sth->fetchrow_hashref() ) {
52    
53 dpavlin 144 warn "# row: ",Dumper($row) if ($c->{debug});
54 dpavlin 131
55     # create document
56     my $doc = new Search::Estraier::Document;
57    
58 dpavlin 147 if (my $id = $row->{$pk_col}) {
59     $doc->add_attr('@uri', $id);
60     } else {
61     die "can't find pk_col column '$pk_col' in results\n";
62     }
63 dpavlin 131
64     printf "%4d ",$i;
65    
66     while (my ($col,$val) = each %{$row}) {
67    
68     if ($val) {
69     # change encoding?
70 dpavlin 145 from_to($val, ($c->{db_encoding} || 'ISO-8859-1'), 'UTF-8');
71 dpavlin 131
72     # add attributes (make column usable from attribute search)
73     $doc->add_attr($col, $val);
74    
75     # add body text to document (make it searchable using full-text index)
76     $doc->add_text($val);
77    
78     print "R";
79     } else {
80     print ".";
81     }
82    
83     }
84    
85 dpavlin 144 warn "# doc draft: ",$doc->dump_draft, "\n" if ($c->{debug});
86 dpavlin 131
87     die "error: ", $node->status,"\n" unless (eval { $node->put_doc($doc) });
88 dpavlin 146
89     printf (" %d%% %.1f/s\n", int(( $i++ / $total) * 100), ( $i / (time() - $t) ) );
90    
91 dpavlin 131 }

Properties

Name Value
svn:executable *

  ViewVC Help
Powered by ViewVC 1.1.26