/[wait]/trunk/script/index_ora
This is repository of my old source code which isn't updated any more. Go to git.rot13.org for current projects!
ViewVC logotype

Diff of /trunk/script/index_ora

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

revision 65 by laperla, Wed Jan 23 12:22:54 2002 UTC revision 71 by laperla, Sun Jan 27 15:27:38 2002 UTC
# Line 1  Line 1 
1  #!/usr/bin/perl -w  #!/usr/bin/perl -w
2  #                              -*- Mode: Perl -*-  #                              -*- Mode: Perl -*-
3  # $Basename$  # $Basename$
4  # $Revision: 1.4 $  # $Revision: 1.8 $
5  # Author          : Ulrich Pfeifer  # Author          : Ulrich Pfeifer
6  # Created On      : Mon Dec 31 13:57:11 2001  # Created On      : Mon Dec 31 13:57:11 2001
7  # Last Modified By: Ulrich Pfeifer  # Last Modified By: Ulrich Pfeifer
# Line 11  Line 11 
11  # (C) Copyright 2001, UUNET Deutschland GmbH, Germany  # (C) Copyright 2001, UUNET Deutschland GmbH, Germany
12  #  #
13    
14    use 5.007;
15    
16  use strict;  use strict;
17    use Devel::Peek qw(Dump);
18    
19  use File::Path;  use File::Path;
20  use DB_File;  use DB_File;
21  use Getopt::Long;  use Getopt::Long;
22  use Cwd;  use Cwd;
23    
24  require WAIT::Config;  BEGIN {require WAIT::Config;}
25  require WAIT::Database;  use WAIT::Database;
26  require WAIT::Parse::Ora;  use WAIT::Parse::Ora;
27  require WAIT::Document::Ora;  use WAIT::Document::Ora;
28  require WAIT::InvertedIndex;  use WAIT::InvertedIndex;
29    
30    
31  $DB_BTREE->{'cachesize'} = 200_000 ;  $DB_BTREE->{'cachesize'} = 200_000 ;
32    
33  my %OPT = (clean    => 0,  my %OPT = (
34             database => 'DB',             database => 'DB',
35             dir      => $WAIT::Config->{WAIT_home} || '/tmp',             dir      => $WAIT::Config->{WAIT_home} || '/tmp',
36             table    => 'ora',             table    => 'ora',
37            );            );
38    
39  GetOptions(\%OPT,  GetOptions(\%OPT,
            'clean!',  
40             'database=s',             'database=s',
41             'dir=s',             'dir=s',
42             'table=s',             'table=s',
43            ) || die "Usage: ...\n";            ) || die "Usage: ...\n";
44    
45  if ($OPT{clean} and -d "$OPT{dir}/$OPT{database}") {  my @localtime = localtime;
46    my $tmp = WAIT::Database->open(name        => $OPT{database},  $localtime[5] += 1900;
47                                   'directory' => $OPT{dir})  $localtime[4]++;
48      or die "Could not open table $OPT{table}: $@\n";  my $jobid = sprintf "%04s-%02s-%02s_%02s:%02s_%d", @localtime[5,4,3,2,1], $$;
49    my $tbl = $tmp->table(name => $OPT{table});  my $db = WAIT::Database->create(name      => "$OPT{database}-$jobid",
50    $tbl->drop if $tbl;                                  directory => $OPT{dir})
51    rmtree("$OPT{dir}/$OPT{database}/$OPT{table}", 1, 1)      or die "Could not create database $OPT{database}: $@\n";
     if -d "$OPT{dir}/$OPT{database}/$OPT{table}";  
   $tmp->close;  
 }  
   
 my $db;  
 unless (-d "$OPT{dir}/$OPT{database}") {  
   $db = WAIT::Database->create(name       => $OPT{database},  
                               'directory' => $OPT{dir})  
     or die "Could not open database $OPT{database}: $@\n";  
 }  
 else {  
   $db = WAIT::Database->open(name        => $OPT{database},  
                              'directory' => $OPT{dir})  
     or die "Could not open table $OPT{table}: $@\n";  
 }  
52    
53  my $layout = new WAIT::Parse::Ora;  my $layout = new WAIT::Parse::Ora;
54    
55  my $stem  = ['isotr', 'isolc', 'split2', 'stop', 'Stem'];  use lib "/usr/local/apache/lib";
56    use oreilly_de_catalog::wait_handler;
57    
58    my $stem  = ['OR_tr_20020124', 'OR_lc_20020124', 'split2', 'stop', 'Stem'];
59  my $text  = [{  my $text  = [{
60                'prefix'    => ['isotr', 'isolc'],                'prefix'    => ['OR_tr_20020124', 'OR_lc_20020124'],
61                'intervall' => ['isotr', 'isolc'],                'intervall' => ['OR_tr_20020124', 'OR_lc_20020124'],
62               },               },
63               'isotr', 'isolc', 'split2', 'stop'];               'OR_tr_20020124', 'OR_lc_20020124', 'split2', 'stop'];
64  my $sound = ['isotr', 'isolc', 'split2', 'Soundex'],;  my $sound = ['OR_tr_20020124', 'OR_lc_20020124', 'split2', 'Soundex'];
65    my $trigr = ['OR_lc_20020124', 'OR_trigrams_20020125'];
66    
67  my $cwd = cwd;  my $cwd = cwd;
68    
# Line 97  my $tb = $db->create_table(name     => $ Line 89  my $tb = $db->create_table(name     => $
89  die "Couldn't create table $OPT{table}: $@\n" unless $tb;  die "Couldn't create table $OPT{table}: $@\n" unless $tb;
90    
91  my ($did, $value);  my ($did, $value);
92    binmode STDOUT, ":utf8";
93  while (($did, $value) = each %D) {  while (($did, $value) = each %D) {
94    my $record   = $layout->split($value);    my $record   = $layout->split($value);
95    my $headline = $record->{title};    my $headline = $record->{title};
# Line 107  while (($did, $value) = each %D) { Line 100  while (($did, $value) = each %D) {
100                %{$record});                %{$record});
101  }  }
102  $tb->set(top=>1);  $tb->set(top=>1);
103    
104    my $tritb = $db->create_table(
105                                  name => "$OPT{table}_fallback",
106                                  attr => [qw(docid headline)],
107                                  invindex => [ headline => $trigr ],
108                                 );
109    my %dict;
110    for my $f ($tb->fields) {
111      my(@idx) = @{$tb->table->{inverted}{$f} || []};
112      for my $idx (@idx) {
113        my @keys = $idx->keys;
114        @dict{@keys} = ();
115      }
116    }
117    my @dictkeys = grep s/^p//, keys %dict;
118    my $maxdebug = 5;
119    for my $headline (@dictkeys) {
120      if ($maxdebug && $headline =~ /[^\040-\177]/) {
121        Dump $headline;
122        $maxdebug--;
123      }
124      $tritb->insert(docid => $headline, headline => $headline);
125    }
126    $tritb->set(top=>1);
127    $tritb->close;
128  $tb->close();  $tb->close();
129  $db->close();  $db->close();
130    
131    # Now we have a new database with a very long name and we want that
132    # database to be accessible with the $OPT{database} name
133    
134    use File::Spec;
135    my $long_dir   = "$OPT{database}-$jobid";
136    my $want_dir   = File::Spec->catdir($OPT{dir}, $OPT{database});
137    my $prel_slink = File::Spec->catdir($OPT{dir}, "$OPT{database}-$$");
138    unlink $prel_slink; # may fail
139    symlink $long_dir, $prel_slink or die "Could not symlink $long_dir, $prel_slink: $!";
140    rename $prel_slink, $want_dir or die "Could not rename $prel_slink, $want_dir: $!";
141    
142  $WAIT::Config = $WAIT::Config; # make perl -w happy  $WAIT::Config = $WAIT::Config; # make perl -w happy
143    
144    
# Line 125  index_ora - generate an WAIT index for O Line 154  index_ora - generate an WAIT index for O
154  =head1 SYNOPSIS  =head1 SYNOPSIS
155    
156  B<index_ora>  B<index_ora>
 [B<-clean>] [B<-noclean>]  
157  [B<-database> I<dbname>]  [B<-database> I<dbname>]
158  [B<-dir> I<directory>]  [B<-dir> I<directory>]
159  [B<-table> I<table name>]  [B<-table> I<table name>]
# Line 137  I<directory> Line 165  I<directory>
165    
166  =over 5  =over 5
167    
 =item B<-clean> / B<-noclean>  
   
 Clean the table before indexing. Default is B<off>.  
   
168  =item B<-database> I<dbname>  =item B<-database> I<dbname>
169    
170  Specify database name. Default is F<DB>.  Specify database name. Default is F<DB>.

Legend:
Removed from v.65  
changed lines
  Added in v.71

  ViewVC Help
Powered by ViewVC 1.1.26