Diff for /nsdl/nsdlloncapaorg/harvester.pl between versions 1.3 and 1.4

version 1.3, 2003/07/29 14:13:36 version 1.4, 2003/07/29 14:37:51
Line 12  use strict; Line 12  use strict;
 use LWP::UserAgent;  use LWP::UserAgent;
 use Getopt::Std;  use Getopt::Std;
 use Digest::MD5 qw(md5_hex);  use Digest::MD5 qw(md5_hex);
   use IO::File;
   
   my $basepath='/home/httpd/cgi-bin/OAI-XMLFile/XMLFile/nsdlexport/data';
   
 my $pub_month;  my $pub_month;
 my $pub_year;  my $pub_year;
Line 60  if ( $response->is_success ) { Line 63  if ( $response->is_success ) {
 #}  #}
   
 my %records = ();;  my %records = ();;
 print '<?xml version="1.0" encoding="UTF-8"?>'."\n\n";  
   
 foreach my $metadata (@loncapa) {  foreach my $metadata (@loncapa) {
  chomp $metadata;   chomp $metadata;
Line 84  foreach my $metadata (@loncapa) { Line 86  foreach my $metadata (@loncapa) {
  next if ( ($subject eq 'Sample') || ($subject eq 'Something') );   next if ( ($subject eq 'Sample') || ($subject eq 'Something') );
  my $resourceurl = 'http://nsdl.lon-capa.org' . $tkline[3];   my $resourceurl = 'http://nsdl.lon-capa.org' . $tkline[3];
         my $baseid=$tkline[3];          my $baseid=$tkline[3];
    my ($adom,$auname)=($baseid=~/^\/res\/(\w+)\/(\w+)\//);
  $baseid=~s/\W/\_/g;   $baseid=~s/\W/\_/g;
  $baseid=~s/^\_res\_//g;   $baseid=~s/^\_res\_//g;
    my $fileid=md5_hex($baseid);
   
  next if ( $resourceurl =~ /(.*)\/demo\/(.*)/ );   next if ( $resourceurl =~ /(.*)\/demo\/(.*)/ );
  my $keywords = $tkline[4];   my $keywords = $tkline[4];
Line 150  foreach my $metadata (@loncapa) { Line 154  foreach my $metadata (@loncapa) {
  # Private means open only to author of material   # Private means open only to author of material
  next if ( $copyright eq 'private');   next if ( $copyright eq 'private');
  my $platform = "5";     # HTML Browser (not specified but construed from metadata)   my $platform = "5";     # HTML Browser (not specified but construed from metadata)
  print (<<ENDMETA);  #
   # Create path
   #
    unless (-e $basepath.'/'.$adom) { mkdir($basepath.'/'.$adom); }
    unless (-e $basepath.'/'.$adom.'/'.$auname) { 
       mkdir($basepath.'/'.$adom.'/'.$auname) || die 'Could not create '.$basepath.'/'.$adom.'/'.$auname;
    }
    open(XML,'>'.$basepath.'/'.$adom.'/'.$auname.'/'.$baseid.'.xml');
    print XML (<<ENDMETA);
   <?xml version="1.0" encoding="UTF-8"?>
   
 <oaidc:dc xmlns="http://purl.org/dc/elements/1.1/  <oaidc:dc xmlns="http://purl.org/dc/elements/1.1/
           xmlns:oaidc="http://www.openarchives.org/OAI/2.0/oai_dc/            xmlns:oaidc="http://www.openarchives.org/OAI/2.0/oai_dc/
           xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance            xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance
Line 166  foreach my $metadata (@loncapa) { Line 180  foreach my $metadata (@loncapa) {
     <description>$abstract</description>      <description>$abstract</description>
     <date>$rev_year-$rev_month-$rev_day</date>      <date>$rev_year-$rev_month-$rev_day</date>
 </oaidc:dc>  </oaidc:dc>
   
 ENDMETA  ENDMETA
         close (XML);
 }  }

Removed from v.1.3  
changed lines
  Added in v.1.4


FreeBSD-CVSweb <freebsd-cvsweb@FreeBSD.org>