[EP-tech] Re: Importing several items from an XML file

[Elizabeth Jennings]

I've attached it.

Lizz

On 15/07/2014 18:42, Sebastien Francois wrote:
> Code? :-)
>
> Or: where does your import code reside (in which function/method)?
>
> Seb.
>
> On 15/07/14 17:00, Lizz Jennings wrote:
>> Hello all,
>>
>> I'm writing an import plugin - I've managed to get it to match the right
>> fields and single records will import fine (this is through command line
>> at present).  However, although it'll iterate through each of the
>> records, it just overwrites the second over the first. So if I put the
>> return statement inside the foreach, I get the first record, and if I
>> put it at the end, I get the second record.
>>
>> How can I make it make a list of items to import?
>>
>> Lizz
>>
>
> *** Options: http://mailman.ecs.soton.ac.uk/mailman/listinfo/eprints-tech
> *** Archive: http://www.eprints.org/tech.php/
> *** EPrints community wiki: http://wiki.eprints.org/
> *** EPrints developers Forum: http://forum.eprints.org/
>
-------------- next part --------------
=head1 NAME

EPrints::Plugin::Import::batb::PureImport

=cut
package EPrints::Plugin::Import::batb::PureImport;

use strict;

use EPrints::Plugin::Import::DefaultXML;

our @ISA = qw/ EPrints::Plugin::Import::DefaultXML /;

sub new
{
	my( $class, %params ) = @_;

	my $self = $class->SUPER::new(%params);

	$self->{name} = "Pure Import";
	$self->{visible} = "all";
	$self->{produce} = [ 'list/eprint', 'dataobj/eprint' ];

	return $self;
}

sub top_level_tag
{
	my( $plugin, $dataset ) = @_;

	return "dataset:GetDataSetsResponse";
}

sub handler_class { "EPrints::Plugin::Import::DefaultXML::DOMHandler" }

sub xml_to_epdata
{
	my( $plugin, $dataset, $xml ) = @_;
	
	my $epdata = {};
	
	my $nscore = "http://atira.dk/schemas/pure4/model/core/stable";
	my $nsstab = "http://atira.dk/schemas/pure4/model/template/dataset/stable";
	my $nspers = "http://atira.dk/schemas/pure4/model/template/abstractperson/stable";
	
	my $puredatasets = $xml->getElementsByTagNameNS($nscore, "result")->item(0);
	return unless defined $puredatasets;
	
	my $purecontent = $puredatasets->getElementsByTagNameNS($nscore, "content"); #->item(0);
	return unless defined $purecontent;
		
	foreach my $pureitem (@$purecontent)
		{
			my $title = $pureitem->getElementsByTagNameNS( $nsstab, "title" )->item(0)->getElementsByTagNameNS( $nscore, "localizedString")->item(0);
			$epdata->{title} = $plugin->xml_to_text( $title ) if defined $title;

			my $pubdate = $pureitem->getElementsByTagNameNS( $nsstab, "dateMadeAvailable" )->item(0);
			if( defined $pubdate )
			{
				my $year  = $pubdate->getElementsByTagNameNS( $nscore, "year" )->item(0);
				my $month = $pubdate->getElementsByTagNameNS( $nscore, "month" )->item(0);
				my $day   = $pubdate->getElementsByTagNameNS( $nscore, "day" )->item(0);
				if (defined $year) # some pubdates have MedlineDate subfield (non parseable date : http://www.nlm.nih.gov/bsd/licensee/elements_descriptions.html#medlinedate)
				{
					my $tmpDate = $plugin->xml_to_text( $year );
					if (defined $month)
					{
						$month = $plugin->xml_to_text( $month );
						if (length $month == 1) # convert 1 to 01
						{
							$month = '0' . $month ;
						}
						$tmpDate .= '-' . $month ; 
						if (defined $day)
						{
							$day = $plugin->xml_to_text( $day );
							if (length $day == 1) # convert 1 to 01
							{
								$day = '0' . $day ;
							}
							$tmpDate .= '-' . $day ;
						}	
					}
					if( defined $tmpDate )
					{
						$epdata->{date} = $tmpDate;
						$epdata->{date_type} = "published";
					}
				}
			}

		my $creatorlist = $pureitem->getElementsByTagNameNS( $nsstab, "persons" )->item(0);
		if( defined $creatorlist )
		{
			foreach my $creator ( $creatorlist->getElementsByTagNameNS($nsstab, "dataSetPersonAssociation") )
			{
				my $role = $creator->getElementsByTagNameNS( $nspers, "personRole")->item(0)->getElementsByTagNameNS( $nscore, "uri")->item(0);
				$role = $plugin->xml_to_text($role);
				if( $role eq "/dk/atira/pure/dataset/roles/dataset/creator" )
					{
						my $name = {};
			
						my $lastname = $creator->getElementsByTagNameNS( $nspers, "name" )->item(0)->getElementsByTagNameNS( $nscore, "lastName" )->item(0);
						$name->{family} = $plugin->xml_to_text( $lastname ) if defined $lastname;
						
						my $forename = $creator->getElementsByTagNameNS( $nspers, "name" )->item(0)->getElementsByTagNameNS( $nscore, "firstName" )->item(0);
						$name->{given} = $plugin->xml_to_text( $forename ) if defined $forename;

					push @{ $epdata->{creators_name} }, $name;
					
					}
			}
		}

		$epdata->{type} = "data_collection";
		$epdata->{eprint_status} = "inbox";
		
		return $epdata;
	}

	#return $epdata;

}

1;