Versions Compared

Old Version 3

changes.mady.by.user Martin Walk

Saved on

compared with

New Version Current

changes.mady.by.user Martin Walk

Saved on

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.
Comment: Add howto label

Motivation: Search for filenames and file descriptions in DSpace

Out of the box discovery in DSpace 6.0 worked quite well. But it did not include filenames and file descriptions in the search index.

As we planned to allow research data publications on our repository, there would be publications with many files (as opposed to text publications which usually consist of only one pdf) and it would be nice to be able to search for filenames and file descriptions.

Add additional fields and values to Solr index

Sometimes extending DSpace turns out to be very easy and straightforward, albeit not very well documented.

To add fields and values to the discovery index, just create a new implementation of the SolrServiceIndexPlugin class and add your fields and values to the document in the additionalIndex method:

Code Block
languagejava
linenumberstrue
package org.dspace.discovery;
import org.apache.solr.common.SolrInputDocument;
import org.dspace.content.DSpaceObject;
import org.dspace.content.Item;
import org.dspace.core.Context;
public class SolrServiceHelloWorldPlugin implements SolrServiceIndexPlugin {
   @Override
   public void additionalIndex(Context context, DSpaceObject dso, SolrInputDocument document) {
      if (dso instanceof Item) {
         Item item = (Item) dso;
         document.addField("greeting","Hello World!");
      }
   }
}

Add your plugin to discovery.xml as a bean:

Code Block
linenumberstrue
<bean id="solrServiceHelloWorldPlugin" class="org.dspace.discovery.SolrServiceHelloWorldPlugin"/>

Recompile, distribute and rebuild the discovery index:

Code Block
linenumberstrue
[dspace]/bin/dspace index-discovery -b


Add multivalued fields

By default, a new field added this way will be multivalued. To add multiple values to such a field in a Solr document you can either call the addField method several times with the same field name:

Code Block
languagejava
linenumberstrue
document.addField("greeting","Hello World!");
document.addField("greeting","Bonjour Monde!");

Or call it once with a collection as value:

Code Block
languagejava
linenumberstrue
String[] greetings = {"Hello World!", "Bonjour Monde!"};
document.addField("greeting", greetings);

SolrServiceFileInfoPlugin

This is what our SolrServiceFileInfoPlugin looks like:

Code Block
languagejava
linenumberstrue
package org.dspace.discovery;

import org.apache.solr.common.SolrInputDocument;
import org.dspace.content.Bitstream;
import org.dspace.content.Bundle;
import org.dspace.content.DSpaceObject;
import org.dspace.content.Item;
import org.dspace.core.Context;

import java.util.List;

public class SolrServiceFileInfoPlugin implements SolrServiceIndexPlugin
{
    private static final String BUNDLE_NAME = "ORIGINAL";
    private static final String SOLR_FIELD_NAME_FOR_FILENAMES = "original_bundle_filenames";
    private static final String SOLR_FIELD_NAME_FOR_DESCRIPTIONS = "original_bundle_descriptions";

    @Override
    public void additionalIndex(Context context, DSpaceObject dso, SolrInputDocument document)
    {
        if (dso instanceof Item)
        {
            Item item = (Item) dso;
            List<Bundle> bundles = item.getBundles();
            if (bundles != null)
            {
                for (Bundle bundle : bundles)
                {
                    String bundleName = bundle.getName();
                    if ((bundleName != null) && bundleName.equals(BUNDLE_NAME))
                    {
                        List<Bitstream> bitstreams = bundle.getBitstreams();
                        if (bitstreams != null)
                        {
                            for (Bitstream bitstream : bitstreams)
                            {
                                document.addField(SOLR_FIELD_NAME_FOR_FILENAMES, bitstream.getName());

                                String description = bitstream.getDescription();
                                if ((description != null) && (!description.isEmpty()))
                                {
                                    document.addField(SOLR_FIELD_NAME_FOR_DESCRIPTIONS, description);
                                }
                            }
                        }
                    }
                }
            }
        }
    }
}