Retrospectively assign DOIs to DSpace items (DSpace 5.x)

RetrospectivelyRegisterIdentifier.java

package nz.co.lernz.db.curation;

import org.apache.commons.lang.StringUtils;
import org.apache.log4j.Logger;
import org.dspace.authorize.AuthorizeException;
import org.dspace.content.DSpaceObject;
import org.dspace.content.Item;
import org.dspace.curate.AbstractCurationTask;
import org.dspace.curate.Curator;
import org.dspace.curate.Distributive;
import org.dspace.curate.Mutative;
import org.dspace.identifier.DOI;
import org.dspace.identifier.Identifier;
import org.dspace.identifier.IdentifierException;
import org.dspace.identifier.IdentifierService;
import org.dspace.utils.DSpace;

import java.io.IOException;
import java.sql.SQLException;

/**
 * @author Andrea Schweer [email protected]
 *         for the University of Waikato's Institutional Research Repositories
 */
@Mutative
@Distributive
public class RetrospectivelyRegisterIdentifier extends AbstractCurationTask {
    private static final Logger log = Logger.getLogger(RetrospectivelyRegisterIdentifier.class);

    private int processedItems;
    private int changedItems;
    private boolean hasError;
    private IdentifierService identifierService;
    private Class<? extends Identifier> identifierClass;

    @Override
    public int perform(DSpaceObject dso) throws IOException {
        processedItems = 0;
        changedItems = 0;
        hasError = false;

        identifierService = new DSpace().getSingletonService(IdentifierService.class);
        if (identifierService == null) {
            String message = "Couldn't obtain DOI provider, aborting";
            report(message);
            setResult(message);
            return Curator.CURATE_ERROR;
        }

        String identifierClassProp = taskProperty("identifier.class");
        identifierClass = null;
        if (StringUtils.isNotBlank(identifierClassProp)) {
            try {
                identifierClass = (Class<? extends Identifier>) Class.forName(identifierClassProp);
            } catch (ClassNotFoundException e) {
                String message = "Specified identifier class (" + identifierClassProp + ") not found: " + e.getMessage();
                log.error(message);
                report(message);
                setResult(message);
                return Curator.CURATE_FAIL;
            } catch (ClassCastException e) {
                String message = "Specified identifier class (" + identifierClassProp + ") is invalid (not a subclass of org.dspace.identifier.Identifier): " + e.getMessage();
                log.error(message);
                report(message);
                setResult(message);
                return Curator.CURATE_FAIL;
            }
        }
        if (identifierClass == null) {
            log.info("No identifier class specified, defaulting to DOI");
            identifierClass = DOI.class;
        }

        distribute(dso);

        if (hasError) {
            String message = "Encountered error, check logs / report for details";
            report(message);
            setResult(message);
            return Curator.CURATE_ERROR;
        }
        if (processedItems == 0) {
            String message = "No items processed";
            report(message);
            setResult(message);
            return Curator.CURATE_SKIP;
        }
        if (changedItems > 0) {
            String message = "Processed " + processedItems + " items, changed " + changedItems;
            report(message);
            setResult(message);
            return Curator.CURATE_SUCCESS;
        } else {
            String message = "Processed " + processedItems + " items but changed none";
            report(message);
            setResult(message);
            return Curator.CURATE_FAIL;
        }
    }

    @Override
    protected void performItem(Item item) throws SQLException, IOException {
        processedItems++;

        String existingIdentifier = null;
        try {
            existingIdentifier = identifierService.lookup(Curator.curationContext(), item, identifierClass);
        } catch (Exception e) {
            // ignore any exceptions here
            log.warn("Ignoring exception during identifier lookup: " + e.getMessage(), e);
        }

        if (StringUtils.isBlank(existingIdentifier)) {
            log.info("No existing identifier found for item id=" + item.getID() + ", trying to register for one");
            try {
                identifierService.register(Curator.curationContext(), item);
                log.info("Registered identifier for item id=" + item.getID());
                changedItems++;
            } catch (IdentifierException | AuthorizeException e) {
                String message = "Problem retrospectively registering DOI for item " + item.getID() + ": " + e.getMessage();
                log.error(message, e);
                report(message);
                hasError = true;
            }
        }
    }
}

Originally written for the LERNZ freshwater quality database. Use at own risk.