package eu.etaxonomy.cdm.api.service;
import java.util.ArrayList;
+import java.util.Arrays;
+import java.util.HashSet;
import java.util.List;
+import java.util.Set;
import org.apache.log4j.Logger;
import org.hibernate.criterion.Criterion;
+import org.springframework.beans.factory.annotation.Autowired;
+import org.springframework.transaction.annotation.Propagation;
import org.springframework.transaction.annotation.Transactional;
import eu.etaxonomy.cdm.api.service.config.IIdentifiableEntityServiceConfigurator;
import eu.etaxonomy.cdm.model.common.LSID;
import eu.etaxonomy.cdm.model.common.UuidAndTitleCache;
import eu.etaxonomy.cdm.model.media.Rights;
-import eu.etaxonomy.cdm.model.reference.ReferenceBase;
+import eu.etaxonomy.cdm.model.reference.Reference;
import eu.etaxonomy.cdm.model.reference.ReferenceFactory;
import eu.etaxonomy.cdm.persistence.dao.common.IIdentifiableDao;
import eu.etaxonomy.cdm.persistence.query.MatchMode;
import eu.etaxonomy.cdm.persistence.query.OrderHint;
+import eu.etaxonomy.cdm.persistence.query.OrderHint.SortOrder;
import eu.etaxonomy.cdm.strategy.cache.common.IIdentifiableEntityCacheStrategy;
+import eu.etaxonomy.cdm.strategy.match.DefaultMatchStrategy;
+import eu.etaxonomy.cdm.strategy.match.IMatchStrategy;
+import eu.etaxonomy.cdm.strategy.match.IMatchable;
+import eu.etaxonomy.cdm.strategy.match.MatchException;
+import eu.etaxonomy.cdm.strategy.merge.IMergable;
+import eu.etaxonomy.cdm.strategy.merge.IMergeStrategy;
+import eu.etaxonomy.cdm.strategy.merge.MergeException;
public abstract class IdentifiableServiceBase<T extends IdentifiableEntity,DAO extends IIdentifiableDao<T>> extends AnnotatableServiceBase<T,DAO>
implements IIdentifiableEntityService<T>{
+ @Autowired
+ protected ICommonService commonService;
+
+
protected static final int UPDATE_TITLE_CACHE_DEFAULT_STEP_SIZE = 1000;
protected static final Logger logger = Logger.getLogger(IdentifiableServiceBase.class);
return new DefaultPagerImpl<T>(pageNumber, numberOfResults, pageSize, results);\r
}
+ @Transactional(readOnly = true)
+ public List<T> listByTitle(Class<? extends T> clazz, String queryString,MatchMode matchmode, List<Criterion> criteria, Integer pageSize, Integer pageNumber, List<OrderHint> orderHints, List<String> propertyPaths) {
+ Integer numberOfResults = dao.countByTitle(clazz, queryString, matchmode, criteria);
+
+ List<T> results = new ArrayList<T>();
+ if(numberOfResults > 0) { // no point checking again
+ results = dao.findByTitle(clazz, queryString, matchmode, criteria, pageSize, pageNumber, orderHints, propertyPaths);
+ }
+ return results;
+ }
+
+ @Transactional(readOnly = true)
+ public List<T> listByReferenceTitle(Class<? extends T> clazz, String queryString,MatchMode matchmode, List<Criterion> criteria, Integer pageSize, Integer pageNumber, List<OrderHint> orderHints, List<String> propertyPaths) {
+ Integer numberOfResults = dao.countByReferenceTitle(clazz, queryString, matchmode, criteria);
+
+ List<T> results = new ArrayList<T>();
+ if(numberOfResults > 0) { // no point checking again
+ results = dao.findByReferenceTitle(clazz, queryString, matchmode, criteria, pageSize, pageNumber, orderHints, propertyPaths);
+ }
+ return results;
+ }
+
@Transactional(readOnly = true)
public T find(LSID lsid) {
return dao.find(lsid);
if (entityCacheStrategy == null){
entityCacheStrategy = entity.getCacheStrategy();
//FIXME find out why the wrong cache strategy is loaded here, see #1876
- if (entity instanceof ReferenceBase){
- entityCacheStrategy = ReferenceFactory.newReference(((ReferenceBase)entity).getType()).getCacheStrategy();
+ if (entity instanceof Reference){
+ entityCacheStrategy = ReferenceFactory.newReference(((Reference)entity).getType()).getCacheStrategy();
}
}
entity.setCacheStrategy(entityCacheStrategy);
- //entity.setTitleCache(null);
- setOtherCachesNull(entity); //TODO find better solution
+ //TODO this won't work for those classes that always generate the title cache new
String titleCache = entity.getTitleCache();
+ setOtherCachesNull(entity); //TODO find better solution
String newTitleCache = entityCacheStrategy.getTitleCache(entity);
if (titleCache == null || titleCache != null && ! titleCache.equals(newTitleCache)){
entity.setTitleCache(null, false);
}
}
}
- saveOrUpdateAll(entitiesToUpdate);
+ saveOrUpdate(entitiesToUpdate);
}
}
+
+
/**
* Needs override if not only the title cache should be set to null to
protected void setOtherCachesNull(T entity) {
return;
}
+
+
+
+ private class DeduplicateState{
+ String lastTitleCache;
+ Integer pageSize = 50;
+ int nPages = 3;
+ int startPage = 0;
+ boolean isCompleted = false;
+ int result;
+ }
+
+ /* (non-Javadoc)
+ * @see eu.etaxonomy.cdm.api.service.IIdentifiableEntityService#deduplicate(java.lang.Class, eu.etaxonomy.cdm.strategy.match.IMatchStrategy, eu.etaxonomy.cdm.strategy.merge.IMergeStrategy)
+ */
+ @Override
+ @Transactional(propagation = Propagation.SUPPORTS, readOnly = false)
+ public int deduplicate(Class<? extends T> clazz, IMatchStrategy matchStrategy, IMergeStrategy mergeStrategy) {
+ DeduplicateState dedupState = new DeduplicateState();
+
+ if (clazz == null){
+ logger.warn("Deduplication clazz must not be null!");
+ return 0;
+ }
+ if (! ( IMatchable.class.isAssignableFrom(clazz) && IMergable.class.isAssignableFrom(clazz) ) ){
+ logger.warn("Deduplication implemented only for classes implementing IMatchable and IMergeable. No deduplication performed!");
+ return 0;
+ }
+ Class matchableClass = clazz;
+ if (matchStrategy == null){
+ matchStrategy = DefaultMatchStrategy.NewInstance(matchableClass);
+ }
+ List<T> nextGroup = new ArrayList<T>();
+
+ int result = 0;
+// double countTotal = count(clazz);
+//
+// Number countPagesN = Math.ceil(countTotal/dedupState.pageSize.doubleValue()) ;
+// int countPages = countPagesN.intValue();
+//
+
+ List<OrderHint> orderHints = Arrays.asList(new OrderHint[]{new OrderHint("titleCache", SortOrder.ASCENDING)});
+
+ while (! dedupState.isCompleted){
+ //get x page sizes
+ List<T> objectList = getPages(clazz, dedupState, orderHints);
+ //after each page check if any changes took place
+ int nUnEqualPages = handleAllPages(objectList, dedupState, nextGroup, matchStrategy, mergeStrategy);
+ nUnEqualPages = nUnEqualPages + dedupState.pageSize * dedupState.startPage;
+ //refresh start page counter
+ int finishedPages = nUnEqualPages / dedupState.pageSize;
+ dedupState.startPage = finishedPages;
+ }
+
+ result += handleLastGroup(nextGroup, matchStrategy, mergeStrategy);
+ return result;
+ }
+
+
+ private int handleAllPages(List<T> objectList, DeduplicateState dedupState, List<T> nextGroup, IMatchStrategy matchStrategy, IMergeStrategy mergeStrategy) {
+ int nUnEqual = 0;
+ for (T object : objectList){
+ String currentTitleCache = object.getTitleCache();
+ if (currentTitleCache != null && currentTitleCache.equals(dedupState.lastTitleCache)){
+ //=titleCache
+ nextGroup.add(object);
+ }else{
+ //<> titleCache
+ dedupState.result += handleLastGroup(nextGroup, matchStrategy, mergeStrategy);
+ nextGroup = new ArrayList<T>();
+ nextGroup.add(object);
+ nUnEqual++;
+ }
+ dedupState.lastTitleCache = currentTitleCache;
+ }
+ handleLastGroup(nextGroup, matchStrategy, mergeStrategy);
+ return nUnEqual;
+ }
+
+ private List<T> getPages(Class<? extends T> clazz, DeduplicateState dedupState, List<OrderHint> orderHints) {
+ List<T> result = new ArrayList<T>();
+ for (int pageNo = dedupState.startPage; pageNo < dedupState.startPage + dedupState.nPages; pageNo++){
+ List<T> objectList = listByTitle(clazz, null, null, null, dedupState.pageSize, pageNo, orderHints, null);
+ result.addAll(objectList);
+ }
+ if (result.size()< dedupState.nPages * dedupState.pageSize ){
+ dedupState.isCompleted = true;
+ }
+ return result;
+ }
+
+ private int handleLastGroup(List<T> group, IMatchStrategy matchStrategy, IMergeStrategy mergeStrategy) {
+ int result = 0;
+ int size = group.size();
+ Set<Integer> exclude = new HashSet<Integer>(); //set to collect all objects, that have been merged already
+ for (int i = 0; i < size - 1; i++){
+ if (exclude.contains(i)){
+ continue;
+ }
+ for (int j = i + 1; j < size; j++){
+ if (exclude.contains(j)){
+ continue;
+ }
+ T firstObject = group.get(i);
+ T secondObject = group.get(j);
+
+ try {
+ if (matchStrategy.invoke((IMatchable)firstObject, (IMatchable)secondObject)){
+ commonService.merge((IMergable)firstObject, (IMergable)secondObject, mergeStrategy);
+ exclude.add(j);
+ result++;
+ }
+ } catch (MatchException e) {
+ logger.warn("MatchException when trying to match " + firstObject.getTitleCache());
+ e.printStackTrace();
+ } catch (MergeException e) {
+ logger.warn("MergeException when trying to merge " + firstObject.getTitleCache());
+ e.printStackTrace();
+ }
+ }
+ }
+ return result;
+ }
+
}\r