cdmlib-io/src/main/java/eu/etaxonomy/cdm/io/dwca/in/DwcTaxonCsv2CdmTaxonRelationConverter.java

   1 // $Id$
   2 /**
   3 * Copyright (C) 2009 EDIT
   4 * European Distributed Institute of Taxonomy
   5 * http://www.e-taxonomy.eu
   6 *
   7 * The contents of this file are subject to the Mozilla Public License Version 1.1
   8 * See LICENSE.TXT at the top of this package for the full license terms.
   9 */
  10 package eu.etaxonomy.cdm.io.dwca.in;
  11
  12 import java.util.ArrayList;
  13 import java.util.HashMap;
  14 import java.util.HashSet;
  15 import java.util.List;
  16 import java.util.Map;
  17 import java.util.Set;
  18
  19 import org.apache.log4j.Logger;
  20
  21 import eu.etaxonomy.cdm.common.CdmUtils;
  22 import eu.etaxonomy.cdm.io.dwca.TermUri;
  23 import eu.etaxonomy.cdm.model.common.CdmBase;
  24 import eu.etaxonomy.cdm.model.reference.Reference;
  25 import eu.etaxonomy.cdm.model.taxon.Classification;
  26 import eu.etaxonomy.cdm.model.taxon.Synonym;
  27 import eu.etaxonomy.cdm.model.taxon.SynonymRelationshipType;
  28 import eu.etaxonomy.cdm.model.taxon.Taxon;
  29 import eu.etaxonomy.cdm.model.taxon.TaxonBase;
  30
  31 /**
  32  * @author a.mueller
  33  * @date 23.11.2011
  34  *
  35  */
  36 public class DwcTaxonCsv2CdmTaxonRelationConverter<STATE extends DwcaImportState> extends PartitionableConverterBase<DwcaImportState>
  37                                                 implements IPartitionableConverter<CsvStreamItem, INamespaceReader<CdmBase>, String>{
  38         private static Logger logger = Logger.getLogger(DwcTaxonCsv2CdmTaxonRelationConverter.class);
  39
  40         private static final String ID = "id";
  41
  42         /**
  43          * @param state
  44          */
  45         public DwcTaxonCsv2CdmTaxonRelationConverter(DwcaImportState state) {
  46                 super();
  47                 this.state = state;
  48         }
  49
  50
  51         public IReader<MappedCdmBase> map(CsvStreamItem item){
  52                 List<MappedCdmBase> resultList = new ArrayList<MappedCdmBase>();
  53
  54                 Map<String, String> csvRecord = item.map;
  55                 Reference<?> sourceReference = null;
  56                 String sourceReferecenDetail = null;
  57
  58                 String id = csvRecord.get(ID);
  59                 TaxonBase<?> taxonBase = getTaxonBase(id, item, null);
  60                 if (taxonBase == null){
  61                         String warning = "Taxon not available for id %s.";
  62                         warning = String.format(warning, id);
  63                         fireWarningEvent(warning, item, 8);
  64                 }else{
  65
  66                         MappedCdmBase mcb = new MappedCdmBase(taxonBase);
  67                         resultList.add(mcb);
  68
  69                         handleAcceptedNameUsage(item, state, taxonBase, id);
  70
  71                         handleParentNameUsage(item, state, taxonBase, resultList);
  72
  73                         handleKingdom(item, state);
  74
  75                         handlePhylum(item, state);
  76
  77                         handleClass(item, state);
  78
  79                         handleOrder(item, state);
  80
  81                         handleFamily(item, state);
  82
  83                         handleGenus(item, state);
  84
  85                         handleSubGenus(item, state);
  86
  87                 }
  88                 csvRecord.remove(ID);
  89
  90
  91 //                  <!-- Top level group; listed as kingdom but may be interpreted as domain or superkingdom
  92 //                       The following eight groups are recognized: Animalia, Archaea, Bacteria, Chromista,
  93 //                       Fungi, Plantae, Protozoa, Viruses -->
  94 //                  <field index='10' term='http://rs.tdwg.org/dwc/terms/kingdom'/>
  95
  96 //                  <!-- Specific epithet; for hybrids, the multiplication symbol is included in the epithet -->
  97 //                  <field index='17' term='http://rs.tdwg.org/dwc/terms/specificEpithet'/>
  98
  99 //                  <!-- Infraspecific epithet -->
 100 //                  <field index='18' term='http://rs.tdwg.org/dwc/terms/infraspecificEpithet'/>
 101
 102 //                      <!-- Acceptance status published in -->
 103 //                  <field index='20' term='http://purl.org/dc/terms/source'/>
 104
 105 //                  <!-- Reference in which the scientific name was first published -->
 106 //                  <field index='21' term='http://rs.tdwg.org/dwc/terms/namePublishedIn'/>
 107
 108 //                  <!-- Scrutiny date -->
 109 //                  <field index='23' term='http://purl.org/dc/terms/modified'/>
 110 //                  <!-- Additional data for the taxon -->
 111
 112 //                  <field index='24' term='http://purl.org/dc/terms/description'/>
 113 //                  </core>
 114
 115                 return new ListReader<MappedCdmBase>(resultList);
 116         }
 117
 118
 119         @Override
 120         public String getSourceId(CsvStreamItem item) {
 121                 String id = item.get(ID);
 122                 return id;
 123         }
 124
 125
 126         private void handleSubGenus(CsvStreamItem item, DwcaImportState state) {
 127                 // TODO Auto-generated method stub
 128
 129         }
 130
 131
 132         private void handleGenus(CsvStreamItem item, DwcaImportState state) {
 133                 // TODO Auto-generated method stub
 134
 135         }
 136
 137
 138         private void handleFamily(CsvStreamItem item, DwcaImportState state) {
 139                 // TODO Auto-generated method stub
 140
 141         }
 142
 143
 144         private void handleOrder(CsvStreamItem item, DwcaImportState state) {
 145                 // TODO Auto-generated method stub
 146
 147         }
 148
 149
 150         private void handleClass(CsvStreamItem item, DwcaImportState state) {
 151                 // TODO Auto-generated method stub
 152
 153         }
 154
 155
 156         private void handlePhylum(CsvStreamItem item, DwcaImportState state) {
 157                 // TODO Auto-generated method stub
 158
 159         }
 160
 161
 162         private void handleKingdom(CsvStreamItem item, DwcaImportState state) {
 163                 // TODO Auto-generated method stub
 164
 165         }
 166
 167
 168         private void handleParentNameUsage(CsvStreamItem item, DwcaImportState state, TaxonBase<?> taxonBase, List<MappedCdmBase> resultList) {
 169                 if (exists(TermUri.DWC_PARENT_NAME_USAGE_ID, item) || exists(TermUri.DWC_PARENT_NAME_USAGE, item)){
 170                         if (taxonBase.isInstanceOf(Taxon.class)){
 171                                 Taxon taxon = CdmBase.deproxy(taxonBase, Taxon.class);
 172                                 String accId = item.get(TermUri.DWC_PARENT_NAME_USAGE_ID);
 173                                 Taxon parentTaxon = getTaxonBase(accId, item, Taxon.class);
 174                                 if (parentTaxon == null){
 175                                                 fireWarningEvent("NON-ID parent Name Usage not yet implemented or parent name usage id not available", item, 4);
 176                                 }else{
 177                                         Classification classification = getClassification(item);
 178                                         Reference<?> citation = null;
 179                                         classification.addParentChild(parentTaxon, taxon, citation, null);
 180                                         resultList.add(new MappedCdmBase(classification));
 181                                 }
 182                         }else{
 183                                 String message = "PARENT_NAME_USAGE given for Synonym. This is not allowed in CDM.";
 184                                 //TODO check "is this Taxon"
 185                                 fireWarningEvent(message, item, 4);
 186                         }
 187                 }
 188
 189
 190         }
 191
 192
 193         private Classification getClassification(CsvStreamItem item) {
 194                 Set<Classification> result = new HashSet<Classification>();
 195                 String datasetKey = item.get(TermUri.DWC_DATASET_ID);
 196                 if (CdmUtils.areBlank(datasetKey,item.get(TermUri.DWC_DATASET_NAME))){
 197                         datasetKey = DwcTaxonCsv2CdmTaxonConverter.NO_DATASET;
 198                 }
 199
 200                 result.addAll(state.get(TermUri.DWC_DATASET_ID.toString(), datasetKey, Classification.class));
 201                 result.addAll(state.get(TermUri.DWC_DATASET_NAME.toString(), item.get(TermUri.DWC_DATASET_NAME), Classification.class));
 202                 if (result.isEmpty()){
 203                         return null;
 204                 }else if (result.size() > 1){
 205                         fireWarningEvent("Dataset is ambigous. I take arbitrary one.", item, 8);
 206                 }
 207                 return result.iterator().next();
 208         }
 209
 210
 211         private void handleAcceptedNameUsage(CsvStreamItem item, DwcaImportState state, TaxonBase taxonBase, String id) {
 212                 if (exists(TermUri.DWC_ACCEPTED_NAME_USAGE_ID, item) || exists(TermUri.DWC_ACCEPTED_NAME_USAGE, item)){
 213                         String accId = item.get(TermUri.DWC_ACCEPTED_NAME_USAGE_ID);
 214                         if (id.equals(accId)){
 215                                 return;   //mapping to itself needs no further handling
 216                         }
 217                         if (taxonBase.isInstanceOf(Synonym.class)){
 218                                 Synonym synonym = CdmBase.deproxy(taxonBase, Synonym.class);
 219                                 Taxon accTaxon = getTaxonBase(accId, item, Taxon.class);
 220                                 if (accTaxon == null){
 221                                                 fireWarningEvent("NON-ID accepted Name Usage not yet implemented or taxon for name usage id not available", item, 4);
 222                                 }else{
 223                                         accTaxon.addSynonym(synonym, SynonymRelationshipType.SYNONYM_OF(),null, null);
 224                                 }
 225                         } else{
 226                                 String message = "Accepted name usage is not of type synonym. This is not allowed in CDM. Can't create realtionship";
 227                                 //TODO check "is this Taxon"
 228                                 fireWarningEvent(message, item, 4);
 229                         }
 230                 }else{
 231                         if (logger.isDebugEnabled()){logger.debug("");}
 232                 }
 233         }
 234
 235
 236         private <T extends TaxonBase> T getTaxonBase(String id, CsvStreamItem item, Class<T> clazz) {
 237                 if (clazz == null){
 238                         clazz = (Class)TaxonBase.class;
 239                 }
 240                 List<T> taxonList = state.get(TermUri.DWC_TAXON.toString(), id, clazz);
 241                 if (taxonList.size() > 1){
 242                         String message = "Undefined taxon mapping for id %s.";
 243                         message = String.format(message, id);
 244                         fireWarningEvent(message, item, 8);
 245                         logger.warn(message);  //TODO remove when events are handled correctly
 246                         return null;
 247                 }else if (taxonList.isEmpty()){
 248                         return null;
 249                 }else{
 250                         return taxonList.get(0);
 251                 }
 252         }
 253
 254
 255 //**************************** PARTITIONABLE ************************************************
 256
 257
 258
 259
 260         protected void makeForeignKeysForItem(CsvStreamItem item, Map<String, Set<String>> fkMap){
 261                 //do nothing, their are no foreign keys yet to handle
 262         }
 263
 264 //************************************* TO STRING ********************************************
 265
 266         @Override
 267         public String toString(){
 268                 return this.getClass().getName();
 269         }
 270
 271
 272 }