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