Changeset 870


Ignore:
Timestamp:
Aug 3, 2012, 8:35:37 PM (4 years ago)
Author:
hailiang.mei@…
Message:

fixed the inconsistency between SP release file and its species list.

File:
1 edited

Legend:

Unmodified
Added
Removed
  • trunk/code/conceptwiki/imports/imports-swissprot/src/main/java/nl/nbic/conceptwiki/imports/swissprot/SwissProtReader.java

    r869 r870  
    422422            if (proteinRecommendedName.isEmpty()) {
    423423                logger.error("proteinRecommendedName is empty: " + entry.getAccession().get(0));
    424             }                               
     424            }   
     425           
     426            logger.info("primary AC: " + entry.getAccession().get(0));
     427            logger.info("proteinRecommendedName: " +proteinRecommendedName);
     428           
    425429            if (!bootstrapGenericProteinTags.containsKey(proteinRecommendedName)) {
    426430                bootstrapGenericProteinTags.put(proteinRecommendedName, new BootstrapConceptEnvelopImpl(genericProteinConcept, CONCEPTWIKI_BRANCH));
     
    524528     */
    525529    private LinkedHashMultiset<Concept> getSpeciesConcepts() throws IOException {
     530       
     531        allSpeciesConcepts.clear();
    526532       
    527533        logger.info("enter SwissProtReader.getSpeciesConcepts()");
     
    582588        // name = Aedes albopictus densovirus (isolate Boublik/1994)
    583589        final String code = line.substring(0, line.indexOf(' '));
    584         final String scientificName = line.substring(line.indexOf(SPECIES_NAME_PREFIX) + SPECIES_NAME_PREFIX.length()).trim();
     590        String scientificName = line.substring(line.indexOf(SPECIES_NAME_PREFIX) + SPECIES_NAME_PREFIX.length()).trim();       
     591       
     592        // Some species name is ended with a ".". This is not consistent with the SwissProt release XML file. Thus the "." should be skipped.
     593        if (scientificName.endsWith(".")) {
     594                scientificName = scientificName.substring(0, scientificName.length()-1);
     595        }
    585596       
    586597        labels.add(new LabelImpl(LabelType.PREFERRED, scientificName, LANG_EN));
     
    630641        while (xsr.nextTag() == XMLStreamConstants.START_ELEMENT) {
    631642            final Entry entry = (Entry)unmarshaller.unmarshal(xsr);
    632             updatedTriples.addAll(getTriplesFromEntry(entry));           
     643            final LinkedHashMultiset<Triple> triples = getTriplesFromEntry(entry);
     644           
     645            if (triples != null) {
     646                updatedTriples.addAll(getTriplesFromEntry(entry));
     647            }           
    633648            if (counter % CHUNK_SIZE == 0) {
    634649                logger.info(Integer.toString(counter));
Note: See TracChangeset for help on using the changeset viewer.