src/jalview/io/BSMLFile.java

   1 /*
   2  * Jalview - A Sequence Alignment Editor and Viewer ($$Version-Rel$$)
   3  * Copyright (C) $$Year-Rel$$ The Jalview Authors
   4  *
   5  * This file is part of Jalview.
   6  *
   7  * Jalview is free software: you can redistribute it and/or
   8  * modify it under the terms of the GNU General Public License
   9  * as published by the Free Software Foundation, either version 3
  10  * of the License, or (at your option) any later version.
  11  *
  12  * Jalview is distributed in the hope that it will be useful, but
  13  * WITHOUT ANY WARRANTY; without even the implied warranty
  14  * of MERCHANTABILITY or FITNESS FOR A PARTICULAR
  15  * PURPOSE.  See the GNU General Public License for more details.
  16  *
  17  * You should have received a copy of the GNU General Public License
  18  * along with Jalview.  If not, see <http://www.gnu.org/licenses/>.
  19  * The Jalview Authors are detailed in the 'AUTHORS' file.
  20  */
  21 package jalview.io;
  22
  23 import jalview.datamodel.Sequence;
  24 import jalview.datamodel.SequenceFeature;
  25 import jalview.datamodel.SequenceI;
  26 import jalview.util.MessageManager;
  27
  28 import java.io.BufferedReader;
  29 import java.io.FileNotFoundException;
  30 import java.io.FileReader;
  31 import java.io.IOException;
  32 import java.util.Hashtable;
  33 import java.util.Map;
  34
  35 import javax.xml.parsers.DocumentBuilder;
  36 import javax.xml.parsers.DocumentBuilderFactory;
  37 import javax.xml.parsers.ParserConfigurationException;
  38
  39 import org.w3c.dom.Document;
  40 import org.w3c.dom.Element;
  41 import org.w3c.dom.NodeList;
  42 import org.xml.sax.InputSource;
  43 import org.xml.sax.SAXException;
  44
  45 import fr.orsay.lri.varna.exceptions.ExceptionFileFormatOrSyntax;
  46 import fr.orsay.lri.varna.exceptions.ExceptionLoadingFailed;
  47 import fr.orsay.lri.varna.exceptions.ExceptionPermissionDenied;
  48
  49 /**
  50  * Preliminary reader for Bioinformatics Sequence Markup Language
  51  * http://www.bsml.org
  52  *
  53  * @author hansonr
  54  *
  55  */
  56 public class BSMLFile extends AlignFile
  57 {
  58
  59   public BSMLFile()
  60   {
  61     super();
  62
  63   }
  64
  65   public BSMLFile(String inFile, DataSourceType type) throws IOException
  66   {
  67     super(inFile, type);
  68
  69   }
  70
  71   public BSMLFile(FileParse source) throws IOException
  72   {
  73     super(source);
  74
  75   }
  76
  77   public BufferedReader CreateReader() throws FileNotFoundException
  78   {
  79     FileReader fr = null;
  80     fr = new FileReader(inFile);
  81
  82     BufferedReader r = new BufferedReader(fr);
  83     return r;
  84   }
  85
  86   /*
  87    * (non-Javadoc)
  88    *
  89    * @see jalview.io.AlignFile#parse()
  90    */
  91   @Override
  92   public void parse() throws IOException
  93   {
  94     try
  95     {
  96       _parse();
  97     } catch (ExceptionPermissionDenied pdx)
  98     {
  99       errormessage = MessageManager.formatMessage(
 100               "exception.BSML_couldnt_access_datasource", new String[]
 101               { pdx.getMessage() });
 102       throw new IOException(pdx);
 103     } catch (ExceptionLoadingFailed lf)
 104     {
 105       errormessage = MessageManager.formatMessage(
 106               "exception.BSML_couldnt_process_data", new String[]
 107               { lf.getMessage() });
 108       throw new IOException(lf);
 109     } catch (ExceptionFileFormatOrSyntax iff)
 110     {
 111       errormessage = MessageManager
 112               .formatMessage("exception.BSML_invalid_file", new String[]
 113               { iff.getMessage() });
 114       throw new IOException(iff);
 115     } catch (Exception x)
 116     {
 117       error = true;
 118       errormessage = MessageManager.formatMessage(
 119               "exception.BSML_problem_parsing_data", new String[]
 120               { x.getMessage() });
 121       throw new IOException(errormessage, x);
 122     }
 123   }
 124
 125   @SuppressWarnings("unchecked")
 126   public void _parse()
 127           throws ExceptionPermissionDenied, ExceptionLoadingFailed,
 128           ExceptionFileFormatOrSyntax, ParserConfigurationException,
 129           SAXException, IOException
 130   {
 131
 132     DocumentBuilderFactory dbf = DocumentBuilderFactory.newInstance();
 133
 134     dbf.setIgnoringElementContentWhitespace(true);
 135     dbf.setIgnoringComments(true);
 136     dbf.setValidating(true);
 137     dbf.setCoalescing(true);
 138     dbf.setNamespaceAware(true);
 139       dbf.setFeature("http://xml.org/sax/features/namespaces", false);
 140       dbf.setFeature("http://xml.org/sax/features/validation", false);
 141       dbf.setFeature(
 142               "http://apache.org/xml/features/nonvalidating/load-dtd-grammar",
 143               false);
 144       dbf.setFeature(
 145               "http://apache.org/xml/features/nonvalidating/load-external-dtd",
 146               false);
 147
 148       DocumentBuilder db = dbf.newDocumentBuilder();
 149
 150       Map<String, SequenceI> htSeq = new Hashtable<>();
 151       InputSource is = new InputSource(getReader());
 152       Document d = db.parse(is);
 153       NodeList sequences = d.getElementsByTagName("Sequence-data");
 154       int n = sequences.getLength();
 155       SequenceI[] sqs = new SequenceI[n];
 156       for (int i = 0; i < n; i++)
 157       {
 158         Element e = (Element) sequences.item(i);
 159         String s = e.getTextContent();
 160         String id = e.getAttribute("seq-name");
 161         SequenceI seq = sqs[i] = new Sequence(id, s, 1, s.length());
 162         htSeq.put(id, seq);
 163         // ?? sqs[i].setEnd(sqs[i].findPosition(sqs[i].getLength()));
 164       }
 165
 166       sequences = d.getElementsByTagName("Sequence");
 167       n = sequences.getLength();
 168       for (int i = 0; i < n; i++)
 169       {
 170         Element e = (Element) sequences.item(i);
 171         String mol = e.getAttribute("molecule"); // dna or rna
 172         if (!"dna".equals(mol))
 173         {
 174           System.err.println("BSML molecule=rna not implemented");
 175           continue;
 176         }
 177         String title = e.getAttribute("title");
 178         SequenceI seq = htSeq.get(title);
 179         if (seq == null)
 180         {
 181           continue;
 182         }
 183       NodeList features = e.getElementsByTagName("Feature");
 184         int featureCount = features.getLength();
 185         for (int f = 0; f < featureCount; f++)
 186         {
 187           Element feature = (Element) features.item(f);
 188           // <Feature class="GENE" title="CPXV-GER_1980_EP4-211">
 189           // <Interval-loc complement="0" endpos="217104" startpos="216643"/>
 190           // <Resource id="GENE-ID:119705"/>
 191           // </Feature>
 192         Element iloc = (Element) feature
 193                 .getElementsByTagName("Interval-loc").item(0);
 194           String complement = iloc.getAttribute("complement");
 195           if (!"0".equals(complement))
 196           {
 197             // Jalview cannot handle complement genes (running backward on the
 198             // complementary strand);
 199             continue;
 200           }
 201           String fclass = feature.getAttribute("class");
 202           if (!"GENE".equals(fclass))
 203           {
 204             // just processing GENE features for now;
 205             continue;
 206           }
 207           String ftitle = feature.getAttribute("title");
 208         int start = Integer.parseInt(iloc.getAttribute("startpos"));
 209         int end = Integer.parseInt(iloc.getAttribute("endpos"));
 210         SequenceFeature sf = new SequenceFeature("GENE", ftitle, start, end,
 211                 null);
 212         seq.addSequenceFeature(sf);
 213         }
 214         setSeqs(sqs);
 215       }
 216
 217   }
 218
 219   @Override
 220   public String print(SequenceI[] s, boolean jvSuffix)
 221   {
 222     return "not yet implemented";
 223   }
 224
 225 }