/*\r
- * Jalview - A Sequence Alignment Editor and Viewer\r
- * Copyright (C) 2007 AM Waterhouse, J Procter, G Barton, M Clamp, S Searle\r
- *\r
+ * Jalview - A Sequence Alignment Editor and Viewer (Version 2.4)\r
+ * Copyright (C) 2008 AM Waterhouse, J Procter, G Barton, M Clamp, S Searle\r
+ * \r
* This program is free software; you can redistribute it and/or\r
* modify it under the terms of the GNU General Public License\r
* as published by the Free Software Foundation; either version 2\r
* of the License, or (at your option) any later version.\r
- *\r
+ * \r
* This program is distributed in the hope that it will be useful,\r
* but WITHOUT ANY WARRANTY; without even the implied warranty of\r
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the\r
* GNU General Public License for more details.\r
- *\r
+ * \r
* You should have received a copy of the GNU General Public License\r
* along with this program; if not, write to the Free Software\r
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA\r
import java.io.*;\r
import java.util.*;\r
\r
+import org.biojava.dasobert.dasregistry.DasSource;\r
import org.exolab.castor.mapping.*;\r
import org.exolab.castor.xml.*;\r
import jalview.analysis.*;\r
import jalview.ws.ebi.EBIFetchClient;\r
\r
/**\r
- * DOCUMENT ME!\r
+ * Implements a runnable for validating a sequence against external databases\r
+ * and then propagating references and features onto the sequence(s)\r
* \r
* @author $author$\r
* @version $Revision$\r
{\r
SequenceI[] dataset;\r
\r
- AlignFrame af;\r
+ IProgressIndicator af;\r
\r
CutAndPasteTransfer output = new CutAndPasteTransfer();\r
\r
}\r
\r
/**\r
- * Creates a new SequenceFeatureFetcher object.\r
+ * Creates a new SequenceFeatureFetcher object and fetches from the\r
+ * currently selected set of databases.\r
* \r
* @param seqs\r
* fetch references for these sequences\r
*/\r
public DBRefFetcher(SequenceI[] seqs, AlignFrame af)\r
{\r
+ this(seqs, af, null);\r
+ }\r
+ /**\r
+ * Creates a new SequenceFeatureFetcher object and fetches from the\r
+ * currently selected set of databases.\r
+ * \r
+ * @param seqs\r
+ * fetch references for these sequences\r
+ * @param af\r
+ * the parent alignframe for progress bar monitoring.\r
+ * @param sources array of database source strings to query references from\r
+ */\r
+ public DBRefFetcher(SequenceI[] seqs, AlignFrame af, String[] sources)\r
+ {\r
this.af = af;\r
SequenceI[] ds = new SequenceI[seqs.length];\r
for (int i = 0; i < seqs.length; i++)\r
ds[i] = seqs[i];\r
}\r
this.dataset = ds;\r
- sfetcher = new SequenceFetcher();\r
- // select appropriate databases based on alignFrame context.\r
- if (af.getViewport().getAlignment().isNucleotide())\r
+ // TODO Jalview 2.5 lots of this code should be in the gui package!\r
+ sfetcher = jalview.gui.SequenceFetcher.getSequenceFetcherSingleton(af);\r
+ if (sources==null)\r
{\r
- dbSources = DBRefSource.DNACODINGDBS;\r
+ // af.featureSettings_actionPerformed(null);\r
+ String[] defdb=null,otherdb = sfetcher.getDbInstances(jalview.ws.dbsources.DasSequenceSource.class);\r
+ Vector selsources = new Vector(), dasselsrc= (af.featureSettings!=null) ? af.featureSettings.getSelectedSources()\r
+ : new jalview.gui.DasSourceBrowser().getSelectedSources();\r
+ Enumeration en = dasselsrc.elements();\r
+ while (en.hasMoreElements())\r
+ {\r
+ DasSource src = (DasSource) en.nextElement();\r
+ selsources.addElement(src.getNickname());\r
+ }\r
+ int osel = 0;\r
+ for (int o=0;otherdb!=null && o<otherdb.length;o++)\r
+ {\r
+ if (!selsources.contains(otherdb[o]))\r
+ {\r
+ otherdb[o] = null;\r
+ } else {\r
+ osel++;\r
+ }\r
+ }\r
+ // select appropriate databases based on alignFrame context.\r
+ if (af.getViewport().getAlignment().isNucleotide())\r
+ {\r
+ defdb = DBRefSource.DNACODINGDBS;\r
+ }\r
+ else\r
+ {\r
+ defdb = DBRefSource.PROTEINDBS;\r
+ }\r
+ // append the selected sequence sources to the default dbs \r
+ dbSources = new String[defdb.length+osel];\r
+ System.arraycopy(defdb, 0, dbSources, 0, defdb.length);\r
+ for (int o=0,op=defdb.length; otherdb!=null && o<otherdb.length; o++)\r
+ {\r
+ if (otherdb[o]!=null)\r
+ {\r
+ dbSources[op++] = otherdb[o];\r
+ }\r
+ }\r
+ } else {\r
+ // we assume the caller knows what they're doing and ensured that all the db source names are valid\r
+ dbSources = sources;\r
}\r
- else\r
+ }\r
+ /**\r
+ * retrieve all the das sequence sources and add them to the list of db sources to retrieve from \r
+ */\r
+ public void appendAllDasSources()\r
+ {\r
+ if (dbSources == null)\r
{\r
- dbSources = DBRefSource.PROTEINDBS;\r
+ dbSources = new String[] {};\r
+ }\r
+ // append additional sources\r
+ String[] otherdb = sfetcher.getDbInstances(jalview.ws.dbsources.DasSequenceSource.class);\r
+ if (otherdb!=null && otherdb.length>0)\r
+ {\r
+ String[] newsrc = new String[dbSources.length+otherdb.length];\r
+ System.arraycopy(dbSources, 0, newsrc,0,dbSources.length);\r
+ System.arraycopy(otherdb, 0, newsrc,dbSources.length, otherdb.length);\r
+ dbSources = newsrc;\r
}\r
}\r
-\r
/**\r
* start the fetcher thread\r
* \r
maxqlen = ((Integer) dbsource.getDbSourceProperties().get(\r
DBRefSource.MULTIACC)).intValue();\r
}\r
+ else\r
+ {\r
+ maxqlen = 1;\r
+ }\r
// iterate through db for each remaining un-verified sequence\r
SequenceI[] currSeqs = new SequenceI[sdataset.size()];\r
sdataset.copyInto(currSeqs);// seqs that are to be validated against\r
{\r
// Still queries to make for current seqIndex\r
StringBuffer queryString = new StringBuffer("");\r
- int nqSize = (maxqlen > queries.size()) ? queries.size()\r
+ int numq=0,nqSize = (maxqlen > queries.size()) ? queries.size()\r
: maxqlen;\r
- for (int nq = 0, numq = 0; nq < nqSize; nq++)\r
+ \r
+ while (queries.size()>0 && numq < nqSize)\r
{\r
- String query = (String) queries.elementAt(nq);\r
+ String query = (String) queries.elementAt(0);\r
if (dbsource.isValidReference(query))\r
{\r
- queryString.append((nq == 0) ? "" : dbsource\r
+ queryString.append((numq == 0) ? "" : dbsource\r
.getAccessionSeparator());\r
queryString.append(query);\r
numq++;\r
}\r
- }\r
- for (int nq = 0; nq < nqSize; nq++)\r
- {\r
+ // remove the extracted query string\r
queries.removeElementAt(0);\r
}\r
// make the queries and process the response\r
AlignmentI retrieved = null;\r
try\r
{\r
+ if (jalview.bin.Cache.log.isDebugEnabled())\r
+ {\r
+ jalview.bin.Cache.log.debug("Querying "+dbsource.getDbName()+" with : '"+queryString.toString()+"'");\r
+ }\r
retrieved = dbsource.getSequenceRecords(queryString.toString());\r
} catch (Exception ex)\r
{\r
ex.printStackTrace();\r
+ } catch (OutOfMemoryError err)\r
+ {\r
+ new OOMWarning("retrieving database references ("\r
+ + queryString.toString() + ")", err);\r
}\r
if (retrieved != null)\r
{\r
{ dbSources[db] }); // jalview.datamodel.DBRefSource.UNIPROT\r
// });\r
// check for existing dbrefs to use\r
- if (uprefs != null)\r
+ if (uprefs != null && uprefs.length>0)\r
{\r
for (int j = 0; j < uprefs.length; j++)\r
{\r
}\r
}\r
}\r
- if (sequenceMatches.size()==0)\r
+ if (sequenceMatches.size() == 0)\r
{\r
- // failed to match directly on accessionId==query so just compare all sequences to entry\r
+ // failed to match directly on accessionId==query so just compare all\r
+ // sequences to entry\r
Enumeration e = seqRefs.keys();\r
while (e.hasMoreElements())\r
{\r
Vector sqs = (Vector) seqRefs.get(e.nextElement());\r
- if (sqs!=null && sqs.size()>0)\r
+ if (sqs != null && sqs.size() > 0)\r
{\r
Enumeration sqe = sqs.elements();\r
while (sqe.hasMoreElements())\r
for (int m = 0; m < sequenceMatches.size(); m++)\r
{\r
sequence = (SequenceI) sequenceMatches.elementAt(m);\r
- // only update start and end positions and shift features if there are no existing references\r
- // TODO: test for legacy where uniprot or EMBL refs exist but no mappings are made (but content matches retrieved set)\r
- boolean updateRefFrame = sequence.getDBRef()==null || sequence.getDBRef().length==0;\r
+ // only update start and end positions and shift features if there are\r
+ // no existing references\r
+ // TODO: test for legacy where uniprot or EMBL refs exist but no\r
+ // mappings are made (but content matches retrieved set)\r
+ boolean updateRefFrame = sequence.getDBRef() == null\r
+ || sequence.getDBRef().length == 0;\r
// verify sequence against the entry sequence\r
\r
String nonGapped = AlignSeq.extractGaps("-. ",\r
+ " SEQUENCE NOT %100 MATCH \n");\r
continue;\r
}\r
- \r
+\r
sbuffer.append(sequence.getName() + " HAS " + absStart\r
- + " PREFIXED RESIDUES COMPARED TO " + dbSource+"\n");\r
+ + " PREFIXED RESIDUES COMPARED TO " + dbSource + "\n");\r
//\r
- // + " - ANY SEQUENCE FEATURES"\r
- // + " HAVE BEEN ADJUSTED ACCORDINGLY \n");\r
+ // + " - ANY SEQUENCE FEATURES"\r
+ // + " HAVE BEEN ADJUSTED ACCORDINGLY \n");\r
// absStart = 0;\r
// create valid mapping between matching region of local sequence and\r
// the mapped sequence\r
mp = new Mapping(null, new int[]\r
- { sequence.getStart()+absStart, sequence.getStart()+absStart+entrySeq.length()-1 }, new int[]\r
- { entry.getStart(),\r
- entry.getStart() + entrySeq.length() - 1 }, 1, 1);\r
- updateRefFrame=false; // mapping is based on current start/end so don't modify start and end\r
+ { sequence.getStart() + absStart,\r
+ sequence.getStart() + absStart + entrySeq.length() - 1 },\r
+ new int[]\r
+ { entry.getStart(),\r
+ entry.getStart() + entrySeq.length() - 1 }, 1, 1);\r
+ updateRefFrame = false; // mapping is based on current start/end so\r
+ // don't modify start and end\r
}\r
else\r
{\r
SequenceFeature[] sf = sequence.getSequenceFeatures();\r
int start = sequence.getStart();\r
int end = sequence.getEnd();\r
- int startShift = 1-absStart-start; // how much the features are to be shifted by\r
+ int startShift = 1 - absStart - start; // how much the features are\r
+ // to be shifted by\r
for (int sfi = 0; sfi < sf.length; sfi++)\r
{\r
if (sf[sfi].getBegin() >= start && sf[sfi].getEnd() <= end)\r
git://source.jalview.org / jalview.git / blobdiff