X-Git-Url: http://source.jalview.org/gitweb/?a=blobdiff_plain;f=src%2Fjalview%2Fdatamodel%2FSequence.java;h=24f904cf8f00bbe72c62c62a3ccad57e3ee66fb8;hb=ba61fc7a3397ce6b57640e0b19c76392a983768a;hp=b0faf21132bf9cdaf21ec9be246833e5cfd2f4d0;hpb=dfa35cc477b0d3af4a3cd25472cb6126626c67d3;p=jalview.git diff --git a/src/jalview/datamodel/Sequence.java b/src/jalview/datamodel/Sequence.java index b0faf21..24f904c 100755 --- a/src/jalview/datamodel/Sequence.java +++ b/src/jalview/datamodel/Sequence.java @@ -22,6 +22,8 @@ package jalview.datamodel; import jalview.analysis.AlignSeq; import jalview.api.DBRefEntryI; +import jalview.datamodel.features.SequenceFeatures; +import jalview.datamodel.features.SequenceFeaturesI; import jalview.util.Comparison; import jalview.util.DBRefUtils; import jalview.util.MapList; @@ -34,6 +36,8 @@ import java.util.Enumeration; import java.util.List; import java.util.Vector; +import com.stevesoft.pat.Regex; + import fr.orsay.lri.varna.models.rna.RNA; /** @@ -45,6 +49,11 @@ import fr.orsay.lri.varna.models.rna.RNA; */ public class Sequence extends ASequence implements SequenceI { + private static final Regex limitrx = new Regex( + "[/][0-9]{1,}[-][0-9]{1,}$"); + + private static final Regex endrx = new Regex("[0-9]{1,}$"); + SequenceI datasetSequence; String name; @@ -78,8 +87,22 @@ public class Sequence extends ASequence implements SequenceI */ int index = -1; - /** array of sequence features - may not be null for a valid sequence object */ - public SequenceFeature[] sequenceFeatures; + private SequenceFeatures sequenceFeatureStore; + + /* + * A cursor holding the approximate current view position to the sequence, + * as determined by findIndex or findPosition or findPositions. + * Using a cursor as a hint allows these methods to be more performant for + * large sequences. + */ + private SequenceCursor cursor; + + /* + * A number that should be incremented whenever the sequence is edited. + * If the value matches the cursor token, then we can trust the cursor, + * if not then it should be recomputed. + */ + private int changeCount; /** * Creates a new Sequence object. @@ -96,11 +119,13 @@ public class Sequence extends ASequence implements SequenceI */ public Sequence(String name, String sequence, int start, int end) { + this(); initSeqAndName(name, sequence.toCharArray(), start, end); } public Sequence(String name, char[] sequence, int start, int end) { + this(); initSeqAndName(name, sequence, start, end); } @@ -124,11 +149,6 @@ public class Sequence extends ASequence implements SequenceI checkValidRange(); } - com.stevesoft.pat.Regex limitrx = new com.stevesoft.pat.Regex( - "[/][0-9]{1,}[-][0-9]{1,}$"); - - com.stevesoft.pat.Regex endrx = new com.stevesoft.pat.Regex("[0-9]{1,}$"); - void parseId() { if (name == null) @@ -175,6 +195,14 @@ public class Sequence extends ASequence implements SequenceI } /** + * default constructor + */ + private Sequence() + { + sequenceFeatureStore = new SequenceFeatures(); + } + + /** * Creates a new Sequence object. * * @param name @@ -213,8 +241,8 @@ public class Sequence extends ASequence implements SequenceI */ public Sequence(SequenceI seq, AlignmentAnnotation[] alAnnotation) { + this(); initSeqFrom(seq, alAnnotation); - } /** @@ -230,33 +258,39 @@ public class Sequence extends ASequence implements SequenceI protected void initSeqFrom(SequenceI seq, AlignmentAnnotation[] alAnnotation) { - { - char[] oseq = seq.getSequence(); - initSeqAndName(seq.getName(), Arrays.copyOf(oseq, oseq.length), - seq.getStart(), seq.getEnd()); - } + char[] oseq = seq.getSequence(); + initSeqAndName(seq.getName(), Arrays.copyOf(oseq, oseq.length), + seq.getStart(), seq.getEnd()); + description = seq.getDescription(); if (seq != datasetSequence) { setDatasetSequence(seq.getDatasetSequence()); } - if (datasetSequence == null && seq.getDBRefs() != null) + + /* + * only copy DBRefs and seqfeatures if we really are a dataset sequence + */ + if (datasetSequence == null) { - // only copy DBRefs and seqfeatures if we really are a dataset sequence - DBRefEntry[] dbr = seq.getDBRefs(); - for (int i = 0; i < dbr.length; i++) - { - addDBRef(new DBRefEntry(dbr[i])); - } - if (seq.getSequenceFeatures() != null) + if (seq.getDBRefs() != null) { - SequenceFeature[] sf = seq.getSequenceFeatures(); - for (int i = 0; i < sf.length; i++) + DBRefEntry[] dbr = seq.getDBRefs(); + for (int i = 0; i < dbr.length; i++) { - addSequenceFeature(new SequenceFeature(sf[i])); + addDBRef(new DBRefEntry(dbr[i])); } } + + /* + * make copies of any sequence features + */ + for (SequenceFeature sf : seq.getSequenceFeatures()) + { + addSequenceFeature(new SequenceFeature(sf)); + } } + if (seq.getAnnotation() != null) { AlignmentAnnotation[] sqann = seq.getAnnotation(); @@ -293,122 +327,67 @@ public class Sequence extends ASequence implements SequenceI } @Override - public void setSequenceFeatures(SequenceFeature[] features) + public void setSequenceFeatures(List features) { - if (datasetSequence == null) - { - sequenceFeatures = features; - } - else + if (datasetSequence != null) { - if (datasetSequence.getSequenceFeatures() != features - && datasetSequence.getSequenceFeatures() != null - && datasetSequence.getSequenceFeatures().length > 0) - { - new Exception( - "Warning: JAL-2046 side effect ? Possible implementation error: overwriting dataset sequence features by setting sequence features on alignment") - .printStackTrace(); - } datasetSequence.setSequenceFeatures(features); + return; } + sequenceFeatureStore = new SequenceFeatures(features); } @Override public synchronized boolean addSequenceFeature(SequenceFeature sf) { - if (sequenceFeatures == null && datasetSequence != null) + if (sf.getType() == null) { - return datasetSequence.addSequenceFeature(sf); - } - if (sequenceFeatures == null) - { - sequenceFeatures = new SequenceFeature[0]; + System.err.println("SequenceFeature type may not be null: " + + sf.toString()); + return false; } - for (int i = 0; i < sequenceFeatures.length; i++) + if (datasetSequence != null) { - if (sequenceFeatures[i].equals(sf)) - { - return false; - } + return datasetSequence.addSequenceFeature(sf); } - SequenceFeature[] temp = new SequenceFeature[sequenceFeatures.length + 1]; - System.arraycopy(sequenceFeatures, 0, temp, 0, sequenceFeatures.length); - temp[sequenceFeatures.length] = sf; - - sequenceFeatures = temp; - return true; + return sequenceFeatureStore.add(sf); } @Override public void deleteFeature(SequenceFeature sf) { - if (sequenceFeatures == null) - { - if (datasetSequence != null) - { - datasetSequence.deleteFeature(sf); - } - return; - } - - int index = 0; - for (index = 0; index < sequenceFeatures.length; index++) - { - if (sequenceFeatures[index].equals(sf)) - { - break; - } - } - - if (index == sequenceFeatures.length) - { - return; - } - - int sfLength = sequenceFeatures.length; - if (sfLength < 2) + if (datasetSequence != null) { - sequenceFeatures = null; + datasetSequence.deleteFeature(sf); } else { - SequenceFeature[] temp = new SequenceFeature[sfLength - 1]; - System.arraycopy(sequenceFeatures, 0, temp, 0, index); - - if (index < sfLength) - { - System.arraycopy(sequenceFeatures, index + 1, temp, index, - sequenceFeatures.length - index - 1); - } - - sequenceFeatures = temp; + sequenceFeatureStore.delete(sf); } } /** - * Returns the sequence features (if any), looking first on the sequence, then - * on its dataset sequence, and so on until a non-null value is found (or - * none). This supports retrieval of sequence features stored on the sequence - * (as in the applet) or on the dataset sequence (as in the Desktop version). + * {@inheritDoc} * * @return */ @Override - public SequenceFeature[] getSequenceFeatures() + public List getSequenceFeatures() { - SequenceFeature[] features = sequenceFeatures; - - SequenceI seq = this; - int count = 0; // failsafe against loop in sequence.datasetsequence... - while (features == null && seq.getDatasetSequence() != null - && count++ < 10) + if (datasetSequence != null) { - seq = seq.getDatasetSequence(); - features = ((Sequence) seq).sequenceFeatures; + return datasetSequence.getSequenceFeatures(); } - return features; + return sequenceFeatureStore.getAllFeatures(); + } + + @Override + public SequenceFeaturesI getFeatures() + { + return datasetSequence != null ? datasetSequence.getFeatures() + : sequenceFeatureStore; } @Override @@ -564,6 +543,7 @@ public class Sequence extends ASequence implements SequenceI { this.sequence = seq.toCharArray(); checkValidRange(); + sequenceChanged(); } @Override @@ -684,58 +664,424 @@ public class Sequence extends ASequence implements SequenceI return this.description; } - /* - * (non-Javadoc) - * - * @see jalview.datamodel.SequenceI#findIndex(int) + /** + * {@inheritDoc} */ @Override public int findIndex(int pos) { - // returns the alignment position for a residue + /* + * use a valid, hopefully nearby, cursor if available + */ + if (isValidCursor(cursor)) + { + return findIndex(pos, cursor); + } + int j = start; int i = 0; // Rely on end being at least as long as the length of the sequence. while ((i < sequence.length) && (j <= end) && (j <= pos)) { - if (!jalview.util.Comparison.isGap(sequence[i])) + if (!Comparison.isGap(sequence[i])) { j++; } - i++; } - if ((j == end) && (j < pos)) + if (j == end && j < pos) { return end + 1; } - else + + updateCursor(pos, i); + return i; + } + + /** + * Updates the cursor to the latest found residue and column position + * + * @param residuePos + * (start..) + * @param column + * (1..) + */ + protected void updateCursor(int residuePos, int column) + { + cursor = new SequenceCursor(this, residuePos, column, this.changeCount); + } + + /** + * Answers the aligned column position (1..) for the given residue position + * (start..) given a 'hint' of a residue/column location in the neighbourhood. + * The hint may be left of, at, or to the right of the required position. + * + * @param pos + * @param curs + * @return + */ + protected int findIndex(int pos, SequenceCursor curs) + { + if (!isValidCursor(curs)) { - return i; + /* + * wrong or invalidated cursor, compute de novo + */ + return findIndex(pos); } + + if (curs.residuePosition == pos) + { + return curs.columnPosition; + } + + /* + * move left or right to find pos from hint.position + */ + int col = curs.columnPosition - 1; // convert from base 1 to 0-based array + // index + int newPos = curs.residuePosition; + int delta = newPos > pos ? -1 : 1; + + while (newPos != pos) + { + col += delta; // shift one column left or right + if (col < 0 || col == sequence.length) + { + break; + } + if (!Comparison.isGap(sequence[col])) + { + newPos += delta; + } + } + + col++; // convert back to base 1 + updateCursor(pos, col); + + return col; } + /** + * {@inheritDoc} + */ @Override - public int findPosition(int i) + public int findPosition(final int column) { + /* + * use a valid, hopefully nearby, cursor if available + */ + if (isValidCursor(cursor)) + { + return findPosition(column + 1, cursor); + } + + // TODO recode this more naturally i.e. count residues only + // as they are found, not 'in anticipation' + + int lastPosFound = 0; + int lastPosFoundColumn = 0; + int seqlen = sequence.length; + if (seqlen > 0 && !Comparison.isGap(sequence[0])) + { + lastPosFound = start; + lastPosFoundColumn = 0; + } + int j = 0; int pos = start; - int seqlen = sequence.length; - while ((j < i) && (j < seqlen)) + + while (j < column && j < seqlen) { - if (!jalview.util.Comparison.isGap(sequence[j])) + if (!Comparison.isGap(sequence[j])) { + lastPosFound = pos; + lastPosFoundColumn = j; pos++; } - j++; } + if (j < seqlen && !Comparison.isGap(sequence[j])) + { + lastPosFound = pos; + lastPosFoundColumn = j; + } + + /* + * update the cursor to the last residue position found (if any) + * (converting column position to base 1) + */ + if (lastPosFound != 0) + { + updateCursor(lastPosFound, lastPosFoundColumn + 1); + } return pos; } /** + * Answers true if the given cursor is not null, is for this sequence object, + * and has a token value that matches this object's changeCount, else false. + * This allows us to ignore a cursor as 'stale' if the sequence has been + * modified since the cursor was created. + * + * @param curs + * @return + */ + protected boolean isValidCursor(SequenceCursor curs) + { + if (curs == null || curs.sequence != this || curs.token != changeCount) + { + return false; + } + /* + * sanity check against range + */ + if (curs.columnPosition < 0 || curs.columnPosition >= sequence.length) + { + return false; + } + if (curs.residuePosition < start || curs.residuePosition > end) + { + return false; + } + return true; + } + + /** + * Answers the sequence position (start..) for the given aligned column + * position (1..), given a hint of a cursor in the neighbourhood. The cursor + * may lie left of, at, or to the right of the column position. + * + * @param col + * @param curs + * @return + */ + protected int findPosition(final int col, SequenceCursor curs) + { + if (!isValidCursor(curs)) + { + /* + * wrong or invalidated cursor, compute de novo + */ + return findPosition(col - 1);// ugh back to base 0 + } + + if (curs.columnPosition == col) + { + cursor = curs; // in case this method becomes public + return curs.residuePosition; // easy case :-) + } + + /* + * move left or right to find pos from cursor position + */ + int column = curs.columnPosition - 1; // to base 0 + int newPos = curs.residuePosition; + int delta = curs.columnPosition > col ? -1 : 1; + boolean gapped = false; + int lastFoundPosition = curs.residuePosition; + int lastFoundPositionColumn = curs.columnPosition; + + while (column != col - 1) + { + column += delta; // shift one column left or right + if (column < 0 || column == sequence.length) + { + break; + } + gapped = Comparison.isGap(sequence[column]); + if (!gapped) + { + newPos += delta; + lastFoundPosition = newPos; + lastFoundPositionColumn = column + 1; + } + } + + if (cursor == null || lastFoundPosition != cursor.residuePosition) + { + updateCursor(lastFoundPosition, lastFoundPositionColumn); + } + + /* + * hack to give position to the right if on a gap + * or beyond the length of the sequence (see JAL-2562) + */ + if (delta > 0 && (gapped || column >= sequence.length)) + { + newPos++; + } + + return newPos; + } + + /** + * {@inheritDoc} + */ + @Override + public Range findPositions(int fromCol, int toCol) + { + if (cursor != null && cursor.sequence == this + && cursor.token == changeCount) + { + return findPositions(fromCol, toCol, cursor); + } + + /* + * count residues before fromCol + */ + int j = 0; + int count = 0; + int seqlen = sequence.length; + while (j < fromCol && j < seqlen) + { + if (!Comparison.isGap(sequence[j])) + { + count++; + } + j++; + } + + /* + * find first and last residues between fromCol and toCol + */ + int firstPos = 0; + int lastPos = 0; + boolean foundFirst = false; + + while (j <= toCol && j < seqlen) + { + if (!Comparison.isGap(sequence[j])) + { + count++; + if (!foundFirst) + { + firstPos = count; + foundFirst = true; + } + lastPos = count; + } + j++; + } + + if (firstPos == 0) + { + /* + * no residues in this range + */ + return null; + } + + /* + * adjust for sequence start coordinate + */ + firstPos += start - 1; + lastPos += start - 1; + + return new Range(firstPos, lastPos); + } + + /** + * Returns the range of sequence positions included in the given alignment + * position range. If no positions are included (the range is entirely gaps), + * then returns null. The cursor parameter may provide a starting position in + * the neighbourhood of the search (which may be left of, right of, or + * overlapping the search region). + * + * @param fromCol + * start column of region (0..) + * @param toCol + * end column of region (0..) + * @param curs + * @return + */ + protected Range findPositions(int fromCol, int toCol, SequenceCursor curs) + { + if (!isValidCursor(curs)) + { + /* + * wrong or invalidated cursor, compute de novo + */ + return findPositions(fromCol, toCol); + } + + /* + * keep this simple...first step from cursor to fromCol... + */ + final int seqlen = sequence.length; + int resNo = curs.residuePosition; + int col = curs.columnPosition - 1; // from base 1 to base 0 + if (col != fromCol) + { + int delta = col > fromCol ? -1 : 1; + while (col != fromCol && col >= 0 && col < seqlen) + { + if (!Comparison.isGap(sequence[col])) + { + resNo += delta; + } + col += delta; + } + } + + if (col < fromCol || col == seqlen) + { + /* + * sequence lies to the left of the target region + */ + return null; + } + + /* + * resNo is now the residue at fromCol (if not gapped), else the one + * before it (if delta == 1), else the one after (if delta == -1); + * we want the residue before fromCol + */ + if (!Comparison.isGap(sequence[fromCol])) + { + resNo--; + } + else if (curs.columnPosition > fromCol) + { + resNo -= 2; + } + + /* + * now first and last residues between fromCol and toCol + */ + int firstPos = 0; + int lastPos = 0; + boolean foundFirst = false; + + while (col <= toCol && col < seqlen) + { + if (!Comparison.isGap(sequence[col])) + { + resNo++; + if (!foundFirst) + { + firstPos = resNo; + foundFirst = true; + } + lastPos = resNo; + } + col++; + } + + if (firstPos == 0) + { + /* + * no residues in this range + */ + return null; + } + + return new Range(firstPos, lastPos); + } + + /** * Returns an int array where indices correspond to each residue in the * sequence and the element value gives its position in the alignment * @@ -891,6 +1237,7 @@ public class Sequence extends ASequence implements SequenceI start = newstart; end = newend; sequence = tmp; + sequenceChanged(); } @Override @@ -921,6 +1268,7 @@ public class Sequence extends ASequence implements SequenceI } sequence = tmp; + sequenceChanged(); } @Override @@ -1154,8 +1502,8 @@ public class Sequence extends ASequence implements SequenceI dsseq.setDescription(description); // move features and database references onto dataset sequence - dsseq.sequenceFeatures = sequenceFeatures; - sequenceFeatures = null; + dsseq.sequenceFeatureStore = sequenceFeatureStore; + sequenceFeatureStore = null; dsseq.dbrefs = dbrefs; dbrefs = null; // TODO: search and replace any references to this sequence with @@ -1214,11 +1562,11 @@ public class Sequence extends ASequence implements SequenceI return null; } - Vector subset = new Vector(); - Enumeration e = annotation.elements(); + Vector subset = new Vector(); + Enumeration e = annotation.elements(); while (e.hasMoreElements()) { - AlignmentAnnotation ann = (AlignmentAnnotation) e.nextElement(); + AlignmentAnnotation ann = e.nextElement(); if (ann.label != null && ann.label.equals(label)) { subset.addElement(ann); @@ -1233,7 +1581,7 @@ public class Sequence extends ASequence implements SequenceI e = subset.elements(); while (e.hasMoreElements()) { - anns[i++] = (AlignmentAnnotation) e.nextElement(); + anns[i++] = e.nextElement(); } subset.removeAllElements(); return anns; @@ -1286,12 +1634,12 @@ public class Sequence extends ASequence implements SequenceI if (entry.getSequenceFeatures() != null) { - SequenceFeature[] sfs = entry.getSequenceFeatures(); - for (int si = 0; si < sfs.length; si++) + List sfs = entry.getSequenceFeatures(); + for (SequenceFeature feature : sfs) { - SequenceFeature sf[] = (mp != null) ? mp.locateFeature(sfs[si]) - : new SequenceFeature[] { new SequenceFeature(sfs[si]) }; - if (sf != null && sf.length > 0) + SequenceFeature sf[] = (mp != null) ? mp.locateFeature(feature) + : new SequenceFeature[] { new SequenceFeature(feature) }; + if (sf != null) { for (int sfi = 0; sfi < sf.length; sfi++) { @@ -1304,10 +1652,10 @@ public class Sequence extends ASequence implements SequenceI // transfer PDB entries if (entry.getAllPDBEntries() != null) { - Enumeration e = entry.getAllPDBEntries().elements(); + Enumeration e = entry.getAllPDBEntries().elements(); while (e.hasMoreElements()) { - PDBEntry pdb = (PDBEntry) e.nextElement(); + PDBEntry pdb = e.nextElement(); addPDBId(pdb); } } @@ -1475,4 +1823,27 @@ public class Sequence extends ASequence implements SequenceI } } + /** + * {@inheritDoc} + */ + @Override + public List findFeatures(int from, int to, + String... types) + { + if (datasetSequence != null) + { + return datasetSequence.findFeatures(from, to, types); + } + return sequenceFeatureStore.findFeatures(from, to, types); + } + + /** + * Invalidates any stale cursors (forcing recalculation) by incrementing the + * token that has to match the one presented by the cursor + */ + @Override + public void sequenceChanged() + { + changeCount++; + } }