/*
- * Jalview - A Sequence Alignment Editor and Viewer (Version 2.9)
- * Copyright (C) 2015 The Jalview Authors
+ * Jalview - A Sequence Alignment Editor and Viewer ($$Version-Rel$$)
+ * Copyright (C) $$Year-Rel$$ The Jalview Authors
*
* This file is part of Jalview.
*
*/
package jalview.analysis;
+import jalview.api.AlignViewportI;
+import jalview.api.FinderI;
import jalview.datamodel.AlignmentI;
+import jalview.datamodel.Range;
+import jalview.datamodel.SearchResultMatchI;
import jalview.datamodel.SearchResults;
-import jalview.datamodel.Sequence;
+import jalview.datamodel.SearchResultsI;
import jalview.datamodel.SequenceGroup;
+import jalview.datamodel.SequenceI;
+import jalview.datamodel.VisibleContigsIterator;
+import jalview.util.Comparison;
+import java.util.List;
import java.util.Vector;
-public class Finder
+import com.stevesoft.pat.Regex;
+
+/**
+ * Implements the search algorithm for the Find dialog
+ */
+public class Finder implements FinderI
{
- /**
- * Implements the search algorithms for the Find dialog box.
+ /*
+ * matched residue locations
*/
- SearchResults searchResults;
+ private SearchResultsI searchResults;
- AlignmentI alignment;
+ /*
+ * sequences matched by id or description
+ */
+ private Vector<SequenceI> idMatches;
- jalview.datamodel.SequenceGroup selection = null;
+ /*
+ * the viewport to search over
+ */
+ private AlignViewportI viewport;
- Vector idMatch = null;
+ /*
+ * sequence index in alignment to search from
+ */
+ private int sequenceIndex;
- boolean caseSensitive = false;
+ /*
+ * column position in sequence to search from, base 0
+ * - absolute column number including any hidden columns
+ * (position after start of last match for a repeat search)
+ */
+ private int columnIndex;
- private boolean includeDescription = false;
+ /**
+ * Constructor for searching a viewport
+ *
+ * @param av
+ */
+ public Finder(AlignViewportI av)
+ {
+ this.viewport = av;
+ this.sequenceIndex = 0;
+ this.columnIndex = -1;
+ }
- boolean findAll = false;
+ @Override
+ public void findAll(String theSearchString, boolean matchCase,
+ boolean searchDescription)
+ {
+ /*
+ * search from the start
+ */
+ sequenceIndex = 0;
+ columnIndex = -1;
- com.stevesoft.pat.Regex regex = null;
+ doFind(theSearchString, matchCase, searchDescription, true);
- /**
- * hold's last-searched position between calles to find(false)
- */
- int seqIndex = 0, resIndex = -1;
+ /*
+ * reset to start for next search
+ */
+ sequenceIndex = 0;
+ columnIndex = -1;
+ }
- public Finder(AlignmentI alignment, SequenceGroup selection)
+ @Override
+ public void findNext(String theSearchString, boolean matchCase,
+ boolean searchDescription)
{
- this.alignment = alignment;
- this.selection = selection;
+ doFind(theSearchString, matchCase, searchDescription, false);
+
+ if (searchResults.isEmpty() && idMatches.isEmpty())
+ {
+ /*
+ * search failed - reset to start for next search
+ */
+ sequenceIndex = 0;
+ columnIndex = -1;
+ }
}
/**
- * restart search at given sequence and residue on alignment and (optionally)
- * contained in selection
+ * Performs a 'find next' or 'find all'
*
- * @param alignment
- * @param selectionGroup
- * @param seqIndex
- * @param resIndex
+ * @param theSearchString
+ * @param matchCase
+ * @param searchDescription
+ * @param findAll
*/
- public Finder(AlignmentI alignment, SequenceGroup selectionGroup,
- int seqIndex, int resIndex)
+ protected void doFind(String theSearchString, boolean matchCase,
+ boolean searchDescription, boolean findAll)
{
- this(alignment, selectionGroup);
- this.seqIndex = seqIndex;
- this.resIndex = resIndex;
- }
+ String searchString = matchCase ? theSearchString
+ : theSearchString.toUpperCase();
+ Regex searchPattern = new Regex(searchString);
+ searchPattern.setIgnoreCase(!matchCase);
- public boolean find(String searchString)
- {
- boolean hasResults = false;
- if (!caseSensitive)
- {
- searchString = searchString.toUpperCase();
- }
- regex = new com.stevesoft.pat.Regex(searchString);
- regex.setIgnoreCase(!caseSensitive);
searchResults = new SearchResults();
- idMatch = new Vector();
- Sequence seq;
- String item = null;
- boolean found = false;
- int end = alignment.getHeight();
-
- // /////////////////////////////////////////////
+ idMatches = new Vector<>();
- if (selection != null)
+ SequenceGroup selection = viewport.getSelectionGroup();
+ if (selection != null && selection.getSize() < 1)
{
- if ((selection.getSize() < 1)
- || ((selection.getEndRes() - selection.getStartRes()) < 2))
- {
- selection = null;
- }
+ selection = null; // ? ignore column-only selection
}
- while (!found && (seqIndex < end))
- {
- seq = (Sequence) alignment.getSequenceAt(seqIndex);
+ AlignmentI alignment = viewport.getAlignment();
+ int end = alignment.getHeight();
- if ((selection != null && selection.getSize() > 0)
- && !selection.getSequences(null).contains(seq))
+ while (sequenceIndex < end)
+ {
+ SequenceI seq = alignment.getSequenceAt(sequenceIndex);
+ boolean found = findNextMatch(seq, searchString, searchPattern,
+ searchDescription);
+ if (found && !findAll)
{
- seqIndex++;
- resIndex = -1;
-
- continue;
+ return;
}
- if (resIndex < 0)
+ if (!found)
{
- resIndex = 0;
- // test for one off matches - sequence position and sequence ID
- // //// is the searchString a residue number?
- try
- {
- int res = Integer.parseInt(searchString);
- // possibly a residue number - check if valid for seq
- if (seq.getEnd() >= res)
- {
- searchResults.addResult(seq, res, res);
- hasResults = true;
- // resIndex=seq.getLength();
- // seqIndex++;
- if (!findAll)
- {
- found = true;
- break;
- }
- }
- } catch (NumberFormatException ex)
- {
- }
-
- if (regex.search(seq.getName()))
- {
- idMatch.addElement(seq);
- hasResults = true;
- if (!findAll)
- {
- // stop and return the match
- found = true;
- break;
- }
- }
-
- if (isIncludeDescription() && seq.getDescription() != null
- && regex.search(seq.getDescription()))
- {
- idMatch.addElement(seq);
- hasResults = true;
- if (!findAll)
- {
- // stop and return the match
- found = true;
- break;
- }
- }
+ sequenceIndex++;
+ columnIndex = -1;
}
- item = seq.getSequenceAsString();
+ }
+ }
- if ((selection != null)
- && (selection.getEndRes() < alignment.getWidth() - 1))
- {
- item = item.substring(0, selection.getEndRes() + 1);
- }
+ /**
+ * Answers the start-end column range of the visible region of
+ * <code>sequence</code> starting at or after the given <code>column</code>.
+ * If there are no hidden columns, this just returns the remaining width of
+ * the sequence. The range is restricted to the current <code>selection</code>
+ * if there is one. Answers null if there are no visible columns at or after
+ * <code>column</code>.
+ */
+ protected Range getNextVisibleSequenceRegion(SequenceI sequence,
+ int column)
+ {
+ int seqColStart = column;
+ int seqColEnd = sequence.getLength() - 1;
- // /Shall we ignore gaps???? - JBPNote: Add Flag for forcing this or not
- StringBuffer noGapsSB = new StringBuffer();
- int insertCount = 0;
- Vector spaces = new Vector();
+ /*
+ * restrict search to (next) visible column region,
+ * in case there are hidden columns
+ */
+ AlignmentI alignment = viewport.getAlignment();
+ VisibleContigsIterator visibleRegions = alignment.getHiddenColumns()
+ .getVisContigsIterator(column, alignment.getWidth(),
+ false);
+ int[] visible = visibleRegions.hasNext() ? visibleRegions.next() : null;
+ if (visible == null)
+ {
+ columnIndex = seqColEnd + 1;
+ return null;
+ }
+ seqColStart = Math.max(seqColStart, visible[0]);
+ seqColEnd = Math.min(seqColEnd, visible[1]);
- for (int j = 0; j < item.length(); j++)
+ /*
+ * restrict search to selected region if there is one
+ */
+ SequenceGroup selection = viewport.getSelectionGroup();
+ if (selection != null)
+ {
+ int selectionStart = selection.getStartRes();
+ int selectionEnd = selection.getEndRes();
+ if (selectionStart > seqColEnd || selectionEnd < seqColStart)
{
- if (!jalview.util.Comparison.isGap(item.charAt(j)))
- {
- noGapsSB.append(item.charAt(j));
- spaces.addElement(new Integer(insertCount));
- }
- else
- {
- insertCount++;
- }
+ /*
+ * sequence region doesn't overlap selection region
+ */
+ columnIndex = seqColEnd + 1;
+ return null;
}
+ seqColStart = Math.max(seqColStart, selectionStart);
+ seqColEnd = Math.min(seqColEnd, selectionEnd);
+ }
- String noGaps = noGapsSB.toString();
-
- for (int r = resIndex; r < noGaps.length(); r++)
- {
+ return new Range(seqColStart, seqColEnd);
+ }
- if (regex.searchFrom(noGaps, r))
- {
- resIndex = regex.matchedFrom();
-
- if ((selection != null && selection.getSize() > 0)
- && ((resIndex + Integer.parseInt(spaces.elementAt(
- resIndex).toString())) < selection.getStartRes()))
- {
- continue;
- }
- // if invalid string used, then regex has no matched to/from
- int sres = seq
- .findPosition(resIndex
- + Integer.parseInt(spaces.elementAt(resIndex)
- .toString()));
- int eres = seq.findPosition(regex.matchedTo()
- - 1
- + Integer.parseInt(spaces
- .elementAt(regex.matchedTo() - 1).toString()));
-
- searchResults.addResult(seq, sres, eres);
- hasResults = true;
- if (!findAll)
- {
- // thats enough, break and display the result
- found = true;
- resIndex++;
-
- break;
- }
-
- r = resIndex;
- }
- else
- {
- break;
- }
- }
+ /**
+ * Finds the next match in the given sequence, starting at column at
+ * <code>columnIndex</code>. Answers true if a match is found, else false. If
+ * a match is found, <code>columnIndex</code> is advanced to the column after
+ * the start of the matched region, ready for a search from the next position.
+ *
+ * @param seq
+ * @param searchString
+ * @param searchPattern
+ * @param matchDescription
+ * @return
+ */
+ protected boolean findNextMatch(SequenceI seq, String searchString,
+ Regex searchPattern, boolean matchDescription)
+ {
+ SequenceGroup selection = viewport.getSelectionGroup();
+ if (selection != null && !selection.contains(seq))
+ {
+ /*
+ * this sequence is not in the selection - advance to next sequence
+ */
+ return false;
+ }
- if (!found)
+ if (columnIndex < 0)
+ {
+ /*
+ * at start of sequence; try find by residue number, in sequence id,
+ * or (optionally) in sequence description
+ */
+ if (doNonMotifSearches(seq, searchString, searchPattern,
+ matchDescription))
{
- seqIndex++;
- resIndex = -1;
+ return true;
}
}
- /**
- * We now search the Id string in the main search loop. for (int id = 0; id
- * < alignment.getHeight(); id++) { if
- * (regex.search(alignment.getSequenceAt(id).getName())) {
- * idMatch.addElement(alignment.getSequenceAt(id)); hasResults = true; } }
+ /*
+ * search for next match in sequence string
*/
- return hasResults;
- }
-
- /**
- * @return the alignment
- */
- public AlignmentI getAlignment()
- {
- return alignment;
+ int end = seq.getLength();
+ while (columnIndex < end)
+ {
+ if (searchNextVisibleRegion(seq, searchPattern))
+ {
+ return true;
+ }
+ }
+ return false;
}
/**
- * @param alignment
- * the alignment to set
+ * Searches the sequence, starting from <code>columnIndex</code>, and adds the
+ * next match (if any) to <code>searchResults</code>. The search is restricted
+ * to the next visible column region, and to the <code>selection</code> region
+ * if there is one. Answers true if a match is added, else false.
+ *
+ * @param seq
+ * @param searchPattern
+ * @return
*/
- public void setAlignment(AlignmentI alignment)
+ protected boolean searchNextVisibleRegion(SequenceI seq, Regex searchPattern)
{
- this.alignment = alignment;
- }
+ Range visible = getNextVisibleSequenceRegion(seq, columnIndex);
+ if (visible == null)
+ {
+ return false;
+ }
+ String seqString = seq.getSequenceAsString(visible.start, visible.end + 1);
+ String noGaps = AlignSeq.extractGaps(Comparison.GapChars, seqString);
- /**
- * @return the caseSensitive
- */
- public boolean isCaseSensitive()
- {
- return caseSensitive;
- }
+ if (searchPattern.search(noGaps))
+ {
+ int sequenceStartPosition = seq.findPosition(visible.start);
+ recordMatch(seq, searchPattern, sequenceStartPosition);
+ return true;
+ }
+ else
+ {
+ /*
+ * no match - advance columnIndex past this visible region
+ * so the next visible region (if any) is searched next
+ */
+ columnIndex = visible.end + 1;
+ }
- /**
- * @param caseSensitive
- * the caseSensitive to set
- */
- public void setCaseSensitive(boolean caseSensitive)
- {
- this.caseSensitive = caseSensitive;
+ return false;
}
/**
- * @return the findAll
+ * Adds the match held in the <code>searchPattern</code> Regex to the
+ * <code>searchResults</code>, unless it is a subregion of the last match
+ * recorded. <code>columnIndex</code> is advanced to the position after the
+ * start of the matched region, ready for the next search. Answers true if a
+ * match was added, else false.
+ *
+ * @param seq
+ * @param searchPattern
+ * @param firstResiduePosition
+ * @return
*/
- public boolean isFindAll()
+ protected boolean recordMatch(SequenceI seq, Regex searchPattern,
+ int firstResiduePosition)
{
- return findAll;
- }
+ /*
+ * get start/end of the match in sequence coordinates
+ */
+ int offset = searchPattern.matchedFrom();
+ int matchStartPosition = firstResiduePosition + offset;
+ int matchEndPosition = matchStartPosition
+ + searchPattern.charsMatched() - 1;
+
+ /*
+ * update columnIndex to next column after the start of the match
+ * (findIndex returns a value base 1, columnIndex is held base 0)
+ */
+ columnIndex = seq.findIndex(matchStartPosition);
- /**
- * @param findAll
- * the findAll to set
- */
- public void setFindAll(boolean findAll)
- {
- this.findAll = findAll;
- }
+ /*
+ * check that this match is not a subset of the previous one (JAL-2302)
+ */
+ List<SearchResultMatchI> matches = searchResults.getResults();
+ SearchResultMatchI lastMatch = matches.isEmpty() ? null
+ : matches.get(matches.size() - 1);
- /**
- * @return the selection
- */
- public jalview.datamodel.SequenceGroup getSelection()
- {
- return selection;
- }
+ if (lastMatch == null || !lastMatch.contains(seq, matchStartPosition,
+ matchEndPosition))
+ {
+ searchResults.addResult(seq, matchStartPosition, matchEndPosition);
+ return true;
+ }
- /**
- * @param selection
- * the selection to set
- */
- public void setSelection(jalview.datamodel.SequenceGroup selection)
- {
- this.selection = selection;
+ return false;
}
/**
- * @return the idMatch
+ * Does searches other than for residue patterns. Currently this includes
+ * <ul>
+ * <li>find residue by position (if search string is a number)</li>
+ * <li>match search string to sequence id</li>
+ * <li>match search string to sequence description (optional)</li>
+ * </ul>
+ * Answers true if a match is found, else false.
+ *
+ * @param seq
+ * @param searchString
+ * @param searchPattern
+ * @param includeDescription
+ * @return
*/
- public Vector getIdMatch()
+ protected boolean doNonMotifSearches(SequenceI seq, String searchString,
+ Regex searchPattern, boolean includeDescription)
{
- return idMatch;
- }
+ /*
+ * position sequence search to start of sequence
+ */
+ columnIndex = 0;
- /**
- * @return the regex
- */
- public com.stevesoft.pat.Regex getRegex()
- {
- return regex;
+ if (searchForResidueNumber(seq, searchString))
+ {
+ return true;
+ }
+ if (searchSequenceName(seq, searchPattern))
+ {
+ return true;
+ }
+ if (includeDescription && searchSequenceDescription(seq, searchPattern))
+ {
+ return true;
+ }
+ return false;
}
/**
- * @return the searchResults
+ * Searches for a match with the sequence description, and if found, adds the
+ * sequence to the list of match ids (but not as a duplicate). Answers true if
+ * a match was added, else false.
+ *
+ * @param seq
+ * @param searchPattern
+ * @return
*/
- public SearchResults getSearchResults()
+ protected boolean searchSequenceDescription(SequenceI seq, Regex searchPattern)
{
- return searchResults;
+ String desc = seq.getDescription();
+ if (desc != null && searchPattern.search(desc) && !idMatches.contains(seq))
+ {
+ idMatches.addElement(seq);
+ return true;
+ }
+ return false;
}
/**
- * @return the resIndex
+ * Searches for a match with the sequence name, and if found, adds the
+ * sequence to the list of match ids (but not as a duplicate). Answers true if
+ * a match was added, else false.
+ *
+ * @param seq
+ * @param searchPattern
+ * @return
*/
- public int getResIndex()
+ protected boolean searchSequenceName(SequenceI seq, Regex searchPattern)
{
- return resIndex;
+ if (searchPattern.search(seq.getName()) && !idMatches.contains(seq))
+ {
+ idMatches.addElement(seq);
+ return true;
+ }
+ return false;
}
/**
- * @param resIndex
- * the resIndex to set
+ * Tries to interpret the search string as a residue position, and if valid,
+ * adds the position to the search results and returns true, else answers
+ * false
*/
- public void setResIndex(int resIndex)
+ protected boolean searchForResidueNumber(SequenceI seq, String searchString)
{
- this.resIndex = resIndex;
+ try
+ {
+ int res = Integer.parseInt(searchString);
+ if (seq.getStart() <= res && seq.getEnd() >= res)
+ {
+ searchResults.addResult(seq, res, res);
+ return true;
+ }
+ } catch (NumberFormatException ex)
+ {
+ }
+ return false;
}
- /**
- * @return the seqIndex
+ /* (non-Javadoc)
+ * @see jalview.analysis.FinderI#getIdMatch()
*/
- public int getSeqIndex()
+ @Override
+ public Vector<SequenceI> getIdMatches()
{
- return seqIndex;
+ return idMatches;
}
- /**
- * @param seqIndex
- * the seqIndex to set
+ /* (non-Javadoc)
+ * @see jalview.analysis.FinderI#getSearchResults()
*/
- public void setSeqIndex(int seqIndex)
- {
- this.seqIndex = seqIndex;
- }
-
- public boolean isIncludeDescription()
+ @Override
+ public SearchResultsI getSearchResults()
{
- return includeDescription;
- }
-
- public void setIncludeDescription(boolean includeDescription)
- {
- this.includeDescription = includeDescription;
+ return searchResults;
}
}