2 * Jalview - A Sequence Alignment Editor and Viewer ($$Version-Rel$$)
3 * Copyright (C) $$Year-Rel$$ The Jalview Authors
5 * This file is part of Jalview.
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.
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.
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.
21 package jalview.analysis;
23 import java.util.Locale;
25 import jalview.api.AlignViewportI;
26 import jalview.api.FinderI;
27 import jalview.datamodel.AlignmentI;
28 import jalview.datamodel.SearchResultMatchI;
29 import jalview.datamodel.SearchResults;
30 import jalview.datamodel.SearchResultsI;
31 import jalview.datamodel.SequenceGroup;
32 import jalview.datamodel.SequenceI;
33 import jalview.util.Comparison;
34 import jalview.util.MapList;
36 import java.util.ArrayList;
37 import java.util.Arrays;
38 import java.util.Iterator;
39 import java.util.List;
41 import com.stevesoft.pat.Regex;
44 * Implements the search algorithm for the Find dialog
46 public class Finder implements FinderI
49 * matched residue locations
51 private SearchResultsI searchResults;
54 * sequences matched by id or description
56 private List<SequenceI> idMatches;
59 * the viewport to search over
61 private AlignViewportI viewport;
64 * sequence index in alignment to search from
66 private int sequenceIndex;
69 * position offset in sequence to search from, base 0
70 * (position after start of last match for a 'find next')
72 private int residueIndex;
75 * the true sequence position of the start of the
76 * last sequence searched (when 'ignore hidden regions' does not apply)
78 private int searchedSequenceStartPosition;
81 * when 'ignore hidden regions' applies, this holds the mapping from
82 * the visible sequence positions (1, 2, ...) to true sequence positions
84 private MapList searchedSequenceMap;
86 private String seqToSearch;
89 * Constructor for searching a viewport
93 public Finder(AlignViewportI av)
96 this.sequenceIndex = 0;
97 this.residueIndex = -1;
101 public void findAll(String theSearchString, boolean matchCase,
102 boolean searchDescription, boolean ignoreHidden)
105 * search from the start
110 doFind(theSearchString, matchCase, searchDescription, true,
114 * reset to start for next search
121 public void findNext(String theSearchString, boolean matchCase,
122 boolean searchDescription, boolean ignoreHidden)
124 doFind(theSearchString, matchCase, searchDescription, false,
127 if (searchResults.isEmpty() && idMatches.isEmpty())
130 * search failed - reset to start for next search
138 * Performs a 'find next' or 'find all'
140 * @param theSearchString
142 * @param searchDescription
144 * @param ignoreHidden
146 protected void doFind(String theSearchString, boolean matchCase,
147 boolean searchDescription, boolean findAll, boolean ignoreHidden)
149 searchResults = new SearchResults();
150 idMatches = new ArrayList<>();
152 String searchString = matchCase ? theSearchString
153 : theSearchString.toUpperCase(Locale.ROOT);
154 Regex searchPattern = new Regex(searchString);
155 searchPattern.setIgnoreCase(!matchCase);
157 SequenceGroup selection = viewport.getSelectionGroup();
158 if (selection != null && selection.getSize() < 1)
160 selection = null; // ? ignore column-only selection
163 AlignmentI alignment = viewport.getAlignment();
164 int end = alignment.getHeight();
166 getSequence(ignoreHidden);
168 boolean found = false;
169 while ((!found || findAll) && sequenceIndex < end)
171 found = findNextMatch(searchString, searchPattern, searchDescription,
177 * Calculates and saves the sequence string to search. The string is
178 * restricted to the current selection region if there is one, and is saved
179 * with all gaps removed.
181 * If there are hidden columns, and option {@ignoreHidden} is selected, then
182 * only visible positions of the sequence are included, and a mapping is also
183 * constructed from the returned string positions to the true sequence
186 * Note we have to do this each time {@code findNext} or {@code findAll} is
187 * called, in case the alignment, selection group or hidden columns have
188 * changed. In particular, if the sequence at offset {@code sequenceIndex} in
189 * the alignment is (no longer) in the selection group, search is advanced to
190 * the next sequence that is.
192 * Sets sequence string to the empty string if there are no more sequences (in
193 * selection group if any) at or after {@code sequenceIndex}.
195 * Returns true if a sequence could be found, false if end of alignment was
198 * @param ignoreHidden
201 private boolean getSequence(boolean ignoreHidden)
203 AlignmentI alignment = viewport.getAlignment();
204 if (sequenceIndex >= alignment.getHeight())
209 SequenceI seq = alignment.getSequenceAt(sequenceIndex);
210 SequenceGroup selection = viewport.getSelectionGroup();
211 if (selection != null && !selection.contains(seq))
213 if (!nextSequence(ignoreHidden))
217 seq = alignment.getSequenceAt(sequenceIndex);
220 String seqString = null;
223 seqString = getVisibleSequence(seq);
224 this.searchedSequenceStartPosition = 1;
229 int endCol = seq.getLength() - 1;
230 this.searchedSequenceStartPosition = seq.getStart();
231 if (selection != null)
233 startCol = selection.getStartRes();
234 endCol = Math.min(endCol, selection.getEndRes());
235 this.searchedSequenceStartPosition = seq.findPosition(startCol);
237 seqString = seq.getSequenceAsString(startCol, endCol + 1);
241 * remove gaps; note that even if this leaves an empty string, we 'search'
242 * the sequence anyway (for possible match on name or description)
244 String ungapped = AlignSeq.extractGaps(Comparison.GapChars, seqString);
245 this.seqToSearch = ungapped;
251 * Returns a string consisting of only the visible residues of {@code seq}
252 * from alignment column {@ fromColumn}, restricted to the current selection
253 * region if there is one.
255 * As a side-effect, also computes the mapping from the true sequence
256 * positions to the positions (1, 2, ...) of the returned sequence. This is to
257 * allow search matches in the visible sequence to be converted to sequence
263 private String getVisibleSequence(SequenceI seq)
266 * get start / end columns of sequence and convert to base 0
267 * (so as to match the visible column ranges)
269 int seqStartCol = seq.findIndex(seq.getStart()) - 1;
270 int seqEndCol = seq.findIndex(seq.getStart() + seq.getLength() - 1) - 1;
271 Iterator<int[]> visibleColumns = viewport.getViewAsVisibleContigs(true);
272 StringBuilder visibleSeq = new StringBuilder(seqEndCol - seqStartCol);
273 List<int[]> fromRanges = new ArrayList<>();
275 while (visibleColumns.hasNext())
277 int[] range = visibleColumns.next();
278 if (range[0] > seqEndCol)
280 // beyond the end of the sequence
283 if (range[1] < seqStartCol)
285 // before the start of the sequence
288 String subseq = seq.getSequenceAsString(range[0], range[1] + 1);
289 String ungapped = AlignSeq.extractGaps(Comparison.GapChars, subseq);
290 visibleSeq.append(ungapped);
291 if (!ungapped.isEmpty())
294 * visible region includes at least one non-gap character,
295 * so add the range to the mapping being constructed
297 int seqResFrom = seq.findPosition(range[0]);
298 int seqResTo = seqResFrom + ungapped.length() - 1;
299 fromRanges.add(new int[] { seqResFrom, seqResTo });
304 * construct the mapping
305 * from: visible sequence positions 1..length
306 * to: true residue positions of the alignment sequence
308 List<int[]> toRange = Arrays
310 { 1, visibleSeq.length() });
311 searchedSequenceMap = new MapList(fromRanges, toRange, 1, 1);
313 return visibleSeq.toString();
317 * Advances the search to the next sequence in the alignment. Sequences not in
318 * the current selection group (if there is one) are skipped. The
319 * (sub-)sequence to be searched is extracted, gaps removed, and saved, or set
320 * to null if there are no more sequences to search.
322 * Returns true if a sequence could be found, false if end of alignment was
325 * @param ignoreHidden
327 private boolean nextSequence(boolean ignoreHidden)
332 return getSequence(ignoreHidden);
336 * Finds the next match in the given sequence, starting at offset
337 * {@code residueIndex}. Answers true if a match is found, else false.
339 * If a match is found, {@code residueIndex} is advanced to the position after
340 * the start of the matched region, ready for the next search.
342 * If no match is found, {@code sequenceIndex} is advanced ready to search the
346 * @param searchString
347 * @param searchPattern
348 * @param matchDescription
349 * @param ignoreHidden
352 protected boolean findNextMatch(String searchString, Regex searchPattern,
353 boolean matchDescription, boolean ignoreHidden)
355 if (residueIndex < 0)
358 * at start of sequence; try find by residue number, in sequence id,
359 * or (optionally) in sequence description
361 if (doNonMotifSearches(searchString, searchPattern, matchDescription))
368 * search for next match in sequence string
370 int end = seqToSearch.length();
371 while (residueIndex < end)
373 boolean matched = searchPattern.searchFrom(seqToSearch, residueIndex);
376 if (recordMatch(searchPattern, ignoreHidden))
383 residueIndex = Integer.MAX_VALUE;
387 nextSequence(ignoreHidden);
392 * Adds the match held in the <code>searchPattern</code> Regex to the
393 * <code>searchResults</code>, unless it is a subregion of the last match
394 * recorded. <code>residueIndex</code> is advanced to the position after the
395 * start of the matched region, ready for the next search. Answers true if a
396 * match was added, else false.
398 * Matches that lie entirely within hidden regions of the alignment are not
401 * @param searchPattern
402 * @param ignoreHidden
405 protected boolean recordMatch(Regex searchPattern, boolean ignoreHidden)
407 SequenceI seq = viewport.getAlignment().getSequenceAt(sequenceIndex);
410 * convert start/end of the match to sequence coordinates
412 int offset = searchPattern.matchedFrom();
413 int matchStartPosition = this.searchedSequenceStartPosition + offset;
414 int matchEndPosition = matchStartPosition + searchPattern.charsMatched()
418 * update residueIndex to next position after the start of the match
419 * (findIndex returns a value base 1, columnIndex is held base 0)
421 residueIndex = searchPattern.matchedFrom() + 1;
424 * return false if the match is entirely in a hidden region
426 if (allHidden(seq, matchStartPosition, matchEndPosition))
432 * check that this match is not a subset of the previous one (JAL-2302)
434 List<SearchResultMatchI> matches = searchResults.getResults();
435 SearchResultMatchI lastMatch = matches.isEmpty() ? null
436 : matches.get(matches.size() - 1);
438 if (lastMatch == null || !lastMatch.contains(seq, matchStartPosition,
441 addMatch(seq, matchStartPosition, matchEndPosition, ignoreHidden);
449 * Adds one match to the stored list. If hidden residues are being skipped,
450 * then the match may need to be split into contiguous positions of the
451 * sequence (so it does not include skipped residues).
454 * @param matchStartPosition
455 * @param matchEndPosition
456 * @param ignoreHidden
458 private void addMatch(SequenceI seq, int matchStartPosition,
459 int matchEndPosition, boolean ignoreHidden)
466 searchResults.addResult(seq, matchStartPosition, matchEndPosition);
471 * get start-end contiguous ranges in underlying sequence
473 int[] truePositions = searchedSequenceMap
474 .locateInFrom(matchStartPosition, matchEndPosition);
475 searchResults.addResult(seq, truePositions);
479 * Returns true if all residues are hidden, else false
486 private boolean allHidden(SequenceI seq, int fromPos, int toPos)
488 if (!viewport.hasHiddenColumns())
492 for (int res = fromPos; res <= toPos; res++)
494 if (isVisible(seq, res))
503 * Does searches other than for residue patterns. Currently this includes
505 * <li>find residue by position (if search string is a number)</li>
506 * <li>match search string to sequence id</li>
507 * <li>match search string to sequence description (optional)</li>
509 * Answers true if a match is found, else false.
511 * @param searchString
512 * @param searchPattern
513 * @param includeDescription
516 protected boolean doNonMotifSearches(String searchString,
517 Regex searchPattern, boolean includeDescription)
519 SequenceI seq = viewport.getAlignment().getSequenceAt(sequenceIndex);
522 * position sequence search to start of sequence
527 int res = Integer.parseInt(searchString);
528 return searchForResidueNumber(seq, res);
529 } catch (NumberFormatException ex)
531 // search pattern is not a number
534 if (searchSequenceName(seq, searchPattern))
538 if (includeDescription && searchSequenceDescription(seq, searchPattern))
546 * Searches for a match with the sequence description, and if found, adds the
547 * sequence to the list of match ids (but not as a duplicate). Answers true if
548 * a match was added, else false.
551 * @param searchPattern
554 protected boolean searchSequenceDescription(SequenceI seq,
557 String desc = seq.getDescription();
558 if (desc != null && searchPattern.search(desc)
559 && !idMatches.contains(seq))
568 * Searches for a match with the sequence name, and if found, adds the
569 * sequence to the list of match ids (but not as a duplicate). Answers true if
570 * a match was added, else false.
573 * @param searchPattern
576 protected boolean searchSequenceName(SequenceI seq, Regex searchPattern)
578 if (searchPattern.search(seq.getName()) && !idMatches.contains(seq))
587 * If the residue position is valid for the sequence, and in a visible column,
588 * adds the position to the search results and returns true, else answers
595 protected boolean searchForResidueNumber(SequenceI seq, int resNo)
597 if (seq.getStart() <= resNo && seq.getEnd() >= resNo)
599 if (isVisible(seq, resNo))
601 searchResults.addResult(seq, resNo, resNo);
609 * Returns true if the residue is in a visible column, else false
615 private boolean isVisible(SequenceI seq, int res)
617 if (!viewport.hasHiddenColumns())
621 int col = seq.findIndex(res); // base 1
622 return viewport.getAlignment().getHiddenColumns().isVisible(col - 1); // base
627 public List<SequenceI> getIdMatches()
633 public SearchResultsI getSearchResults()
635 return searchResults;