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.Platform;
35 import jalview.util.MapList;
37 import java.util.ArrayList;
38 import java.util.Arrays;
39 import java.util.Iterator;
40 import java.util.List;
42 import com.stevesoft.pat.Regex;
45 * Implements the search algorithm for the Find dialog
47 public class Finder implements FinderI
50 * matched residue locations
52 private SearchResultsI searchResults;
55 * sequences matched by id or description
57 private List<SequenceI> idMatches;
60 * the viewport to search over
62 private AlignViewportI viewport;
65 * sequence index in alignment to search from
67 private int sequenceIndex;
70 * position offset in sequence to search from, base 0
71 * (position after start of last match for a 'find next')
73 private int residueIndex;
76 * the true sequence position of the start of the
77 * last sequence searched (when 'ignore hidden regions' does not apply)
79 private int searchedSequenceStartPosition;
82 * when 'ignore hidden regions' applies, this holds the mapping from
83 * the visible sequence positions (1, 2, ...) to true sequence positions
85 private MapList searchedSequenceMap;
87 private String seqToSearch;
90 * Constructor for searching a viewport
94 public Finder(AlignViewportI av)
97 this.sequenceIndex = 0;
98 this.residueIndex = -1;
102 public void findAll(String theSearchString, boolean matchCase,
103 boolean searchDescription, boolean ignoreHidden)
106 * search from the start
111 doFind(theSearchString, matchCase, searchDescription, true,
115 * reset to start for next search
122 public void findNext(String theSearchString, boolean matchCase,
123 boolean searchDescription, boolean ignoreHidden)
125 doFind(theSearchString, matchCase, searchDescription, false,
128 if (searchResults.isEmpty() && idMatches.isEmpty())
131 * search failed - reset to start for next search
139 * Performs a 'find next' or 'find all'
141 * @param theSearchString
143 * @param searchDescription
145 * @param ignoreHidden
147 protected void doFind(String theSearchString, boolean matchCase,
148 boolean searchDescription, boolean findAll, boolean ignoreHidden)
150 searchResults = new SearchResults();
151 idMatches = new ArrayList<>();
153 String searchString = matchCase ? theSearchString
154 : theSearchString.toUpperCase(Locale.ROOT);
155 Regex searchPattern = Platform.newRegex(searchString);
156 searchPattern.setIgnoreCase(!matchCase);
158 SequenceGroup selection = viewport.getSelectionGroup();
159 if (selection != null && selection.getSize() < 1)
161 selection = null; // ? ignore column-only selection
164 AlignmentI alignment = viewport.getAlignment();
165 int end = alignment.getHeight();
167 getSequence(ignoreHidden);
169 boolean found = false;
170 while ((!found || findAll) && sequenceIndex < end)
172 found = findNextMatch(searchString, searchPattern, searchDescription,
178 * Calculates and saves the sequence string to search. The string is restricted
179 * to the current selection region if there is one, and is saved with all gaps
182 * If there are hidden columns, and option {@ignoreHidden} is selected, then
183 * only visible positions of the sequence are included, and a mapping is also
184 * constructed from the returned string positions to the true sequence
187 * Note we have to do this each time {@code findNext} or {@code findAll} is
188 * called, in case the alignment, selection group or hidden columns have
189 * changed. In particular, if the sequence at offset {@code sequenceIndex} in
190 * the alignment is (no longer) in the selection group, search is advanced to
191 * the next sequence that is.
193 * Sets sequence string to the empty string if there are no more sequences (in
194 * selection group if any) at or after {@code sequenceIndex}.
196 * Returns true if a sequence could be found, false if end of alignment was
199 * @param ignoreHidden
202 private boolean getSequence(boolean ignoreHidden)
204 AlignmentI alignment = viewport.getAlignment();
205 if (sequenceIndex >= alignment.getHeight())
210 SequenceI seq = alignment.getSequenceAt(sequenceIndex);
211 SequenceGroup selection = viewport.getSelectionGroup();
212 if (selection != null && !selection.contains(seq))
214 if (!nextSequence(ignoreHidden))
218 seq = alignment.getSequenceAt(sequenceIndex);
221 String seqString = null;
224 seqString = getVisibleSequence(seq);
225 this.searchedSequenceStartPosition = 1;
230 int endCol = seq.getLength() - 1;
231 this.searchedSequenceStartPosition = seq.getStart();
232 if (selection != null)
234 startCol = selection.getStartRes();
235 endCol = Math.min(endCol, selection.getEndRes());
236 this.searchedSequenceStartPosition = seq.findPosition(startCol);
238 seqString = seq.getSequenceAsString(startCol, endCol + 1);
242 * remove gaps; note that even if this leaves an empty string, we 'search'
243 * the sequence anyway (for possible match on name or description)
245 String ungapped = AlignSeq.extractGaps(Comparison.GapChars, seqString);
246 this.seqToSearch = ungapped;
252 * Returns a string consisting of only the visible residues of {@code seq} from
253 * alignment column {@ fromColumn}, restricted to the current selection region
256 * As a side-effect, also computes the mapping from the true sequence positions
257 * to the positions (1, 2, ...) of the returned sequence. This is to allow
258 * search matches in the visible sequence to be converted to sequence positions.
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 (sub-)sequence
319 * to be searched is extracted, gaps removed, and saved, or set to null if there
320 * 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,
353 Regex searchPattern, boolean matchDescription,
354 boolean ignoreHidden)
356 if (residueIndex < 0)
359 * at start of sequence; try find by residue number, in sequence id,
360 * or (optionally) in sequence description
362 if (doNonMotifSearches(searchString, searchPattern,
370 * search for next match in sequence string
372 int end = seqToSearch.length();
373 while (residueIndex < end)
375 boolean matched = searchPattern.searchFrom(seqToSearch, residueIndex);
378 if (recordMatch(searchPattern, ignoreHidden))
385 residueIndex = Integer.MAX_VALUE;
389 nextSequence(ignoreHidden);
394 * Adds the match held in the <code>searchPattern</code> Regex to the
395 * <code>searchResults</code>, unless it is a subregion of the last match
396 * recorded. <code>residueIndex</code> is advanced to the position after the
397 * start of the matched region, ready for the next search. Answers true if a
398 * match was added, else false.
400 * Matches that lie entirely within hidden regions of the alignment are not
403 * @param searchPattern
404 * @param ignoreHidden
407 protected boolean recordMatch(Regex searchPattern, boolean ignoreHidden)
409 SequenceI seq = viewport.getAlignment().getSequenceAt(sequenceIndex);
412 * convert start/end of the match to sequence coordinates
414 int offset = searchPattern.matchedFrom();
415 int matchStartPosition = this.searchedSequenceStartPosition + offset;
416 int matchEndPosition = matchStartPosition
417 + searchPattern.charsMatched() - 1;
420 * update residueIndex to next position after the start of the match
421 * (findIndex returns a value base 1, columnIndex is held base 0)
423 residueIndex = searchPattern.matchedFrom()+1;
426 * return false if the match is entirely in a hidden region
428 if (allHidden(seq, matchStartPosition, matchEndPosition))
434 * check that this match is not a subset of the previous one (JAL-2302)
436 List<SearchResultMatchI> matches = searchResults.getResults();
437 SearchResultMatchI lastMatch = matches.isEmpty() ? null
438 : matches.get(matches.size() - 1);
440 if (lastMatch == null || !lastMatch.contains(seq, matchStartPosition,
443 addMatch(seq, matchStartPosition, matchEndPosition, ignoreHidden);
451 * Adds one match to the stored list. If hidden residues are being skipped, then
452 * the match may need to be split into contiguous positions of the sequence (so
453 * it does not include skipped residues).
456 * @param matchStartPosition
457 * @param matchEndPosition
458 * @param ignoreHidden
460 private void addMatch(SequenceI seq, int matchStartPosition,
461 int matchEndPosition, boolean ignoreHidden)
468 searchResults.addResult(seq, matchStartPosition, matchEndPosition);
473 * get start-end contiguous ranges in underlying sequence
475 int[] truePositions = searchedSequenceMap
476 .locateInFrom(matchStartPosition, matchEndPosition);
477 searchResults.addResult(seq, truePositions);
481 * Returns true if all residues are hidden, else false
488 private boolean allHidden(SequenceI seq, int fromPos, int toPos)
490 if (!viewport.hasHiddenColumns())
494 for (int res = fromPos; res <= toPos; res++)
496 if (isVisible(seq, res))
505 * Does searches other than for residue patterns. Currently this includes
507 * <li>find residue by position (if search string is a number)</li>
508 * <li>match search string to sequence id</li>
509 * <li>match search string to sequence description (optional)</li>
511 * Answers true if a match is found, else false.
513 * @param searchString
514 * @param searchPattern
515 * @param includeDescription
518 protected boolean doNonMotifSearches(String searchString,
519 Regex searchPattern, boolean includeDescription)
521 SequenceI seq = viewport.getAlignment().getSequenceAt(sequenceIndex);
524 * position sequence search to start of sequence
529 int res = Integer.parseInt(searchString);
530 return searchForResidueNumber(seq, res);
531 } catch (NumberFormatException ex)
533 // search pattern is not a number
536 if (searchSequenceName(seq, searchPattern))
540 if (includeDescription && searchSequenceDescription(seq, searchPattern))
548 * Searches for a match with the sequence description, and if found, adds the
549 * sequence to the list of match ids (but not as a duplicate). Answers true if
550 * a match was added, else false.
553 * @param searchPattern
556 protected boolean searchSequenceDescription(SequenceI seq, Regex searchPattern)
558 String desc = seq.getDescription();
559 if (desc != null && searchPattern.search(desc) && !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 false.
594 protected boolean searchForResidueNumber(SequenceI seq, int resNo)
596 if (seq.getStart() <= resNo && seq.getEnd() >= resNo)
598 if (isVisible(seq, resNo))
600 searchResults.addResult(seq, resNo, resNo);
608 * Returns true if the residue is in a visible column, else false
614 private boolean isVisible(SequenceI seq, int res)
616 if (!viewport.hasHiddenColumns())
620 int col = seq.findIndex(res); // base 1
621 return viewport.getAlignment().getHiddenColumns().isVisible(col - 1); // base 0
625 public List<SequenceI> getIdMatches()
631 public SearchResultsI getSearchResults()
633 return searchResults;