X-Git-Url: http://source.jalview.org/gitweb/?a=blobdiff_plain;ds=sidebyside;f=src%2Fjalview%2Fanalysis%2FAlignmentSorter.java;h=bb3dbcc21bdf9cbed856d95b6502228554ec6cbb;hb=b56c8f4155ed8cdf730ad0d11cfe95fb4f6af2ba;hp=e9fb3355a40aac50059a6674e1723b18db7d0791;hpb=f4e78f8f0fa25050e2f164d93cf81d8bea69e4e7;p=jalview.git diff --git a/src/jalview/analysis/AlignmentSorter.java b/src/jalview/analysis/AlignmentSorter.java index e9fb335..bb3dbcc 100755 --- a/src/jalview/analysis/AlignmentSorter.java +++ b/src/jalview/analysis/AlignmentSorter.java @@ -29,12 +29,17 @@ import jalview.datamodel.SequenceFeature; import jalview.datamodel.SequenceGroup; import jalview.datamodel.SequenceI; import jalview.datamodel.SequenceNode; -import jalview.util.MessageManager; import jalview.util.QuickSort; import java.util.ArrayList; -import java.util.Arrays; +import java.util.Collections; +import java.util.Iterator; import java.util.List; +import java.util.Map; + +import org.forester.phylogeny.Phylogeny; +import org.forester.phylogeny.PhylogenyNode; +import org.forester.phylogeny.iterators.PhylogenyNodeIterator; /** * Routines for manipulating the order of a multiple sequence alignment TODO: @@ -53,7 +58,7 @@ import java.util.List; */ public class AlignmentSorter { - /** + /* * todo: refactor searches to follow a basic pattern: (search property, last * search state, current sort direction) */ @@ -69,21 +74,22 @@ public class AlignmentSorter static TreeModel lastTree = null; + static Phylogeny lastAptxTree = null; + static boolean sortTreeAscending = true; - /** - * last Annotation Label used by sortByScore + /* + * last Annotation Label used for sort by Annotation score */ - private static String lastSortByScore; + private static String lastSortByAnnotation; - private static boolean sortByScoreAscending = true; - - /** - * compact representation of last arguments to SortByFeatureScore + /* + * string hash of last arguments to sortByFeature + * (sort order toggles if this is unchanged between sorts) */ - private static String lastSortByFeatureScore; + private static String sortByFeatureCriteria; - private static boolean sortByFeatureScoreAscending = true; + private static boolean sortByFeatureAscending = true; private static boolean sortLengthAscending; @@ -109,8 +115,9 @@ public class AlignmentSorter true); for (int i = 0; i < nSeq; i++) { - scores[i] = (float) PIDModel.computePID(align.getSequenceAt(i) - .getSequenceAsString(), refSeq, pidParams); + scores[i] = (float) PIDModel.computePID( + align.getSequenceAt(i).getSequenceAsString(), refSeq, + pidParams); seqs[i] = align.getSequenceAt(i); } @@ -182,7 +189,7 @@ public class AlignmentSorter List algn; synchronized (algn = align.getSequences()) { - List tmp = new ArrayList(); + List tmp = new ArrayList<>(); for (int i = 0; i < seqs.length; i++) { @@ -279,7 +286,7 @@ public class AlignmentSorter { // MAINTAINS ORIGNAL SEQUENCE ORDER, // ORDERS BY GROUP SIZE - List groups = new ArrayList(); + List groups = new ArrayList<>(); if (groups.hashCode() != lastGroupHash) { @@ -315,7 +322,7 @@ public class AlignmentSorter // NOW ADD SEQUENCES MAINTAINING ALIGNMENT ORDER // ///////////////////////////////////////////// - List seqs = new ArrayList(); + List seqs = new ArrayList<>(); for (int i = 0; i < groups.size(); i++) { @@ -357,7 +364,7 @@ public class AlignmentSorter // tmp2 = tmp.retainAll(mask); // return tmp2.addAll(mask.removeAll(tmp2)) - ArrayList seqs = new ArrayList(); + ArrayList seqs = new ArrayList<>(); int i, idx; boolean[] tmask = new boolean[mask.size()]; @@ -416,7 +423,8 @@ public class AlignmentSorter } else { - setReverseOrder(align, vectorSubsetToArray(tmp, align.getSequences())); + setReverseOrder(align, + vectorSubsetToArray(tmp, align.getSequences())); } } @@ -435,9 +443,9 @@ public class AlignmentSorter { int nSeq = align.getHeight(); - List tmp = new ArrayList(); + List tmp = new ArrayList<>(); - tmp = _sortByTree(tree.getTopNode(), tmp, align.getSequences()); + tmp = _sortByTree(tree.getTopNode(), tmp); if (tmp.size() != nSeq) { @@ -451,18 +459,36 @@ public class AlignmentSorter if (tmp.size() != nSeq) { - System.err - .println("WARNING: tmp.size()=" - + tmp.size() - + " != nseq=" - + nSeq - + " in getOrderByTree - tree contains sequences not in alignment"); + System.err.println("WARNING: tmp.size()=" + tmp.size() + " != nseq=" + + nSeq + + " in getOrderByTree - tree contains sequences not in alignment"); } } return tmp; } + + + private static List getOrderByTree(Phylogeny aptxTree, + Map nodesWithBoundSeqs) + { + List seqsByTreeOrder = new ArrayList<>(); + if (!aptxTree.isEmpty()) + { + for (final PhylogenyNodeIterator iter = aptxTree + .iteratorPreorder(); iter.hasNext();) + { + PhylogenyNode treeNode = iter.next(); + seqsByTreeOrder.add(nodesWithBoundSeqs.get(treeNode)); + } + + } + return seqsByTreeOrder; + + + } + /** * Sorts the alignment by a given tree * @@ -492,7 +518,50 @@ public class AlignmentSorter } else { - setReverseOrder(align, vectorSubsetToArray(tmp, align.getSequences())); + setReverseOrder(align, + vectorSubsetToArray(tmp, align.getSequences())); + } + } + + /** + * Sorts the alignment by a given tree from Archaeopteryx + * + * @param align + * alignment to order + * @param tree + * tree which has + */ + public static void sortByTree(AlignmentI align, + Map aptxNodesWithSeqs, + Phylogeny aptxTree) throws IllegalArgumentException + { + List tmp = getOrderByTree(aptxTree, aptxNodesWithSeqs); + + if (!tmp.isEmpty()) + { + if (lastAptxTree != aptxTree) + { + sortTreeAscending = true; + lastAptxTree = aptxTree; + } + else + { + sortTreeAscending = !sortTreeAscending; + } + + if (sortTreeAscending) + { + setOrder(align, tmp); + } + else + { + setReverseOrder(align, + vectorSubsetToArray(tmp, align.getSequences())); + } + } + else + { + throw new IllegalArgumentException(); } } @@ -536,7 +605,7 @@ public class AlignmentSorter * @return DOCUMENT ME! */ private static List _sortByTree(SequenceNode node, - List tmp, List seqset) + List tmp) { if (node == null) { @@ -565,13 +634,15 @@ public class AlignmentSorter } else { - _sortByTree(left, tmp, seqset); - _sortByTree(right, tmp, seqset); + _sortByTree(left, tmp); + _sortByTree(right, tmp); } return tmp; } + + // Ordering Objects // Alignment.sortBy(OrderObj) - sequence of sequence pointer refs in // appropriate order @@ -659,9 +730,9 @@ public class AlignmentSorter } jalview.util.QuickSort.sort(scores, seqs); - if (lastSortByScore != scoreLabel) + if (lastSortByAnnotation != scoreLabel) { - lastSortByScore = scoreLabel; + lastSortByAnnotation = scoreLabel; setOrder(alignment, seqs); } else @@ -682,72 +753,41 @@ public class AlignmentSorter public static String FEATURE_DENSITY = "density"; - private static boolean containsIgnoreCase(final String lab, - final List labs) - { - if (labs == null) - { - return true; - } - if (lab == null) - { - return false; - } - for (String label : labs) - { - if (lab.equalsIgnoreCase(label)) - { - return true; - } - } - return false; - } - - public static void sortByFeature(List featureLabels, - List groupLabels, int start, int stop, + /** + * Sort sequences by feature score or density, optionally restricted by + * feature types, feature groups, or alignment start/end positions. + *

+ * If the sort is repeated for the same combination of types and groups, sort + * order is reversed. + * + * @param featureTypes + * a list of feature types to include (or null for all) + * @param groups + * a list of feature groups to include (or null for all) + * @param startCol + * start column position to include (base zero) + * @param endCol + * end column position to include (base zero) + * @param alignment + * the alignment to be sorted + * @param method + * either "average_score" or "density" ("text" not yet implemented) + */ + public static void sortByFeature(List featureTypes, + List groups, final int startCol, final int endCol, AlignmentI alignment, String method) { if (method != FEATURE_SCORE && method != FEATURE_LABEL && method != FEATURE_DENSITY) { - throw new Error( - MessageManager - .getString("error.implementation_error_sortbyfeature")); - } - - boolean ignoreScore = method != FEATURE_SCORE; - StringBuffer scoreLabel = new StringBuffer(); - scoreLabel.append(start + stop + method); - // This doesn't quite work yet - we'd like to have a canonical ordering that - // can be preserved from call to call - if (featureLabels != null) - { - for (String label : featureLabels) - { - scoreLabel.append(label); - } - } - if (groupLabels != null) - { - for (String label : groupLabels) - { - scoreLabel.append(label); - } + String msg = String + .format("Implementation Error - sortByFeature method must be either '%s' or '%s'", + FEATURE_SCORE, FEATURE_DENSITY); + System.err.println(msg); + return; } - /* - * if resorting the same feature, toggle sort order - */ - if (lastSortByFeatureScore == null - || !scoreLabel.toString().equals(lastSortByFeatureScore)) - { - sortByFeatureScoreAscending = true; - } - else - { - sortByFeatureScoreAscending = !sortByFeatureScoreAscending; - } - lastSortByFeatureScore = scoreLabel.toString(); + flipFeatureSortIfUnchanged(method, featureTypes, groups, startCol, endCol); SequenceI[] seqs = alignment.getSequencesArray(); @@ -756,52 +796,44 @@ public class AlignmentSorter int hasScores = 0; // number of scores present on set double[] scores = new double[seqs.length]; int[] seqScores = new int[seqs.length]; - Object[] feats = new Object[seqs.length]; - double min = 0, max = 0; + Object[][] feats = new Object[seqs.length][]; + double min = 0d; + double max = 0d; + for (int i = 0; i < seqs.length; i++) { - SequenceFeature[] sf = seqs[i].getSequenceFeatures(); - if (sf == null) - { - sf = new SequenceFeature[0]; - } - else - { - SequenceFeature[] tmp = new SequenceFeature[sf.length]; - for (int s = 0; s < tmp.length; s++) - { - tmp[s] = sf[s]; - } - sf = tmp; - } - int sstart = (start == -1) ? start : seqs[i].findPosition(start); - int sstop = (stop == -1) ? stop : seqs[i].findPosition(stop); + /* + * get sequence residues overlapping column region + * and features for residue positions and specified types + */ + String[] types = featureTypes == null ? null : featureTypes + .toArray(new String[featureTypes.size()]); + List sfs = seqs[i].findFeatures(startCol + 1, + endCol + 1, types); + seqScores[i] = 0; scores[i] = 0.0; - int n = sf.length; - for (int f = 0; f < sf.length; f++) + + Iterator it = sfs.listIterator(); + while (it.hasNext()) { - // filter for selection criteria - if ( - // ignore features outwith alignment start-stop positions. - (sf[f].end < sstart || sf[f].begin > sstop) || - // or ignore based on selection criteria - (featureLabels != null && !AlignmentSorter - .containsIgnoreCase(sf[f].type, featureLabels)) - || (groupLabels != null - // problem here: we cannot eliminate null feature group features - && (sf[f].getFeatureGroup() != null && !AlignmentSorter - .containsIgnoreCase(sf[f].getFeatureGroup(), - groupLabels)))) + SequenceFeature sf = it.next(); + + /* + * accept all features with null or empty group, otherwise + * check group is one of the currently visible groups + */ + String featureGroup = sf.getFeatureGroup(); + if (groups != null && featureGroup != null + && !"".equals(featureGroup) + && !groups.contains(featureGroup)) { - // forget about this feature - sf[f] = null; - n--; + it.remove(); } else { - // or, also take a look at the scores if necessary. - if (!ignoreScore && !Float.isNaN(sf[f].getScore())) + float score = sf.getScore(); + if (FEATURE_SCORE.equals(method) && !Float.isNaN(score)) { if (seqScores[i] == 0) { @@ -809,33 +841,26 @@ public class AlignmentSorter } seqScores[i]++; hasScore[i] = true; - scores[i] += sf[f].getScore(); // take the first instance of this - // score. + scores[i] += score; + // take the first instance of this score // ?? } } } - SequenceFeature[] fs; - feats[i] = fs = new SequenceFeature[n]; - if (n > 0) + + feats[i] = sfs.toArray(new SequenceFeature[sfs.size()]); + if (!sfs.isEmpty()) { - n = 0; - for (int f = 0; f < sf.length; f++) - { - if (sf[f] != null) - { - ((SequenceFeature[]) feats[i])[n++] = sf[f]; - } - } if (method == FEATURE_LABEL) { - // order the labels by alphabet - String[] labs = new String[fs.length]; - for (int l = 0; l < labs.length; l++) + // order the labels by alphabet (not yet implemented) + String[] labs = new String[sfs.size()]; + for (int l = 0; l < sfs.size(); l++) { - labs[l] = (fs[l].getDescription() != null ? fs[l] - .getDescription() : fs[l].getType()); + SequenceFeature sf = sfs.get(l); + String description = sf.getDescription(); + labs[l] = (description != null ? description : sf.getType()); } - QuickSort.sort(labs, ((Object[]) feats[i])); + QuickSort.sort(labs, feats[i]); } } if (hasScore[i]) @@ -845,23 +870,18 @@ public class AlignmentSorter // update the score bounds. if (hasScores == 1) { - max = min = scores[i]; + min = scores[i]; + max = min; } else { - if (max < scores[i]) - { - max = scores[i]; - } - if (min > scores[i]) - { - min = scores[i]; - } + max = Math.max(max, scores[i]); + min = Math.min(min, scores[i]); } } } - if (method == FEATURE_SCORE) + if (FEATURE_SCORE.equals(method)) { if (hasScores == 0) { @@ -886,9 +906,9 @@ public class AlignmentSorter } } } - QuickSort.sortByDouble(scores, seqs, sortByFeatureScoreAscending); + QuickSort.sortByDouble(scores, seqs, sortByFeatureAscending); } - else if (method == FEATURE_DENSITY) + else if (FEATURE_DENSITY.equals(method)) { for (int i = 0; i < seqs.length; i++) { @@ -898,18 +918,53 @@ public class AlignmentSorter // System.err.println("Sorting on Density: seq "+seqs[i].getName()+ // " Feats: "+featureCount+" Score : "+scores[i]); } - QuickSort.sortByDouble(scores, seqs, sortByFeatureScoreAscending); + QuickSort.sortByDouble(scores, seqs, sortByFeatureAscending); } - else + + setOrder(alignment, seqs); + } + + /** + * Builds a string hash of criteria for sorting, and if unchanged from last + * time, reverse the sort order + * + * @param method + * @param featureTypes + * @param groups + * @param startCol + * @param endCol + */ + protected static void flipFeatureSortIfUnchanged(String method, + List featureTypes, List groups, + final int startCol, final int endCol) + { + StringBuilder sb = new StringBuilder(64); + sb.append(startCol).append(method).append(endCol); + if (featureTypes != null) { - if (method == FEATURE_LABEL) - { - throw new Error( - MessageManager.getString("error.not_yet_implemented")); - } + Collections.sort(featureTypes); + sb.append(featureTypes.toString()); + } + if (groups != null) + { + Collections.sort(groups); + sb.append(groups.toString()); } + String scoreCriteria = sb.toString(); - setOrder(alignment, seqs); + /* + * if resorting on the same criteria, toggle sort order + */ + if (sortByFeatureCriteria == null + || !scoreCriteria.equals(sortByFeatureCriteria)) + { + sortByFeatureAscending = true; + } + else + { + sortByFeatureAscending = !sortByFeatureAscending; + } + sortByFeatureCriteria = scoreCriteria; } }