From ace99d5556a3f546b98d32f73d7cf78cbe9bcb32 Mon Sep 17 00:00:00 2001 From: BobHanson Date: Fri, 5 Jun 2020 09:30:55 -0500 Subject: [PATCH] JAL-3446 AlignmentSorter/Test more documentation; bit cleaner code --- src/jalview/analysis/AlignmentSorter.java | 203 ++++++++++++++++------------- 1 file changed, 112 insertions(+), 91 deletions(-) diff --git a/src/jalview/analysis/AlignmentSorter.java b/src/jalview/analysis/AlignmentSorter.java index af9b5df..af7db0a 100755 --- a/src/jalview/analysis/AlignmentSorter.java +++ b/src/jalview/analysis/AlignmentSorter.java @@ -34,6 +34,7 @@ import jalview.datamodel.SequenceNode; import jalview.util.QuickSort; import java.util.ArrayList; +import java.util.BitSet; import java.util.Collections; import java.util.Iterator; import java.util.List; @@ -97,7 +98,6 @@ public class AlignmentSorter implements ApplicationSingletonI boolean sortTreeAscending = true; - /** * last Annotation Label used for sort by Annotation score */ @@ -167,7 +167,7 @@ public class AlignmentSorter implements ApplicationSingletonI QuickSort.sort(ids, seqs); AlignmentSorter as = getInstance(); as.sortIdAscending = !as.sortIdAscending; - sort(align, seqs, as.sortIdAscending); + set(align, seqs, as.sortIdAscending); } /** @@ -192,7 +192,7 @@ public class AlignmentSorter implements ApplicationSingletonI QuickSort.sort(length, seqs); AlignmentSorter as = getInstance(); as.sortLengthAscending = !as.sortLengthAscending; - sort(align, seqs, as.sortLengthAscending); + set(align, seqs, as.sortLengthAscending); } /** @@ -257,56 +257,7 @@ public class AlignmentSorter implements ApplicationSingletonI tmp.add(orderedseqs[j]); } } - sort(align, tmp, as.sortGroupAscending); - } - - /** - * Select sequences in order from tmp that is present in mask, and any - * remaining sequences in mask not in tmp - * - * @param tmp - * thread safe collection of sequences - * @param mask - * thread safe collection of sequences - * - * @return intersect(tmp,mask)+intersect(complement(tmp),mask) - */ - private static SequenceI[] vectorSubsetToArray(List tmp, - List mask) - { - // or? - // tmp2 = tmp.retainAll(mask); - // return tmp2.addAll(mask.removeAll(tmp2)) - - ArrayList seqs = new ArrayList<>(); - int i, idx; - boolean[] tmask = new boolean[mask.size()]; - - for (i = 0; i < mask.size(); i++) - { - tmask[i] = true; - } - - for (i = 0; i < tmp.size(); i++) - { - SequenceI sq = tmp.get(i); - idx = mask.indexOf(sq); - if (idx > -1 && tmask[idx]) - { - tmask[idx] = false; - seqs.add(sq); - } - } - - for (i = 0; i < tmask.length; i++) - { - if (tmask[i]) - { - seqs.add(mask.get(i)); - } - } - - return seqs.toArray(new SequenceI[seqs.size()]); + set(align, tmp, as.sortGroupAscending); } /** @@ -332,7 +283,7 @@ public class AlignmentSorter implements ApplicationSingletonI { as.sortOrderAscending = true; } - sort(align, tmp, as.sortOrderAscending); + set(align, tmp, as.sortOrderAscending); } /** @@ -399,7 +350,7 @@ public class AlignmentSorter implements ApplicationSingletonI { as.sortTreeAscending = !as.sortTreeAscending; } - sort(align, tmp, as.sortTreeAscending); + set(align, tmp, as.sortTreeAscending); } /** @@ -606,14 +557,15 @@ public class AlignmentSorter implements ApplicationSingletonI if (method != FEATURE_SCORE && method != FEATURE_LABEL && method != FEATURE_DENSITY) { - String msg = String - .format("Implementation Error - sortByFeature method must be either '%s' or '%s'", - FEATURE_SCORE, FEATURE_DENSITY); + String msg = String.format( + "Implementation Error - sortByFeature method must be either '%s' or '%s'", + FEATURE_SCORE, FEATURE_DENSITY); System.err.println(msg); return; } - flipFeatureSortIfUnchanged(method, featureTypes, groups, startCol, endCol); + flipFeatureSortIfUnchanged(method, featureTypes, groups, startCol, + endCol); SequenceI[] seqs = alignment.getSequencesArray(); @@ -632,8 +584,8 @@ public class AlignmentSorter implements ApplicationSingletonI * 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()]); + String[] types = featureTypes == null ? null + : featureTypes.toArray(new String[featureTypes.size()]); List sfs = seqs[i].findFeatures(startCol + 1, endCol + 1, types); @@ -750,7 +702,8 @@ public class AlignmentSorter implements ApplicationSingletonI } if (doSort) { - QuickSort.sortByDouble(scores, seqs, getInstance().sortByFeatureAscending); + QuickSort.sortByDouble(scores, seqs, + getInstance().sortByFeatureAscending); } setOrder(alignment, seqs); } @@ -799,13 +752,38 @@ public class AlignmentSorter implements ApplicationSingletonI as.sortByFeatureCriteria = scoreCriteria; } - private static void sort(AlignmentI align, List tmp, + /** + * Set the alignment's sequences list to contain the sequences from a + * temporary list, first adding all the elements from the tmp list, then adding all sequences in the alignment that + * are not in the list. Option to do the final sort either in order or in reverse order. + * + * @param align The alignment being sorted + * @param tmp + * the temporary sequence list + * @param ascending + * false for reversed order; only sequences already in + * the alignment will be used (which is actually already guaranteed + * by vectorSubsetToArray) + */ + private static void set(AlignmentI align, List tmp, boolean ascending) { - sort(align, vectorSubsetToArray(tmp, align.getSequences()), ascending); + set(align, vectorSubsetToArray(align.getSequences(), tmp), ascending); } - private static void sort(AlignmentI align, SequenceI[] seqs, + /** + * Set the alignment's sequences list to contain these sequences, either in + * this order or its reverse. + * + * @param align + * @param seqs + * the new sequence array + * @param ascending + * false for reversed order; if ascending, only sequences already in + * the alignment will be used; if descending, then a direct 1:1 + * replacement is made + */ + private static void set(AlignmentI align, SequenceI[] seqs, boolean ascending) { if (ascending) @@ -819,76 +797,119 @@ public class AlignmentSorter implements ApplicationSingletonI } - /** - * Sets the Alignment object with the given sequences + * Replace the alignment's sequences with values in an array, clearing the + * alignment's sequence list and filtering for sequences that are actually in + * the alignment already. * * @param align - * DOCUMENT ME! + * the Alignment * @param seqs - * sequences as an array + * the array of replacement values, of any length */ public static void setOrder(AlignmentI align, SequenceI[] seqs) { // NOTE: DO NOT USE align.setSequenceAt() here - it will NOT work - List algn = align.getSequences(); - synchronized (algn) + List seqList = align.getSequences(); + synchronized (seqList) { List tmp = new ArrayList<>(); for (int i = 0; i < seqs.length; i++) { - if (algn.contains(seqs[i])) + if (seqList.contains(seqs[i])) { tmp.add(seqs[i]); } } - algn.clear(); + seqList.clear(); // User may have hidden seqs, then clicked undo or redo for (int i = 0; i < tmp.size(); i++) { - algn.add(tmp.get(i)); + seqList.add(tmp.get(i)); } } } /** - * Reverse the order of the sort + * Replace the alignment's sequences or a subset of those sequences with + * values in an array in reverse order. All sequences are replaced; no check + * is made that these sequences are in the alignment already. * * @param align - * DOCUMENT ME! + * the Alignment * @param seqs - * DOCUMENT ME! + * the array of replacement values, length must be less than or equal + * to Alignment.sequences.size() */ private static void setReverseOrder(AlignmentI align, SequenceI[] seqs) { int nSeq = seqs.length; int len = (nSeq + (nSeq % 2)) / 2; -// int len = 0; -// -// if ((nSeq % 2) == 0) -// { -// len = nSeq / 2; -// } -// else -// { -// len = (nSeq + 1) / 2; -// } + // int len = 0; + // + // if ((nSeq % 2) == 0) + // { + // len = nSeq / 2; + // } + // else + // { + // len = (nSeq + 1) / 2; + // } // NOTE: DO NOT USE align.setSequenceAt() here - it will NOT work - List asq = align.getSequences(); - synchronized (asq) + List seqList = align.getSequences(); + synchronized (seqList) { for (int i = 0; i < len; i++) { // SequenceI tmp = seqs[i]; - asq.set(i, seqs[nSeq - i - 1]); - asq.set(nSeq - i - 1, seqs[i]); + seqList.set(i, seqs[nSeq - i - 1]); + seqList.set(nSeq - i - 1, seqs[i]); } } } + /** + * Create and array of reordered sequences in order first from tmp that are + * present in seqList already, then, after that, any remaining sequences in + * seqList not in tmp. Any sequences in tmp that are not in seqList already + * are discarded. + * + * @param seqList + * thread safe collection of sequences originally in the alignment + * @param tmp + * thread safe collection of sequences or subsequences possibly in + * seqList + * + * @return intersect(tmp,seqList)+intersect(complement(tmp),seqList) + */ + private static SequenceI[] vectorSubsetToArray(List seqList, + List tmp) + { + ArrayList seqs = new ArrayList<>(); + int n = seqList.size(); + BitSet bs = new BitSet(n); + bs.set(0, n); + for (int i = 0, nt = tmp.size(); i < nt; i++) + { + SequenceI sq = tmp.get(i); + int idx = seqList.indexOf(sq); + if (idx >= 0 && bs.get(idx)) + { + seqs.add(sq); + bs.clear(idx); + } + } + + for (int i = bs.nextSetBit(0); i >= 0; i = bs.nextSetBit(i + 1)) + { + seqs.add(seqList.get(i)); + } + + return seqs.toArray(new SequenceI[seqs.size()]); + } } -- 1.7.10.2