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.
23 import java.util.ArrayList;
24 import java.util.Arrays;
25 import java.util.List;
28 * A simple way of bijectively mapping a non-contiguous linear range to another
29 * non-contiguous linear range.
31 * Use at your own risk!
33 * TODO: efficient implementation of private posMap method
35 * TODO: test/ensure that sense of from and to ratio start position is conserved
36 * (codon start position recovery)
42 * Subregions (base 1) described as { [start1, end1], [start2, end2], ...}
44 private List<int[]> fromShifts;
47 * Same format as fromShifts, for the 'mapped to' sequence
49 private List<int[]> toShifts;
52 * number of steps in fromShifts to one toRatio unit
54 private int fromRatio;
57 * number of steps in toShifts to one fromRatio
62 * lowest and highest value in the from Map
64 private int fromLowest;
66 private int fromHighest;
69 * lowest and highest value in the to Map
73 private int toHighest;
80 fromShifts = new ArrayList<>();
81 toShifts = new ArrayList<>();
85 * Two MapList objects are equal if they are the same object, or they both
86 * have populated shift ranges and all values are the same.
89 public boolean equals(Object o)
91 if (o == null || !(o instanceof MapList))
96 MapList obj = (MapList) o;
101 if (obj.fromRatio != fromRatio || obj.toRatio != toRatio
102 || obj.fromShifts == null || obj.toShifts == null)
106 return Arrays.deepEquals(fromShifts.toArray(), obj.fromShifts.toArray())
107 && Arrays.deepEquals(toShifts.toArray(),
108 obj.toShifts.toArray());
112 * Returns a hashcode made from the fromRatio, toRatio, and from/to ranges
115 public int hashCode()
117 int hashCode = 31 * fromRatio;
118 hashCode = 31 * hashCode + toRatio;
119 for (int[] shift : fromShifts)
121 hashCode = 31 * hashCode + shift[0];
122 hashCode = 31 * hashCode + shift[1];
124 for (int[] shift : toShifts)
126 hashCode = 31 * hashCode + shift[0];
127 hashCode = 31 * hashCode + shift[1];
134 * Returns the 'from' ranges as {[start1, end1], [start2, end2], ...}
138 public List<int[]> getFromRanges()
144 * Returns the 'to' ranges as {[start1, end1], [start2, end2], ...}
148 public List<int[]> getToRanges()
154 * Flattens a list of [start, end] into a single [start1, end1, start2,
160 protected static int[] getRanges(List<int[]> shifts)
162 int[] rnges = new int[2 * shifts.size()];
164 for (int[] r : shifts)
174 * @return length of mapped phrase in from
176 public int getFromRatio()
183 * @return length of mapped phrase in to
185 public int getToRatio()
190 public int getFromLowest()
195 public int getFromHighest()
200 public int getToLowest()
205 public int getToHighest()
211 * Constructor given from and to ranges as [start1, end1, start2, end2,...].
212 * If any end is equal to the next start, the ranges will be merged. There is
213 * no validation check that the ranges do not overlap each other.
216 * contiguous regions as [start1, end1, start2, end2, ...]
218 * same format as 'from'
220 * phrase length in 'from' (e.g. 3 for dna)
222 * phrase length in 'to' (e.g. 1 for protein)
224 public MapList(int from[], int to[], int fromRatio, int toRatio)
227 this.fromRatio = fromRatio;
228 this.toRatio = toRatio;
229 fromLowest = Integer.MAX_VALUE;
230 fromHighest = Integer.MIN_VALUE;
233 for (int i = 0; i < from.length; i += 2)
236 * note lowest and highest values - bearing in mind the
237 * direction may be reversed
239 fromLowest = Math.min(fromLowest, Math.min(from[i], from[i + 1]));
240 fromHighest = Math.max(fromHighest, Math.max(from[i], from[i + 1]));
241 if (added > 0 && from[i] == fromShifts.get(added - 1)[1])
244 * this range starts where the last ended - just extend it
246 fromShifts.get(added - 1)[1] = from[i + 1];
250 fromShifts.add(new int[] { from[i], from[i + 1] });
255 toLowest = Integer.MAX_VALUE;
256 toHighest = Integer.MIN_VALUE;
258 for (int i = 0; i < to.length; i += 2)
260 toLowest = Math.min(toLowest, Math.min(to[i], to[i + 1]));
261 toHighest = Math.max(toHighest, Math.max(to[i], to[i + 1]));
262 if (added > 0 && to[i] == toShifts.get(added - 1)[1])
264 toShifts.get(added - 1)[1] = to[i + 1];
268 toShifts.add(new int[] { to[i], to[i + 1] });
275 * Copy constructor. Creates an identical mapping.
279 public MapList(MapList map)
282 // TODO not used - remove?
283 this.fromLowest = map.fromLowest;
284 this.fromHighest = map.fromHighest;
285 this.toLowest = map.toLowest;
286 this.toHighest = map.toHighest;
288 this.fromRatio = map.fromRatio;
289 this.toRatio = map.toRatio;
290 if (map.fromShifts != null)
292 for (int[] r : map.fromShifts)
294 fromShifts.add(new int[] { r[0], r[1] });
297 if (map.toShifts != null)
299 for (int[] r : map.toShifts)
301 toShifts.add(new int[] { r[0], r[1] });
307 * Constructor given ranges as lists of [start, end] positions. There is no
308 * validation check that the ranges do not overlap each other.
315 public MapList(List<int[]> fromRange, List<int[]> toRange, int fromRatio,
319 fromRange = coalesceRanges(fromRange);
320 toRange = coalesceRanges(toRange);
321 this.fromShifts = fromRange;
322 this.toShifts = toRange;
323 this.fromRatio = fromRatio;
324 this.toRatio = toRatio;
326 fromLowest = Integer.MAX_VALUE;
327 fromHighest = Integer.MIN_VALUE;
328 for (int[] range : fromRange)
330 fromLowest = Math.min(fromLowest, Math.min(range[0], range[1]));
331 fromHighest = Math.max(fromHighest, Math.max(range[0], range[1]));
334 toLowest = Integer.MAX_VALUE;
335 toHighest = Integer.MIN_VALUE;
336 for (int[] range : toRange)
338 toLowest = Math.min(toLowest, Math.min(range[0], range[1]));
339 toHighest = Math.max(toHighest, Math.max(range[0], range[1]));
344 * Consolidates a list of ranges so that any contiguous ranges are merged.
345 * This assumes the ranges are already in start order (does not sort them).
348 * @return the same list (if unchanged), else a new merged list, leaving the
349 * input list unchanged
351 public static List<int[]> coalesceRanges(final List<int[]> ranges)
353 if (ranges == null || ranges.size() < 2)
358 boolean changed = false;
359 List<int[]> merged = new ArrayList<>();
360 int[] lastRange = ranges.get(0);
361 int lastDirection = lastRange[1] >= lastRange[0] ? 1 : -1;
362 lastRange = new int[] { lastRange[0], lastRange[1] };
363 merged.add(lastRange);
364 boolean first = true;
366 for (final int[] range : ranges)
373 if (range[0] == lastRange[0] && range[1] == lastRange[1])
375 // drop duplicate range
381 * drop this range if it lies within the last range
383 if ((lastDirection == 1 && range[0] >= lastRange[0]
384 && range[0] <= lastRange[1] && range[1] >= lastRange[0]
385 && range[1] <= lastRange[1])
386 || (lastDirection == -1 && range[0] <= lastRange[0]
387 && range[0] >= lastRange[1]
388 && range[1] <= lastRange[0]
389 && range[1] >= lastRange[1]))
395 int direction = range[1] >= range[0] ? 1 : -1;
398 * if next range is in the same direction as last and contiguous,
399 * just update the end position of the last range
401 boolean sameDirection = range[1] == range[0]
402 || direction == lastDirection;
403 boolean extending = range[0] == lastRange[1] + lastDirection;
404 boolean overlapping = (lastDirection == 1 && range[0] >= lastRange[0]
405 && range[0] <= lastRange[1])
406 || (lastDirection == -1 && range[0] <= lastRange[0]
407 && range[0] >= lastRange[1]);
408 if (sameDirection && (overlapping || extending))
410 lastRange[1] = range[1];
415 lastRange = new int[] { range[0], range[1] };
416 merged.add(lastRange);
417 // careful: merging [5, 5] after [7, 6] should keep negative direction
418 lastDirection = (range[1] == range[0]) ? lastDirection : direction;
422 return changed ? merged : ranges;
426 * get all mapped positions from 'from' to 'to'
428 * @return int[][] { int[] { fromStart, fromFinish, toStart, toFinish }, int
429 * [fromFinish-fromStart+2] { toStart..toFinish mappings}}
431 protected int[][] makeFromMap()
433 // TODO not used - remove??
434 return posMap(fromShifts, fromRatio, toShifts, toRatio);
438 * get all mapped positions from 'to' to 'from'
440 * @return int[to position]=position mapped in from
442 protected int[][] makeToMap()
444 // TODO not used - remove??
445 return posMap(toShifts, toRatio, fromShifts, fromRatio);
449 * construct an int map for intervals in intVals
452 * @return int[] { from, to pos in range }, int[range.to-range.from+1]
453 * returning mapped position
455 private int[][] posMap(List<int[]> shiftTo, int ratio,
456 List<int[]> shiftFrom, int toRatio)
458 // TODO not used - remove??
459 int iv = 0, ivSize = shiftTo.size();
464 int[] intv = shiftTo.get(iv++);
465 int from = intv[0], to = intv[1];
473 intv = shiftTo.get(iv++);
492 int mp[][] = new int[to - from + 2][];
493 for (int i = 0; i < mp.length; i++)
495 int[] m = shift(i + from, shiftTo, ratio, shiftFrom, toRatio);
516 int[][] map = new int[][] { new int[] { from, to, tF, tT },
517 new int[to - from + 2] };
522 for (int i = 0; i < mp.length; i++)
526 map[1][i] = mp[i][0] - tF;
530 map[1][i] = -1; // indicates an out of range mapping
540 * start position for shift (in original reference frame)
544 * public void addShift(int pos, int shift) { int sidx = 0; int[]
545 * rshift=null; while (sidx<shifts.size() && (rshift=(int[])
546 * shifts.elementAt(sidx))[0]<pos) sidx++; if (sidx==shifts.size())
547 * shifts.insertElementAt(new int[] { pos, shift}, sidx); else
548 * rshift[1]+=shift; }
552 * shift from pos to To(pos)
556 * @return int shifted position in To, frameshift in From, direction of mapped
559 public int[] shiftFrom(int pos)
561 return shift(pos, fromShifts, fromRatio, toShifts, toRatio);
565 * inverse of shiftFrom - maps pos in To to a position in From
569 * @return shifted position in From, frameshift in To, direction of mapped
572 public int[] shiftTo(int pos)
574 return shift(pos, toShifts, toRatio, fromShifts, fromRatio);
585 protected static int[] shift(int pos, List<int[]> shiftTo, int fromRatio,
586 List<int[]> shiftFrom, int toRatio)
588 // TODO: javadoc; tests
589 int[] fromCount = countPos(shiftTo, pos);
590 if (fromCount == null)
594 int fromRemainder = (fromCount[0] - 1) % fromRatio;
595 int toCount = 1 + (((fromCount[0] - 1) / fromRatio) * toRatio);
596 int[] toPos = countToPos(shiftFrom, toCount);
599 return null; // throw new Error("Bad Mapping!");
601 // System.out.println(fromCount[0]+" "+fromCount[1]+" "+toCount);
602 return new int[] { toPos[0], fromRemainder, toPos[1] };
606 * count how many positions pos is along the series of intervals.
610 * @return number of positions or null if pos is not within intervals
612 protected static int[] countPos(List<int[]> shiftTo, int pos)
614 int count = 0, intv[], iv = 0, ivSize = shiftTo.size();
617 intv = shiftTo.get(iv++);
618 if (intv[0] <= intv[1])
620 if (pos >= intv[0] && pos <= intv[1])
622 return new int[] { count + pos - intv[0] + 1, +1 };
626 count += intv[1] - intv[0] + 1;
631 if (pos >= intv[1] && pos <= intv[0])
633 return new int[] { count + intv[0] - pos + 1, -1 };
637 count += intv[0] - intv[1] + 1;
645 * count out pos positions into a series of intervals and return the position
649 * @return position pos in interval set
651 protected static int[] countToPos(List<int[]> shiftFrom, int pos)
653 int count = 0, diff = 0, iv = 0, ivSize = shiftFrom.size();
654 int[] intv = { 0, 0 };
657 intv = shiftFrom.get(iv++);
658 diff = intv[1] - intv[0];
661 if (pos <= count + 1 + diff)
663 return new int[] { pos - count - 1 + intv[0], +1 };
672 if (pos <= count + 1 - diff)
674 return new int[] { intv[0] - (pos - count - 1), -1 };
682 return null;// (diff<0) ? (intv[1]-1) : (intv[0]+1);
686 * find series of intervals mapping from start-end in the From map.
689 * position mapped 'to'
691 * position mapped 'to'
692 * @return series of [start, end] ranges in sequence mapped 'from'
694 public int[] locateInFrom(int start, int end)
696 // inefficient implementation
697 int fromStart[] = shiftTo(start);
698 // needs to be inclusive of end of symbol position
699 int fromEnd[] = shiftTo(end);
701 return getIntervals(fromShifts, fromStart, fromEnd, fromRatio);
705 * find series of intervals mapping from start-end in the to map.
708 * position mapped 'from'
710 * position mapped 'from'
711 * @return series of [start, end] ranges in sequence mapped 'to'
713 public int[] locateInTo(int start, int end)
715 int toStart[] = shiftFrom(start);
716 int toEnd[] = shiftFrom(end);
717 return getIntervals(toShifts, toStart, toEnd, toRatio);
721 * like shift - except returns the intervals in the given vector of shifts
722 * which were spanned in traversing fromStart to fromEnd
728 * @return series of from,to intervals from from first position of starting
729 * region to final position of ending region inclusive
731 protected static int[] getIntervals(List<int[]> shiftFrom,
732 int[] fromStart, int[] fromEnd, int fromRatio2)
734 if (fromStart == null || fromEnd == null)
738 int startpos, endpos;
739 startpos = fromStart[0]; // first position in fromStart
740 endpos = fromEnd[0]; // last position in fromEnd
741 int endindx = (fromRatio2 - 1); // additional positions to get to last
742 // position from endpos
743 int intv = 0, intvSize = shiftFrom.size();
744 int iv[], i = 0, fs = -1, fe_s = -1, fe = -1; // containing intervals
745 // search intervals to locate ones containing startpos and count endindx
746 // positions on from endpos
747 while (intv < intvSize && (fs == -1 || fe == -1))
749 iv = shiftFrom.get(intv++);
752 endpos = iv[0]; // start counting from beginning of interval
753 endindx--; // inclusive of endpos
757 if (fs == -1 && startpos >= iv[0] && startpos <= iv[1])
761 if (endpos >= iv[0] && endpos <= iv[1])
769 if (endpos + endindx <= iv[1])
772 endpos = endpos + endindx; // end of end token is within this
777 endindx -= iv[1] - endpos; // skip all this interval too
784 if (fs == -1 && startpos <= iv[0] && startpos >= iv[1])
788 if (endpos <= iv[0] && endpos >= iv[1])
796 if (endpos - endindx >= iv[1])
799 endpos = endpos - endindx; // end of end token is within this
804 endindx -= endpos - iv[1]; // skip all this interval too
811 if (fs == fe && fe == -1)
815 List<int[]> ranges = new ArrayList<>();
820 // truncate initial interval
821 iv = shiftFrom.get(intv++);
822 iv = new int[] { iv[0], iv[1] };// clone
829 ranges.add(iv); // add initial range
830 iv = shiftFrom.get(intv++); // get next interval
831 iv = new int[] { iv[0], iv[1] };// clone
838 ranges.add(iv); // add only - or final range
842 // walk from end of interval.
843 i = shiftFrom.size() - 1;
848 iv = shiftFrom.get(i);
849 iv = new int[] { iv[1], iv[0] };// reverse and clone
850 // truncate initial interval
856 { // fix apparent logic bug when fe==-1
857 ranges.add(iv); // add (truncated) reversed interval
858 iv = shiftFrom.get(i);
859 iv = new int[] { iv[1], iv[0] }; // reverse and clone
863 // interval is already reversed
866 ranges.add(iv); // add only - or final range
868 // create array of start end intervals.
870 if (ranges != null && ranges.size() > 0)
872 range = new int[ranges.size() * 2];
874 intvSize = ranges.size();
876 while (intv < intvSize)
878 iv = ranges.get(intv);
881 ranges.set(intv++, null); // remove
888 * get the 'initial' position of mpos in To
892 * @return position of first word in to reference frame
894 public int getToPosition(int mpos)
896 // TODO not used - remove??
897 int[] mp = shiftTo(mpos);
906 * get range of positions in To frame for the mpos word in From
910 * @return null or int[] first position in To for mpos, last position in to
913 public int[] getToWord(int mpos)
915 int[] mp = shiftTo(mpos);
918 return new int[] { mp[0], mp[0] + mp[2] * (getFromRatio() - 1) };
924 * get From position in the associated reference frame for position pos in the
925 * associated sequence.
930 public int getMappedPosition(int pos)
932 // TODO not used - remove??
933 int[] mp = shiftFrom(pos);
941 public int[] getMappedWord(int pos)
943 // TODO not used - remove??
944 int[] mp = shiftFrom(pos);
947 return new int[] { mp[0], mp[0] + mp[2] * (getToRatio() - 1) };
954 * @return a MapList whose From range is this maplist's To Range, and vice
957 public MapList getInverse()
959 return new MapList(getToRanges(), getFromRanges(), getToRatio(),
964 * test for containment rather than equivalence to another mapping
967 * to be tested for containment
968 * @return true if local or mapped range map contains or is contained by this
971 public boolean containsEither(boolean local, MapList map)
973 // TODO not used - remove?
976 return ((getFromLowest() >= map.getFromLowest()
977 && getFromHighest() <= map.getFromHighest())
978 || (getFromLowest() <= map.getFromLowest()
979 && getFromHighest() >= map.getFromHighest()));
983 return ((getToLowest() >= map.getToLowest()
984 && getToHighest() <= map.getToHighest())
985 || (getToLowest() <= map.getToLowest()
986 && getToHighest() >= map.getToHighest()));
991 * String representation - for debugging, not guaranteed not to change
994 public String toString()
996 StringBuilder sb = new StringBuilder(64);
998 for (int[] shift : fromShifts)
1000 sb.append(" ").append(Arrays.toString(shift));
1003 sb.append(fromRatio).append(":").append(toRatio);
1005 for (int[] shift : toShifts)
1007 sb.append(" ").append(Arrays.toString(shift));
1010 return sb.toString();
1014 * Extend this map list by adding the given map's ranges. There is no
1015 * validation check that the ranges do not overlap existing ranges (or each
1016 * other), but contiguous ranges are merged.
1020 public void addMapList(MapList map)
1022 if (this.equals(map))
1026 this.fromLowest = Math.min(fromLowest, map.fromLowest);
1027 this.toLowest = Math.min(toLowest, map.toLowest);
1028 this.fromHighest = Math.max(fromHighest, map.fromHighest);
1029 this.toHighest = Math.max(toHighest, map.toHighest);
1031 for (int[] range : map.getFromRanges())
1033 addRange(range, fromShifts);
1035 for (int[] range : map.getToRanges())
1037 addRange(range, toShifts);
1042 * Adds the given range to a list of ranges. If the new range just extends
1043 * existing ranges, the current endpoint is updated instead.
1048 static void addRange(int[] range, List<int[]> addTo)
1051 * list is empty - add to it!
1053 if (addTo.size() == 0)
1059 int[] last = addTo.get(addTo.size() - 1);
1060 boolean lastForward = last[1] >= last[0];
1061 boolean newForward = range[1] >= range[0];
1064 * contiguous range in the same direction - just update endpoint
1066 if (lastForward == newForward && last[1] == range[0])
1073 * next range starts at +1 in forward sense - update endpoint
1075 if (lastForward && newForward && range[0] == last[1] + 1)
1082 * next range starts at -1 in reverse sense - update endpoint
1084 if (!lastForward && !newForward && range[0] == last[1] - 1)
1091 * just add the new range
1097 * Returns true if mapping is from forward strand, false if from reverse
1098 * strand. Result is just based on the first 'from' range that is not a single
1099 * position. Default is true unless proven to be false. Behaviour is not well
1100 * defined if the mapping has a mixture of forward and reverse ranges.
1104 public boolean isFromForwardStrand()
1106 boolean forwardStrand = true;
1107 for (int[] range : getFromRanges())
1109 if (range[1] > range[0])
1111 break; // forward strand confirmed
1113 else if (range[1] < range[0])
1115 forwardStrand = false;
1116 break; // reverse strand confirmed
1119 return forwardStrand;
1124 * @return true if from, or to is a three to 1 mapping
1126 public boolean isTripletMap()
1128 return (toRatio == 3 && fromRatio == 1)
1129 || (fromRatio == 3 && toRatio == 1);