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<int[]>();
81 toShifts = new ArrayList<int[]>();
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 hashCode = 31 * hashCode + fromShifts.toArray().hashCode();
120 hashCode = 31 * hashCode + toShifts.toArray().hashCode();
125 * Returns the 'from' ranges as {[start1, end1], [start2, end2], ...}
129 public List<int[]> getFromRanges()
135 * Returns the 'to' ranges as {[start1, end1], [start2, end2], ...}
139 public List<int[]> getToRanges()
145 * Flattens a list of [start, end] into a single [start1, end1, start2,
151 protected static int[] getRanges(List<int[]> shifts)
153 int[] rnges = new int[2 * shifts.size()];
155 for (int[] r : shifts)
165 * @return length of mapped phrase in from
167 public int getFromRatio()
174 * @return length of mapped phrase in to
176 public int getToRatio()
181 public int getFromLowest()
186 public int getFromHighest()
191 public int getToLowest()
196 public int getToHighest()
202 * Constructor given from and to ranges as [start1, end1, start2, end2,...].
203 * If any end is equal to the next start, the ranges will be merged. There is
204 * no validation check that the ranges do not overlap each other.
207 * contiguous regions as [start1, end1, start2, end2, ...]
209 * same format as 'from'
211 * phrase length in 'from' (e.g. 3 for dna)
213 * phrase length in 'to' (e.g. 1 for protein)
215 public MapList(int from[], int to[], int fromRatio, int toRatio)
218 this.fromRatio = fromRatio;
219 this.toRatio = toRatio;
220 fromLowest = Integer.MAX_VALUE;
221 fromHighest = Integer.MIN_VALUE;
224 for (int i = 0; i < from.length; i += 2)
227 * note lowest and highest values - bearing in mind the
228 * direction may be reversed
230 fromLowest = Math.min(fromLowest, Math.min(from[i], from[i + 1]));
231 fromHighest = Math.max(fromHighest, Math.max(from[i], from[i + 1]));
232 if (added > 0 && from[i] == fromShifts.get(added - 1)[1])
235 * this range starts where the last ended - just extend it
237 fromShifts.get(added - 1)[1] = from[i + 1];
241 fromShifts.add(new int[] { from[i], from[i + 1] });
246 toLowest = Integer.MAX_VALUE;
247 toHighest = Integer.MIN_VALUE;
249 for (int i = 0; i < to.length; i += 2)
251 toLowest = Math.min(toLowest, Math.min(to[i], to[i + 1]));
252 toHighest = Math.max(toHighest, Math.max(to[i], to[i + 1]));
253 if (added > 0 && to[i] == toShifts.get(added - 1)[1])
255 toShifts.get(added - 1)[1] = to[i + 1];
259 toShifts.add(new int[] { to[i], to[i + 1] });
266 * Copy constructor. Creates an identical mapping.
270 public MapList(MapList map)
273 // TODO not used - remove?
274 this.fromLowest = map.fromLowest;
275 this.fromHighest = map.fromHighest;
276 this.toLowest = map.toLowest;
277 this.toHighest = map.toHighest;
279 this.fromRatio = map.fromRatio;
280 this.toRatio = map.toRatio;
281 if (map.fromShifts != null)
283 for (int[] r : map.fromShifts)
285 fromShifts.add(new int[] { r[0], r[1] });
288 if (map.toShifts != null)
290 for (int[] r : map.toShifts)
292 toShifts.add(new int[] { r[0], r[1] });
298 * Constructor given ranges as lists of [start, end] positions. There is no
299 * validation check that the ranges do not overlap each other.
306 public MapList(List<int[]> fromRange, List<int[]> toRange, int fromRatio,
310 fromRange = coalesceRanges(fromRange);
311 toRange = coalesceRanges(toRange);
312 this.fromShifts = fromRange;
313 this.toShifts = toRange;
314 this.fromRatio = fromRatio;
315 this.toRatio = toRatio;
317 fromLowest = Integer.MAX_VALUE;
318 fromHighest = Integer.MIN_VALUE;
319 for (int[] range : fromRange)
321 fromLowest = Math.min(fromLowest, Math.min(range[0], range[1]));
322 fromHighest = Math.max(fromHighest, Math.max(range[0], range[1]));
325 toLowest = Integer.MAX_VALUE;
326 toHighest = Integer.MIN_VALUE;
327 for (int[] range : toRange)
329 toLowest = Math.min(toLowest, Math.min(range[0], range[1]));
330 toHighest = Math.max(toHighest, Math.max(range[0], range[1]));
335 * Consolidates a list of ranges so that any contiguous ranges are merged.
336 * This assumes the ranges are already in start order (does not sort them).
339 * @return the same list (if unchanged), else a new merged list, leaving the
340 * input list unchanged
342 public static List<int[]> coalesceRanges(final List<int[]> ranges)
344 if (ranges == null || ranges.size() < 2)
349 boolean changed = false;
350 List<int[]> merged = new ArrayList<int[]>();
351 int[] lastRange = ranges.get(0);
352 int lastDirection = lastRange[1] >= lastRange[0] ? 1 : -1;
353 lastRange = new int[] { lastRange[0], lastRange[1] };
354 merged.add(lastRange);
355 boolean first = true;
357 for (final int[] range : ranges)
364 if (range[0] == lastRange[0] && range[1] == lastRange[1])
366 // drop duplicate range
372 * drop this range if it lies within the last range
374 if ((lastDirection == 1 && range[0] >= lastRange[0]
375 && range[0] <= lastRange[1] && range[1] >= lastRange[0]
376 && range[1] <= lastRange[1])
377 || (lastDirection == -1 && range[0] <= lastRange[0]
378 && range[0] >= lastRange[1]
379 && range[1] <= lastRange[0]
380 && range[1] >= lastRange[1]))
386 int direction = range[1] >= range[0] ? 1 : -1;
389 * if next range is in the same direction as last and contiguous,
390 * just update the end position of the last range
392 boolean sameDirection = range[1] == range[0]
393 || direction == lastDirection;
394 boolean extending = range[0] == lastRange[1] + lastDirection;
395 boolean overlapping = (lastDirection == 1 && range[0] >= lastRange[0]
396 && range[0] <= lastRange[1])
397 || (lastDirection == -1 && range[0] <= lastRange[0]
398 && range[0] >= lastRange[1]);
399 if (sameDirection && (overlapping || extending))
401 lastRange[1] = range[1];
406 lastRange = new int[] { range[0], range[1] };
407 merged.add(lastRange);
408 // careful: merging [5, 5] after [7, 6] should keep negative direction
409 lastDirection = (range[1] == range[0]) ? lastDirection : direction;
413 return changed ? merged : ranges;
417 * get all mapped positions from 'from' to 'to'
419 * @return int[][] { int[] { fromStart, fromFinish, toStart, toFinish }, int
420 * [fromFinish-fromStart+2] { toStart..toFinish mappings}}
422 protected int[][] makeFromMap()
424 // TODO not used - remove??
425 return posMap(fromShifts, fromRatio, toShifts, toRatio);
429 * get all mapped positions from 'to' to 'from'
431 * @return int[to position]=position mapped in from
433 protected int[][] makeToMap()
435 // TODO not used - remove??
436 return posMap(toShifts, toRatio, fromShifts, fromRatio);
440 * construct an int map for intervals in intVals
443 * @return int[] { from, to pos in range }, int[range.to-range.from+1]
444 * returning mapped position
446 private int[][] posMap(List<int[]> shiftTo, int ratio,
447 List<int[]> shiftFrom, int toRatio)
449 // TODO not used - remove??
450 int iv = 0, ivSize = shiftTo.size();
455 int[] intv = shiftTo.get(iv++);
456 int from = intv[0], to = intv[1];
464 intv = shiftTo.get(iv++);
483 int mp[][] = new int[to - from + 2][];
484 for (int i = 0; i < mp.length; i++)
486 int[] m = shift(i + from, shiftTo, ratio, shiftFrom, toRatio);
507 int[][] map = new int[][] { new int[] { from, to, tF, tT },
508 new int[to - from + 2] };
513 for (int i = 0; i < mp.length; i++)
517 map[1][i] = mp[i][0] - tF;
521 map[1][i] = -1; // indicates an out of range mapping
531 * start position for shift (in original reference frame)
535 * public void addShift(int pos, int shift) { int sidx = 0; int[]
536 * rshift=null; while (sidx<shifts.size() && (rshift=(int[])
537 * shifts.elementAt(sidx))[0]<pos) sidx++; if (sidx==shifts.size())
538 * shifts.insertElementAt(new int[] { pos, shift}, sidx); else
539 * rshift[1]+=shift; }
543 * shift from pos to To(pos)
547 * @return int shifted position in To, frameshift in From, direction of mapped
550 public int[] shiftFrom(int pos)
552 return shift(pos, fromShifts, fromRatio, toShifts, toRatio);
556 * inverse of shiftFrom - maps pos in To to a position in From
560 * @return shifted position in From, frameshift in To, direction of mapped
563 public int[] shiftTo(int pos)
565 return shift(pos, toShifts, toRatio, fromShifts, fromRatio);
576 protected static int[] shift(int pos, List<int[]> shiftTo, int fromRatio,
577 List<int[]> shiftFrom, int toRatio)
579 // TODO: javadoc; tests
580 int[] fromCount = countPos(shiftTo, pos);
581 if (fromCount == null)
585 int fromRemainder = (fromCount[0] - 1) % fromRatio;
586 int toCount = 1 + (((fromCount[0] - 1) / fromRatio) * toRatio);
587 int[] toPos = countToPos(shiftFrom, toCount);
590 return null; // throw new Error("Bad Mapping!");
592 // System.out.println(fromCount[0]+" "+fromCount[1]+" "+toCount);
593 return new int[] { toPos[0], fromRemainder, toPos[1] };
597 * count how many positions pos is along the series of intervals.
601 * @return number of positions or null if pos is not within intervals
603 protected static int[] countPos(List<int[]> shiftTo, int pos)
605 int count = 0, intv[], iv = 0, ivSize = shiftTo.size();
608 intv = shiftTo.get(iv++);
609 if (intv[0] <= intv[1])
611 if (pos >= intv[0] && pos <= intv[1])
613 return new int[] { count + pos - intv[0] + 1, +1 };
617 count += intv[1] - intv[0] + 1;
622 if (pos >= intv[1] && pos <= intv[0])
624 return new int[] { count + intv[0] - pos + 1, -1 };
628 count += intv[0] - intv[1] + 1;
636 * count out pos positions into a series of intervals and return the position
640 * @return position pos in interval set
642 protected static int[] countToPos(List<int[]> shiftFrom, int pos)
644 int count = 0, diff = 0, iv = 0, ivSize = shiftFrom.size();
645 int[] intv = { 0, 0 };
648 intv = shiftFrom.get(iv++);
649 diff = intv[1] - intv[0];
652 if (pos <= count + 1 + diff)
654 return new int[] { pos - count - 1 + intv[0], +1 };
663 if (pos <= count + 1 - diff)
665 return new int[] { intv[0] - (pos - count - 1), -1 };
673 return null;// (diff<0) ? (intv[1]-1) : (intv[0]+1);
677 * find series of intervals mapping from start-end in the From map.
680 * position mapped 'to'
682 * position mapped 'to'
683 * @return series of [start, end] ranges in sequence mapped 'from'
685 public int[] locateInFrom(int start, int end)
687 // inefficient implementation
688 int fromStart[] = shiftTo(start);
689 // needs to be inclusive of end of symbol position
690 int fromEnd[] = shiftTo(end);
692 return getIntervals(fromShifts, fromStart, fromEnd, fromRatio);
696 * find series of intervals mapping from start-end in the to map.
699 * position mapped 'from'
701 * position mapped 'from'
702 * @return series of [start, end] ranges in sequence mapped 'to'
704 public int[] locateInTo(int start, int end)
706 int toStart[] = shiftFrom(start);
707 int toEnd[] = shiftFrom(end);
708 return getIntervals(toShifts, toStart, toEnd, toRatio);
712 * like shift - except returns the intervals in the given vector of shifts
713 * which were spanned in traversing fromStart to fromEnd
719 * @return series of from,to intervals from from first position of starting
720 * region to final position of ending region inclusive
722 protected static int[] getIntervals(List<int[]> shiftFrom,
723 int[] fromStart, int[] fromEnd, int fromRatio2)
725 if (fromStart == null || fromEnd == null)
729 int startpos, endpos;
730 startpos = fromStart[0]; // first position in fromStart
731 endpos = fromEnd[0]; // last position in fromEnd
732 int endindx = (fromRatio2 - 1); // additional positions to get to last
733 // position from endpos
734 int intv = 0, intvSize = shiftFrom.size();
735 int iv[], i = 0, fs = -1, fe_s = -1, fe = -1; // containing intervals
736 // search intervals to locate ones containing startpos and count endindx
737 // positions on from endpos
738 while (intv < intvSize && (fs == -1 || fe == -1))
740 iv = shiftFrom.get(intv++);
743 endpos = iv[0]; // start counting from beginning of interval
744 endindx--; // inclusive of endpos
748 if (fs == -1 && startpos >= iv[0] && startpos <= iv[1])
752 if (endpos >= iv[0] && endpos <= iv[1])
760 if (endpos + endindx <= iv[1])
763 endpos = endpos + endindx; // end of end token is within this
768 endindx -= iv[1] - endpos; // skip all this interval too
775 if (fs == -1 && startpos <= iv[0] && startpos >= iv[1])
779 if (endpos <= iv[0] && endpos >= iv[1])
787 if (endpos - endindx >= iv[1])
790 endpos = endpos - endindx; // end of end token is within this
795 endindx -= endpos - iv[1]; // skip all this interval too
802 if (fs == fe && fe == -1)
806 List<int[]> ranges = new ArrayList<int[]>();
811 // truncate initial interval
812 iv = shiftFrom.get(intv++);
813 iv = new int[] { iv[0], iv[1] };// clone
820 ranges.add(iv); // add initial range
821 iv = shiftFrom.get(intv++); // get next interval
822 iv = new int[] { iv[0], iv[1] };// clone
829 ranges.add(iv); // add only - or final range
833 // walk from end of interval.
834 i = shiftFrom.size() - 1;
839 iv = shiftFrom.get(i);
840 iv = new int[] { iv[1], iv[0] };// reverse and clone
841 // truncate initial interval
847 { // fix apparent logic bug when fe==-1
848 ranges.add(iv); // add (truncated) reversed interval
849 iv = shiftFrom.get(i);
850 iv = new int[] { iv[1], iv[0] }; // reverse and clone
854 // interval is already reversed
857 ranges.add(iv); // add only - or final range
859 // create array of start end intervals.
861 if (ranges != null && ranges.size() > 0)
863 range = new int[ranges.size() * 2];
865 intvSize = ranges.size();
867 while (intv < intvSize)
869 iv = ranges.get(intv);
872 ranges.set(intv++, null); // remove
879 * get the 'initial' position of mpos in To
883 * @return position of first word in to reference frame
885 public int getToPosition(int mpos)
887 // TODO not used - remove??
888 int[] mp = shiftTo(mpos);
897 * get range of positions in To frame for the mpos word in From
901 * @return null or int[] first position in To for mpos, last position in to
904 public int[] getToWord(int mpos)
906 int[] mp = shiftTo(mpos);
909 return new int[] { mp[0], mp[0] + mp[2] * (getFromRatio() - 1) };
915 * get From position in the associated reference frame for position pos in the
916 * associated sequence.
921 public int getMappedPosition(int pos)
923 // TODO not used - remove??
924 int[] mp = shiftFrom(pos);
932 public int[] getMappedWord(int pos)
934 // TODO not used - remove??
935 int[] mp = shiftFrom(pos);
938 return new int[] { mp[0], mp[0] + mp[2] * (getToRatio() - 1) };
945 * @return a MapList whose From range is this maplist's To Range, and vice
948 public MapList getInverse()
950 return new MapList(getToRanges(), getFromRanges(), getToRatio(),
955 * test for containment rather than equivalence to another mapping
958 * to be tested for containment
959 * @return true if local or mapped range map contains or is contained by this
962 public boolean containsEither(boolean local, MapList map)
964 // TODO not used - remove?
967 return ((getFromLowest() >= map.getFromLowest()
968 && getFromHighest() <= map.getFromHighest())
969 || (getFromLowest() <= map.getFromLowest()
970 && getFromHighest() >= map.getFromHighest()));
974 return ((getToLowest() >= map.getToLowest()
975 && getToHighest() <= map.getToHighest())
976 || (getToLowest() <= map.getToLowest()
977 && getToHighest() >= map.getToHighest()));
982 * String representation - for debugging, not guaranteed not to change
985 public String toString()
987 StringBuilder sb = new StringBuilder(64);
989 for (int[] shift : fromShifts)
991 sb.append(" ").append(Arrays.toString(shift));
994 sb.append(fromRatio).append(":").append(toRatio);
996 for (int[] shift : toShifts)
998 sb.append(" ").append(Arrays.toString(shift));
1001 return sb.toString();
1005 * Extend this map list by adding the given map's ranges. There is no
1006 * validation check that the ranges do not overlap existing ranges (or each
1007 * other), but contiguous ranges are merged.
1011 public void addMapList(MapList map)
1013 if (this.equals(map))
1017 this.fromLowest = Math.min(fromLowest, map.fromLowest);
1018 this.toLowest = Math.min(toLowest, map.toLowest);
1019 this.fromHighest = Math.max(fromHighest, map.fromHighest);
1020 this.toHighest = Math.max(toHighest, map.toHighest);
1022 for (int[] range : map.getFromRanges())
1024 addRange(range, fromShifts);
1026 for (int[] range : map.getToRanges())
1028 addRange(range, toShifts);
1033 * Adds the given range to a list of ranges. If the new range just extends
1034 * existing ranges, the current endpoint is updated instead.
1039 static void addRange(int[] range, List<int[]> addTo)
1042 * list is empty - add to it!
1044 if (addTo.size() == 0)
1050 int[] last = addTo.get(addTo.size() - 1);
1051 boolean lastForward = last[1] >= last[0];
1052 boolean newForward = range[1] >= range[0];
1055 * contiguous range in the same direction - just update endpoint
1057 if (lastForward == newForward && last[1] == range[0])
1064 * next range starts at +1 in forward sense - update endpoint
1066 if (lastForward && newForward && range[0] == last[1] + 1)
1073 * next range starts at -1 in reverse sense - update endpoint
1075 if (!lastForward && !newForward && range[0] == last[1] - 1)
1082 * just add the new range
1088 * Returns true if mapping is from forward strand, false if from reverse
1089 * strand. Result is just based on the first 'from' range that is not a single
1090 * position. Default is true unless proven to be false. Behaviour is not well
1091 * defined if the mapping has a mixture of forward and reverse ranges.
1095 public boolean isFromForwardStrand()
1097 boolean forwardStrand = true;
1098 for (int[] range : getFromRanges())
1100 if (range[1] > range[0])
1102 break; // forward strand confirmed
1104 else if (range[1] < range[0])
1106 forwardStrand = false;
1107 break; // reverse strand confirmed
1110 return forwardStrand;
1115 * @return true if from, or to is a three to 1 mapping
1117 public boolean isTripletMap()
1119 return (toRatio == 3 && fromRatio == 1)
1120 || (fromRatio == 3 && toRatio == 1);
1124 * Returns a map which is the composite of this one and the input map. That
1125 * is, the output map has the fromRanges of this map, and its toRanges are the
1126 * toRanges of this map as transformed by the input map.
1128 * Returns null if the mappings cannot be traversed (not all toRanges of this
1129 * map correspond to fromRanges of the input), or if this.toRatio does not
1130 * match map.fromRatio.
1134 * this: from [1-100] to [501-600]
1135 * input: from [10-40] to [60-90]
1136 * output: from [10-40] to [560-590]
1137 * Example 2 ('reverse strand exons'):
1138 * this: from [1-100] to [2000-1951], [1000-951] // transcript to loci
1139 * input: from [1-50] to [41-90] // CDS to transcript
1140 * output: from [10-40] to [1960-1951], [1000-971] // CDS to gene loci
1146 public MapList traverse(MapList map)
1154 * compound the ratios by this rule:
1155 * A:B with M:N gives A*M:B*N
1156 * reduced by greatest common divisor
1157 * so 1:3 with 3:3 is 3:9 or 1:3
1158 * 1:3 with 3:1 is 3:3 or 1:1
1159 * 1:3 with 1:3 is 1:9
1160 * 2:5 with 3:7 is 6:35
1162 int outFromRatio = getFromRatio() * map.getFromRatio();
1163 int outToRatio = getToRatio() * map.getToRatio();
1164 int gcd = MathUtils.gcd(outFromRatio, outToRatio);
1165 outFromRatio /= gcd;
1168 List<int[]> toRanges = new ArrayList<>();
1169 for (int[] range : getToRanges())
1171 int[] transferred = map.locateInTo(range[0], range[1]);
1172 if (transferred == null)
1176 toRanges.add(transferred);
1179 return new MapList(getFromRanges(), toRanges, outFromRatio, outToRatio);