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 * There is no validation check that the ranges do not overlap each other.
215 * contiguous regions as [start1, end1, start2, end2, ...]
217 * same format as 'from'
219 * phrase length in 'from' (e.g. 3 for dna)
221 * phrase length in 'to' (e.g. 1 for protein)
223 public MapList(int from[], int to[], int fromRatio, int toRatio)
226 this.fromRatio = fromRatio;
227 this.toRatio = toRatio;
228 fromLowest = Integer.MAX_VALUE;
229 fromHighest = Integer.MIN_VALUE;
231 for (int i = 0; i < from.length; i += 2)
234 * note lowest and highest values - bearing in mind the
235 * direction may be reversed
237 fromLowest = Math.min(fromLowest, Math.min(from[i], from[i + 1]));
238 fromHighest = Math.max(fromHighest, Math.max(from[i], from[i + 1]));
239 fromShifts.add(new int[] { from[i], from[i + 1] });
242 toLowest = Integer.MAX_VALUE;
243 toHighest = Integer.MIN_VALUE;
244 for (int i = 0; i < to.length; i += 2)
246 toLowest = Math.min(toLowest, Math.min(to[i], to[i + 1]));
247 toHighest = Math.max(toHighest, Math.max(to[i], to[i + 1]));
248 toShifts.add(new int[] { to[i], to[i + 1] });
253 * Copy constructor. Creates an identical mapping.
257 public MapList(MapList map)
260 // TODO not used - remove?
261 this.fromLowest = map.fromLowest;
262 this.fromHighest = map.fromHighest;
263 this.toLowest = map.toLowest;
264 this.toHighest = map.toHighest;
266 this.fromRatio = map.fromRatio;
267 this.toRatio = map.toRatio;
268 if (map.fromShifts != null)
270 for (int[] r : map.fromShifts)
272 fromShifts.add(new int[] { r[0], r[1] });
275 if (map.toShifts != null)
277 for (int[] r : map.toShifts)
279 toShifts.add(new int[] { r[0], r[1] });
285 * Constructor given ranges as lists of [start, end] positions. There is no
286 * validation check that the ranges do not overlap each other.
293 public MapList(List<int[]> fromRange, List<int[]> toRange, int fromRatio,
297 fromRange = coalesceRanges(fromRange);
298 toRange = coalesceRanges(toRange);
299 this.fromShifts = fromRange;
300 this.toShifts = toRange;
301 this.fromRatio = fromRatio;
302 this.toRatio = toRatio;
304 fromLowest = Integer.MAX_VALUE;
305 fromHighest = Integer.MIN_VALUE;
306 for (int[] range : fromRange)
308 if (range.length != 2)
310 // throw new IllegalArgumentException(range);
311 System.err.println("Invalid format for fromRange "
312 + Arrays.toString(range) + " may cause errors");
314 fromLowest = Math.min(fromLowest, Math.min(range[0], range[1]));
315 fromHighest = Math.max(fromHighest, Math.max(range[0], range[1]));
318 toLowest = Integer.MAX_VALUE;
319 toHighest = Integer.MIN_VALUE;
320 for (int[] range : toRange)
322 if (range.length != 2)
324 // throw new IllegalArgumentException(range);
325 System.err.println("Invalid format for toRange "
326 + Arrays.toString(range) + " may cause errors");
328 toLowest = Math.min(toLowest, Math.min(range[0], range[1]));
329 toHighest = Math.max(toHighest, Math.max(range[0], range[1]));
334 * Consolidates a list of ranges so that any contiguous ranges are merged.
335 * This assumes the ranges are already in start order (does not sort them).
337 * The main use case for this method is when mapping cDNA sequence to its
338 * protein product, based on CDS feature ranges which derive from spliced
339 * exons, but are contiguous on the cDNA sequence. For example
342 * CDS 1-20 // from exon1
343 * CDS 21-35 // from exon2
344 * CDS 36-71 // from exon3
345 * 'coalesce' to range 1-71
349 * @return the same list (if unchanged), else a new merged list, leaving the
350 * input list unchanged
352 public static List<int[]> coalesceRanges(final List<int[]> ranges)
354 if (ranges == null || ranges.size() < 2)
359 boolean changed = false;
360 List<int[]> merged = new ArrayList<>();
361 int[] lastRange = ranges.get(0);
362 int lastDirection = lastRange[1] >= lastRange[0] ? 1 : -1;
363 lastRange = new int[] { lastRange[0], lastRange[1] };
364 merged.add(lastRange);
365 boolean first = true;
367 for (final int[] range : ranges)
375 int direction = range[1] >= range[0] ? 1 : -1;
378 * if next range is in the same direction as last and contiguous,
379 * just update the end position of the last range
381 boolean sameDirection = range[1] == range[0]
382 || direction == lastDirection;
383 boolean extending = range[0] == lastRange[1] + lastDirection;
384 if (sameDirection && extending)
386 lastRange[1] = range[1];
391 lastRange = new int[] { range[0], range[1] };
392 merged.add(lastRange);
393 // careful: merging [5, 5] after [7, 6] should keep negative direction
394 lastDirection = (range[1] == range[0]) ? lastDirection : direction;
398 return changed ? merged : ranges;
402 * get all mapped positions from 'from' to 'to'
404 * @return int[][] { int[] { fromStart, fromFinish, toStart, toFinish }, int
405 * [fromFinish-fromStart+2] { toStart..toFinish mappings}}
407 protected int[][] makeFromMap()
409 // TODO not used - remove??
410 return posMap(fromShifts, fromRatio, toShifts, toRatio);
414 * get all mapped positions from 'to' to 'from'
416 * @return int[to position]=position mapped in from
418 protected int[][] makeToMap()
420 // TODO not used - remove??
421 return posMap(toShifts, toRatio, fromShifts, fromRatio);
425 * construct an int map for intervals in intVals
428 * @return int[] { from, to pos in range }, int[range.to-range.from+1]
429 * returning mapped position
431 private int[][] posMap(List<int[]> shiftTo, int ratio,
432 List<int[]> shiftFrom, int toRatio)
434 // TODO not used - remove??
435 int iv = 0, ivSize = shiftTo.size();
440 int[] intv = shiftTo.get(iv++);
441 int from = intv[0], to = intv[1];
449 intv = shiftTo.get(iv++);
468 int mp[][] = new int[to - from + 2][];
469 for (int i = 0; i < mp.length; i++)
471 int[] m = shift(i + from, shiftTo, ratio, shiftFrom, toRatio);
492 int[][] map = new int[][] { new int[] { from, to, tF, tT },
493 new int[to - from + 2] };
498 for (int i = 0; i < mp.length; i++)
502 map[1][i] = mp[i][0] - tF;
506 map[1][i] = -1; // indicates an out of range mapping
516 * start position for shift (in original reference frame)
520 * public void addShift(int pos, int shift) { int sidx = 0; int[]
521 * rshift=null; while (sidx<shifts.size() && (rshift=(int[])
522 * shifts.elementAt(sidx))[0]<pos) sidx++; if (sidx==shifts.size())
523 * shifts.insertElementAt(new int[] { pos, shift}, sidx); else
524 * rshift[1]+=shift; }
528 * shift from pos to To(pos)
532 * @return int shifted position in To, frameshift in From, direction of mapped
535 public int[] shiftFrom(int pos)
537 return shift(pos, fromShifts, fromRatio, toShifts, toRatio);
541 * inverse of shiftFrom - maps pos in To to a position in From
545 * @return shifted position in From, frameshift in To, direction of mapped
548 public int[] shiftTo(int pos)
550 return shift(pos, toShifts, toRatio, fromShifts, fromRatio);
561 protected static int[] shift(int pos, List<int[]> shiftTo, int fromRatio,
562 List<int[]> shiftFrom, int toRatio)
564 // TODO: javadoc; tests
565 int[] fromCount = countPositions(shiftTo, pos);
566 if (fromCount == null)
570 int fromRemainder = (fromCount[0] - 1) % fromRatio;
571 int toCount = 1 + (((fromCount[0] - 1) / fromRatio) * toRatio);
572 int[] toPos = traverseToPosition(shiftFrom, toCount);
577 return new int[] { toPos[0], fromRemainder, toPos[1] };
581 * Counts how many positions pos is along the series of intervals. Returns an
582 * array of two values:
584 * <li>the number of positions traversed (inclusive) to reach {@code pos}</li>
585 * <li>+1 if the last interval traversed is forward, -1 if in a negative
588 * Returns null if {@code pos} does not lie in any of the given intervals.
591 * a list of start-end intervals
593 * a position that may lie in one (or more) of the intervals
596 protected static int[] countPositions(List<int[]> intervals, int pos)
600 int ivSize = intervals.size();
604 int[] intv = intervals.get(iv++);
605 if (intv[0] <= intv[1])
610 if (pos >= intv[0] && pos <= intv[1])
612 return new int[] { count + pos - intv[0] + 1, +1 };
616 count += intv[1] - intv[0] + 1;
624 if (pos >= intv[1] && pos <= intv[0])
626 return new int[] { count + intv[0] - pos + 1, -1 };
630 count += intv[0] - intv[1] + 1;
638 * Reads through the given intervals until {@code count} positions have been
639 * traversed, and returns an array consisting of two values:
641 * <li>the value at the {@code count'th} position</li>
642 * <li>+1 if the last interval read is forwards, -1 if reverse direction</li>
644 * Returns null if the ranges include less than {@code count} positions, or if
648 * a list of [start, end] ranges
650 * the number of positions to traverse
653 protected static int[] traverseToPosition(List<int[]> intervals,
657 int ivSize = intervals.size();
667 int[] intv = intervals.get(iv++);
668 int diff = intv[1] - intv[0];
671 if (count <= traversed + 1 + diff)
673 return new int[] { intv[0] + (count - traversed - 1), +1 };
677 traversed += 1 + diff;
682 if (count <= traversed + 1 - diff)
684 return new int[] { intv[0] - (count - traversed - 1), -1 };
688 traversed += 1 - diff;
696 * find series of intervals mapping from start-end in the From map.
699 * position mapped 'to'
701 * position mapped 'to'
702 * @return series of [start, end] ranges in sequence mapped 'from'
704 public int[] locateInFrom(int start, int end)
706 return locateInFrom2(start, end);
708 // inefficient implementation
709 // int fromStart[] = shiftTo(start);
710 // needs to be inclusive of end of symbol position
711 // int fromEnd[] = shiftTo(end);
712 // return getIntervals(fromShifts, fromStart, fromEnd, fromRatio);
716 * find series of intervals mapping from start-end in the to map.
719 * position mapped 'from'
721 * position mapped 'from'
722 * @return series of [start, end] ranges in sequence mapped 'to'
724 public int[] locateInTo(int start, int end)
726 return locateInTo2(start, end);
728 // int toStart[] = shiftFrom(start);
729 // int toEnd[] = shiftFrom(end);
730 // return getIntervals(toShifts, toStart, toEnd, toRatio);
734 * like shift - except returns the intervals in the given vector of shifts
735 * which were spanned in traversing fromStart to fromEnd
741 * @return series of from,to intervals from from first position of starting
742 * region to final position of ending region inclusive
744 protected static int[] getIntervals(List<int[]> shiftFrom,
745 int[] fromStart, int[] fromEnd, int fromRatio2)
747 if (fromStart == null || fromEnd == null)
751 int startpos, endpos;
752 startpos = fromStart[0]; // first position in fromStart
753 endpos = fromEnd[0]; // last position in fromEnd
754 int endindx = (fromRatio2 - 1); // additional positions to get to last
755 // position from endpos
756 int intv = 0, intvSize = shiftFrom.size();
757 int iv[], i = 0, fs = -1, fe_s = -1, fe = -1; // containing intervals
758 // search intervals to locate ones containing startpos and count endindx
759 // positions on from endpos
760 while (intv < intvSize && (fs == -1 || fe == -1))
762 iv = shiftFrom.get(intv++);
765 endpos = iv[0]; // start counting from beginning of interval
766 endindx--; // inclusive of endpos
770 if (fs == -1 && startpos >= iv[0] && startpos <= iv[1])
774 if (endpos >= iv[0] && endpos <= iv[1])
782 if (endpos + endindx <= iv[1])
785 endpos = endpos + endindx; // end of end token is within this
790 endindx -= iv[1] - endpos; // skip all this interval too
797 if (fs == -1 && startpos <= iv[0] && startpos >= iv[1])
801 if (endpos <= iv[0] && endpos >= iv[1])
809 if (endpos - endindx >= iv[1])
812 endpos = endpos - endindx; // end of end token is within this
817 endindx -= endpos - iv[1]; // skip all this interval too
824 if (fs == fe && fe == -1)
828 List<int[]> ranges = new ArrayList<>();
833 // truncate initial interval
834 iv = shiftFrom.get(intv++);
835 iv = new int[] { iv[0], iv[1] };// clone
842 ranges.add(iv); // add initial range
843 iv = shiftFrom.get(intv++); // get next interval
844 iv = new int[] { iv[0], iv[1] };// clone
851 ranges.add(iv); // add only - or final range
855 // walk from end of interval.
856 i = shiftFrom.size() - 1;
861 iv = shiftFrom.get(i);
862 iv = new int[] { iv[1], iv[0] };// reverse and clone
863 // truncate initial interval
869 { // fix apparent logic bug when fe==-1
870 ranges.add(iv); // add (truncated) reversed interval
871 iv = shiftFrom.get(i);
872 iv = new int[] { iv[1], iv[0] }; // reverse and clone
876 // interval is already reversed
879 ranges.add(iv); // add only - or final range
881 // create array of start end intervals.
883 if (ranges != null && ranges.size() > 0)
885 range = new int[ranges.size() * 2];
887 intvSize = ranges.size();
889 while (intv < intvSize)
891 iv = ranges.get(intv);
894 ranges.set(intv++, null); // remove
901 * get the 'initial' position of mpos in To
905 * @return position of first word in to reference frame
907 public int getToPosition(int mpos)
909 // TODO not used - remove??
910 int[] mp = shiftTo(mpos);
919 * get range of positions in To frame for the mpos word in From
923 * @return null or int[] first position in To for mpos, last position in to
926 public int[] getToWord(int mpos)
928 int[] mp = shiftTo(mpos);
931 return new int[] { mp[0], mp[0] + mp[2] * (getFromRatio() - 1) };
937 * get From position in the associated reference frame for position pos in the
938 * associated sequence.
943 public int getMappedPosition(int pos)
945 // TODO not used - remove??
946 int[] mp = shiftFrom(pos);
954 public int[] getMappedWord(int pos)
956 // TODO not used - remove??
957 int[] mp = shiftFrom(pos);
960 return new int[] { mp[0], mp[0] + mp[2] * (getToRatio() - 1) };
967 * @return a MapList whose From range is this maplist's To Range, and vice
970 public MapList getInverse()
972 return new MapList(getToRanges(), getFromRanges(), getToRatio(),
977 * test for containment rather than equivalence to another mapping
980 * to be tested for containment
981 * @return true if local or mapped range map contains or is contained by this
984 public boolean containsEither(boolean local, MapList map)
986 // TODO not used - remove?
989 return ((getFromLowest() >= map.getFromLowest()
990 && getFromHighest() <= map.getFromHighest())
991 || (getFromLowest() <= map.getFromLowest()
992 && getFromHighest() >= map.getFromHighest()));
996 return ((getToLowest() >= map.getToLowest()
997 && getToHighest() <= map.getToHighest())
998 || (getToLowest() <= map.getToLowest()
999 && getToHighest() >= map.getToHighest()));
1004 * String representation - for debugging, not guaranteed not to change
1007 public String toString()
1009 StringBuilder sb = new StringBuilder(64);
1011 for (int[] shift : fromShifts)
1013 sb.append(" ").append(Arrays.toString(shift));
1016 sb.append(fromRatio).append(":").append(toRatio);
1018 for (int[] shift : toShifts)
1020 sb.append(" ").append(Arrays.toString(shift));
1023 return sb.toString();
1027 * Extend this map list by adding the given map's ranges. There is no
1028 * validation check that the ranges do not overlap existing ranges (or each
1029 * other), but contiguous ranges are merged.
1033 public void addMapList(MapList map)
1035 if (this.equals(map))
1039 this.fromLowest = Math.min(fromLowest, map.fromLowest);
1040 this.toLowest = Math.min(toLowest, map.toLowest);
1041 this.fromHighest = Math.max(fromHighest, map.fromHighest);
1042 this.toHighest = Math.max(toHighest, map.toHighest);
1044 for (int[] range : map.getFromRanges())
1046 addRange(range, fromShifts);
1048 for (int[] range : map.getToRanges())
1050 addRange(range, toShifts);
1055 * Adds the given range to a list of ranges. If the new range just extends
1056 * existing ranges, the current endpoint is updated instead.
1061 static void addRange(int[] range, List<int[]> addTo)
1064 * list is empty - add to it!
1066 if (addTo.size() == 0)
1072 int[] last = addTo.get(addTo.size() - 1);
1073 boolean lastForward = last[1] >= last[0];
1074 boolean newForward = range[1] >= range[0];
1077 * contiguous range in the same direction - just update endpoint
1079 if (lastForward == newForward && last[1] == range[0])
1086 * next range starts at +1 in forward sense - update endpoint
1088 if (lastForward && newForward && range[0] == last[1] + 1)
1095 * next range starts at -1 in reverse sense - update endpoint
1097 if (!lastForward && !newForward && range[0] == last[1] - 1)
1104 * just add the new range
1110 * Returns true if mapping is from forward strand, false if from reverse
1111 * strand. Result is just based on the first 'from' range that is not a single
1112 * position. Default is true unless proven to be false. Behaviour is not well
1113 * defined if the mapping has a mixture of forward and reverse ranges.
1117 public boolean isFromForwardStrand()
1119 return isForwardStrand(getFromRanges());
1123 * Returns true if mapping is to forward strand, false if to reverse strand.
1124 * Result is just based on the first 'to' range that is not a single position.
1125 * Default is true unless proven to be false. Behaviour is not well defined if
1126 * the mapping has a mixture of forward and reverse ranges.
1130 public boolean isToForwardStrand()
1132 return isForwardStrand(getToRanges());
1136 * A helper method that returns true unless at least one range has start >
1137 * end. Behaviour is undefined for a mixture of forward and reverse ranges.
1142 private boolean isForwardStrand(List<int[]> ranges)
1144 boolean forwardStrand = true;
1145 for (int[] range : ranges)
1147 if (range[1] > range[0])
1149 break; // forward strand confirmed
1151 else if (range[1] < range[0])
1153 forwardStrand = false;
1154 break; // reverse strand confirmed
1157 return forwardStrand;
1162 * @return true if from, or to is a three to 1 mapping
1164 public boolean isTripletMap()
1166 return (toRatio == 3 && fromRatio == 1)
1167 || (fromRatio == 3 && toRatio == 1);
1171 * Returns a map which is the composite of this one and the input map. That
1172 * is, the output map has the fromRanges of this map, and its toRanges are the
1173 * toRanges of this map as transformed by the input map.
1175 * Returns null if the mappings cannot be traversed (not all toRanges of this
1176 * map correspond to fromRanges of the input), or if this.toRatio does not
1177 * match map.fromRatio.
1181 * this: from [1-100] to [501-600]
1182 * input: from [10-40] to [60-90]
1183 * output: from [10-40] to [560-590]
1184 * Example 2 ('reverse strand exons'):
1185 * this: from [1-100] to [2000-1951], [1000-951] // transcript to loci
1186 * input: from [1-50] to [41-90] // CDS to transcript
1187 * output: from [10-40] to [1960-1951], [1000-971] // CDS to gene loci
1193 public MapList traverse(MapList map)
1201 * compound the ratios by this rule:
1202 * A:B with M:N gives A*M:B*N
1203 * reduced by greatest common divisor
1204 * so 1:3 with 3:3 is 3:9 or 1:3
1205 * 1:3 with 3:1 is 3:3 or 1:1
1206 * 1:3 with 1:3 is 1:9
1207 * 2:5 with 3:7 is 6:35
1209 int outFromRatio = getFromRatio() * map.getFromRatio();
1210 int outToRatio = getToRatio() * map.getToRatio();
1211 int gcd = MathUtils.gcd(outFromRatio, outToRatio);
1212 outFromRatio /= gcd;
1215 List<int[]> toRanges = new ArrayList<>();
1216 for (int[] range : getToRanges())
1218 int fromLength = Math.abs(range[1] - range[0]) + 1;
1219 int[] transferred = map.locateInTo(range[0], range[1]);
1220 if (transferred == null || transferred.length % 2 != 0)
1226 * convert [start1, end1, start2, end2, ...]
1227 * to [[start1, end1], [start2, end2], ...]
1230 for (int i = 0; i < transferred.length;)
1232 toRanges.add(new int[] { transferred[i], transferred[i + 1] });
1233 toLength += Math.abs(transferred[i + 1] - transferred[i]) + 1;
1238 * check we mapped the full range - if not, abort
1240 if (fromLength * map.getToRatio() != toLength * map.getFromRatio())
1246 return new MapList(getFromRanges(), toRanges, outFromRatio, outToRatio);
1250 * Answers true if the mapping is from one contiguous range to another, else
1255 public boolean isContiguous()
1257 return fromShifts.size() == 1 && toShifts.size() == 1;
1261 * Returns the [start, end, start, end, ...] ranges in the 'from' range that
1262 * map to positions between {@code start} and {@code end} in the 'to' range.
1263 * Returns null if no mapped positions are found in start-end.
1269 public int[] locateInFrom2(int start, int end)
1271 List<int[]> ranges = mapBetween(start, end, toShifts, fromShifts,
1272 toRatio, fromRatio);
1274 // TODO: or just return the List and adjust calling code to match
1275 return ranges.isEmpty() ? null : MappingUtils.rangeListToArray(ranges);
1279 * Returns the [start, end, start, end, ...] ranges in the 'to' range that map
1280 * to the given start-end in the 'from' range. Returns null if either
1281 * {@code start} or {@code end} is not a mapped 'from' range position.
1287 public int[] locateInTo2(int start, int end)
1289 List<int[]> ranges = mapBetween(start, end, fromShifts, toShifts,
1290 fromRatio, toRatio);
1292 return ranges.isEmpty() ? null : MappingUtils.rangeListToArray(ranges);
1296 * A helper method for navigating the mapping. Returns a (possibly empty) list
1297 * of [start-end] positions in {@code ranges2} that map to positions in
1298 * {@code ranges1} between {@code start} and {@code end}.
1304 * @param wordLength1
1305 * @param wordLength2
1308 final static List<int[]> mapBetween(int start, int end,
1309 List<int[]> ranges1, List<int[]> ranges2, int wordLength1,
1320 * first traverse ranges1 and record count of mapped positions
1321 * to any that overlap start-end
1323 List<int[]> overlaps = findOverlapPositions(ranges1, start, end);
1324 if (overlaps.isEmpty())
1330 * convert positions to equivalent 'word' positions in ranges2
1332 mapWords(overlaps, wordLength1, wordLength2);
1335 * walk ranges2 and record the values found at
1336 * the offsets in 'overlaps'
1338 List<int[]> mapped = new ArrayList<>();
1339 final int s1 = overlaps.size();
1340 final int s2 = ranges2.size();
1341 int ranges2Index = 0;
1344 * count of mapped positions preceding ranges2[ranges2Index]
1349 * for each [from-to] range in overlaps:
1350 * - walk (what remains of) ranges2
1351 * - record the values at offsets [from-to]
1352 * - stop when past 'to' positions (or at end of ranges2)
1354 for (int i = 0; i < s1; i++)
1356 int[] overlap = overlaps.get(i);
1357 final int toAdd = overlap[1] - overlap[0] + 1;
1358 int added = 0; // how much of overlap has been 'found'
1359 for (; added < toAdd && ranges2Index < s2; ranges2Index++)
1361 int[] range2 = ranges2.get(ranges2Index);
1362 int rangeStart = range2[0];
1363 int rangeEnd = range2[1];
1364 boolean reverseStrand = range2[1] < range2[0];
1365 int rangeLength = Math.abs(rangeEnd - rangeStart) + 1;
1366 if (traversed + rangeLength <= overlap[0])
1369 * precedes overlap - keep looking
1371 traversed += rangeLength;
1374 int overlapStart = overlap[0] - traversed;
1375 int overlapEnd = Math.min(overlapStart + toAdd - added - 1,
1377 int mappedFrom = range2[0] + (reverseStrand ? - overlapStart : overlapStart);
1378 int mappedTo = range2[0] + (reverseStrand ? - overlapEnd : overlapEnd);
1379 mapped.add(new int[] { mappedFrom, mappedTo });
1380 int found = overlapEnd - overlapStart + 1;
1382 overlap[0] += found;
1383 traversed += rangeLength;
1391 * Converts the start-end positions (counted from zero) in the {@code ranges}
1392 * list from one word length to another. Start-end positions are expanded if
1393 * necessary to cover a whole word of length {@code wordLength1}. Positions
1394 * are then divided by {@code wordLength1} and multiplied by
1395 * {@code wordLength2} to give equivalent mapped words.
1397 * Put simply, this converts peptide residue positions to the corresponding
1398 * codon ranges, and codons - including partial codons - to the corresponding
1399 * peptide positions; for example
1402 * [1, 10] with word lengths 3:1 converts (as if bases [0-11]) to [1, 4]
1406 * @param wordLength1
1407 * @param wordLength2
1410 final static void mapWords(List<int[]> ranges, int wordLength1,
1413 if (wordLength1 == 1 && wordLength2 == 1)
1415 return; // nothing to do here
1417 int s = ranges.size();
1418 for (int i = 0; i < s; i++)
1420 int[] range = ranges.get(i);
1423 * expand range start to the start of a word,
1424 * and convert to wordLength2
1426 range[0] -= range[0] % wordLength1;
1427 range[0] = range[0] / wordLength1 * wordLength2;
1430 * similar calculation for range end, adding
1431 * (wordLength2 - 1) for end of mapped word
1433 range[1] -= range[1] % wordLength1;
1434 range[1] = range[1] / wordLength1 * wordLength2;
1435 range[1] += wordLength2 - 1;
1440 * Helper method that returns a (possibly empty) list of offsets in
1441 * {@code ranges} to subranges that overlap {@code start-end} (where start <=
1442 * end}. The list returned holds counts of the number of positions traversed
1443 * (exclusive) to reach the overlapping positions, not the overlapping values.
1444 * Returns null if there are no overlaps.
1451 final static List<int[]> findOverlapPositions(List<int[]> ranges,
1454 List<int[]> positions = new ArrayList<>();
1456 int s = ranges.size();
1457 for (int i = 0; i < s; i++)
1459 int[] range = ranges.get(i);
1460 addOverlap(positions, pos, range, start, end);
1461 pos += 1 + Math.abs(range[1] - range[0]);
1467 * A helper method that checks whether {@code range} overlaps
1468 * {@code start-end}, and if so adds the offset of the overlap in
1469 * {@code range}, plus {@code pos}, to {@code positions}.
1472 * a list of map offsets to add to
1474 * the number of mapped positions already visited
1476 * a from-to range (may be forward or reverse)
1478 * position to test for overlap in range
1480 * position to test for overlap in range
1483 final static void addOverlap(List<int[]> positions, int pos, int[] range,
1486 if (range[1] >= range[0])
1489 * forward direction range
1491 if (start <= range[1] && end >= range[0])
1496 int overlapStart = Math.max(start, range[0]);
1497 int overlapStartOffset = pos + overlapStart - range[0];
1498 int overlapEnd = Math.min(end, range[1]);
1499 int overlapEndOffset = pos + overlapEnd - range[0];
1500 int[] lastOverlap = positions.isEmpty() ? null
1501 : positions.get(positions.size() - 1);
1502 if (lastOverlap != null && overlapStartOffset == lastOverlap[1] + 1)
1505 * just extending the last overlap range
1507 lastOverlap[1] = overlapEndOffset;
1512 * add a new (discontiguous) overlap range
1514 positions.add(new int[] { overlapStartOffset, overlapEndOffset });
1521 * reverse direction range
1523 if (start <= range[0] && end >= range[1])
1528 int overlapStart = Math.max(start, range[1]);
1529 int overlapEnd = Math.min(end, range[0]);
1532 { pos + range[0] - overlapEnd,
1533 pos + range[0] - overlapStart });