scripts/last-dotplot
author Martin C. Frith
Fri Jul 31 15:04:20 2020 +0900 (21 months ago)
changeset 1076 3b8d229f8359
parent 1075 70b0e512d696
child 1077 5209deca661c
permissions -rwxr-xr-x
last-dotplot: implementing annotation names...
     1 #! /usr/bin/env python
     2 # Author: Martin C. Frith 2008
     3 # SPDX-License-Identifier: GPL-3.0-or-later
     4 
     5 # Read pair-wise alignments in MAF or LAST tabular format: write an
     6 # "Oxford grid", a.k.a. dotplot.
     7 
     8 # TODO: Currently, pixels with zero aligned nt-pairs are white, and
     9 # pixels with one or more aligned nt-pairs are black.  This can look
    10 # too crowded for large genome alignments.  I tried shading each pixel
    11 # according to the number of aligned nt-pairs within it, but the
    12 # result is too faint.  How can this be done better?
    13 
    14 import collections
    15 import functools
    16 import gzip
    17 from fnmatch import fnmatchcase
    18 import logging
    19 from operator import itemgetter
    20 import subprocess
    21 import itertools, optparse, os, re, sys
    22 
    23 # Try to make PIL/PILLOW work:
    24 try:
    25     from PIL import Image, ImageDraw, ImageFont, ImageColor
    26 except ImportError:
    27     import Image, ImageDraw, ImageFont, ImageColor
    28 
    29 try:
    30     from future_builtins import zip
    31 except ImportError:
    32     pass
    33 
    34 def myOpen(fileName):  # faster than fileinput
    35     if fileName is None:
    36         return []
    37     if fileName == "-":
    38         return sys.stdin
    39     if fileName.endswith(".gz"):
    40         return gzip.open(fileName, "rt")  # xxx dubious for Python2
    41     return open(fileName)
    42 
    43 def groupByFirstItem(things):
    44     for k, v in itertools.groupby(things, itemgetter(0)):
    45         yield k, [i[1:] for i in v]
    46 
    47 def croppedBlocks(blocks, ranges1, ranges2):
    48     headBeg1, headBeg2, headSize = blocks[0]
    49     for r1 in ranges1:
    50         for r2 in ranges2:
    51             cropBeg1, cropEnd1 = r1
    52             if headBeg1 < 0:
    53                 cropBeg1, cropEnd1 = -cropEnd1, -cropBeg1
    54             cropBeg2, cropEnd2 = r2
    55             if headBeg2 < 0:
    56                 cropBeg2, cropEnd2 = -cropEnd2, -cropBeg2
    57             for beg1, beg2, size in blocks:
    58                 b1 = max(cropBeg1, beg1)
    59                 e1 = min(cropEnd1, beg1 + size)
    60                 if b1 >= e1: continue
    61                 offset = beg2 - beg1
    62                 b2 = max(cropBeg2, b1 + offset)
    63                 e2 = min(cropEnd2, e1 + offset)
    64                 if b2 >= e2: continue
    65                 yield b2 - offset, b2, e2 - b2
    66 
    67 def tabBlocks(beg1, beg2, blocks):
    68     '''Get the gapless blocks of an alignment, from LAST tabular format.'''
    69     for i in blocks.split(","):
    70         if ":" in i:
    71             x, y = i.split(":")
    72             beg1 += int(x)
    73             beg2 += int(y)
    74         else:
    75             size = int(i)
    76             yield beg1, beg2, size
    77             beg1 += size
    78             beg2 += size
    79 
    80 def mafBlocks(beg1, beg2, seq1, seq2):
    81     '''Get the gapless blocks of an alignment, from MAF format.'''
    82     size = 0
    83     for x, y in zip(seq1, seq2):
    84         if x == "-":
    85             if size:
    86                 yield beg1, beg2, size
    87                 beg1 += size
    88                 beg2 += size
    89                 size = 0
    90             beg2 += 1
    91         elif y == "-":
    92             if size:
    93                 yield beg1, beg2, size
    94                 beg1 += size
    95                 beg2 += size
    96                 size = 0
    97             beg1 += 1
    98         else:
    99             size += 1
   100     if size: yield beg1, beg2, size
   101 
   102 def alignmentFromSegment(qrySeqName, qrySeqLen, segment):
   103     refSeqLen = sys.maxsize  # XXX
   104     refSeqName, refSeqBeg, qrySeqBeg, size = segment
   105     block = refSeqBeg, qrySeqBeg, size
   106     return refSeqName, refSeqLen, qrySeqName, qrySeqLen, [block]
   107 
   108 def alignmentInput(lines):
   109     '''Get alignments and sequence lengths, from MAF or tabular format.'''
   110     mafCount = 0
   111     qrySeqName = ""
   112     segments = []
   113     for line in lines:
   114         w = line.split()
   115         if line[0] == "#":
   116             pass
   117         elif len(w) == 1:
   118             for i in segments:
   119                 yield alignmentFromSegment(qrySeqName, qrySeqLen, i)
   120             qrySeqName = w[0]
   121             qrySeqLen = 0
   122             segments = []
   123         elif len(w) == 2 and qrySeqName and w[1].isdigit():
   124             qrySeqLen += int(w[1])
   125         elif len(w) == 4 and qrySeqName and w[1].isdigit() and w[3].isdigit():
   126             refSeqName, refSeqBeg, refSeqEnd = w[0], int(w[1]), int(w[3])
   127             size = abs(refSeqEnd - refSeqBeg)
   128             if refSeqBeg > refSeqEnd:
   129                 refSeqBeg = -refSeqBeg
   130             segments.append((refSeqName, refSeqBeg, qrySeqLen, size))
   131             qrySeqLen += size
   132         elif line[0].isdigit():  # tabular format
   133             chr1, beg1, seqlen1 = w[1], int(w[2]), int(w[5])
   134             if w[4] == "-": beg1 -= seqlen1
   135             chr2, beg2, seqlen2 = w[6], int(w[7]), int(w[10])
   136             if w[9] == "-": beg2 -= seqlen2
   137             blocks = tabBlocks(beg1, beg2, w[11])
   138             yield chr1, seqlen1, chr2, seqlen2, blocks
   139         elif line[0] == "s":  # MAF format
   140             if mafCount == 0:
   141                 chr1, beg1, seqlen1, seq1 = w[1], int(w[2]), int(w[5]), w[6]
   142                 if w[4] == "-": beg1 -= seqlen1
   143                 mafCount = 1
   144             else:
   145                 chr2, beg2, seqlen2, seq2 = w[1], int(w[2]), int(w[5]), w[6]
   146                 if w[4] == "-": beg2 -= seqlen2
   147                 blocks = mafBlocks(beg1, beg2, seq1, seq2)
   148                 yield chr1, seqlen1, chr2, seqlen2, blocks
   149                 mafCount = 0
   150     for i in segments:
   151         yield alignmentFromSegment(qrySeqName, qrySeqLen, i)
   152 
   153 def seqRequestFromText(text):
   154     if ":" in text:
   155         pattern, interval = text.rsplit(":", 1)
   156         if "-" in interval:
   157             beg, end = interval.rsplit("-", 1)
   158             return pattern, int(beg), int(end)  # beg may be negative
   159     return text, 0, sys.maxsize
   160 
   161 def rangesFromSeqName(seqRequests, name, seqLen):
   162     if seqRequests:
   163         base = name.split(".")[-1]  # allow for names like hg19.chr7
   164         for pat, beg, end in seqRequests:
   165             if fnmatchcase(name, pat) or fnmatchcase(base, pat):
   166                 yield max(beg, 0), min(end, seqLen)
   167     else:
   168         yield 0, seqLen
   169 
   170 def updateSeqs(coverDict, seqRanges, seqName, ranges, coveredRange):
   171     beg, end = coveredRange
   172     if beg < 0:
   173         coveredRange = -end, -beg
   174     if seqName in coverDict:
   175         coverDict[seqName].append(coveredRange)
   176     else:
   177         coverDict[seqName] = [coveredRange]
   178         for beg, end in ranges:
   179             r = seqName, beg, end
   180             seqRanges.append(r)
   181 
   182 def readAlignments(fileName, opts):
   183     '''Get alignments and sequence limits, from MAF or tabular format.'''
   184     seqRequests1 = [seqRequestFromText(i) for i in opts.seq1]
   185     seqRequests2 = [seqRequestFromText(i) for i in opts.seq2]
   186 
   187     alignments = []
   188     seqRanges1 = []
   189     seqRanges2 = []
   190     coverDict1 = {}
   191     coverDict2 = {}
   192     lines = myOpen(fileName)
   193     for seqName1, seqLen1, seqName2, seqLen2, blocks in alignmentInput(lines):
   194         ranges1 = sorted(rangesFromSeqName(seqRequests1, seqName1, seqLen1))
   195         if not ranges1: continue
   196         ranges2 = sorted(rangesFromSeqName(seqRequests2, seqName2, seqLen2))
   197         if not ranges2: continue
   198         b = list(croppedBlocks(list(blocks), ranges1, ranges2))
   199         if not b: continue
   200         aln = seqName1, seqName2, b
   201         alignments.append(aln)
   202         coveredRange1 = b[0][0], b[-1][0] + b[-1][2]
   203         updateSeqs(coverDict1, seqRanges1, seqName1, ranges1, coveredRange1)
   204         coveredRange2 = b[0][1], b[-1][1] + b[-1][2]
   205         updateSeqs(coverDict2, seqRanges2, seqName2, ranges2, coveredRange2)
   206     return alignments, seqRanges1, coverDict1, seqRanges2, coverDict2
   207 
   208 def nameAndRangesFromDict(cropDict, seqName):
   209     if seqName in cropDict:
   210         return seqName, cropDict[seqName]
   211     n = seqName.split(".")[-1]
   212     if n in cropDict:
   213         return n, cropDict[n]
   214     return seqName, []
   215 
   216 def rangesForSecondaryAlignments(primaryRanges, seqLen):
   217     if primaryRanges:
   218         return primaryRanges
   219     return [(0, seqLen)]
   220 
   221 def readSecondaryAlignments(opts, cropRanges1, cropRanges2):
   222     cropDict1 = dict(groupByFirstItem(cropRanges1))
   223     cropDict2 = dict(groupByFirstItem(cropRanges2))
   224 
   225     alignments = []
   226     seqRanges1 = []
   227     seqRanges2 = []
   228     coverDict1 = {}
   229     coverDict2 = {}
   230     lines = myOpen(opts.alignments)
   231     for seqName1, seqLen1, seqName2, seqLen2, blocks in alignmentInput(lines):
   232         seqName1, ranges1 = nameAndRangesFromDict(cropDict1, seqName1)
   233         seqName2, ranges2 = nameAndRangesFromDict(cropDict2, seqName2)
   234         if not ranges1 and not ranges2:
   235             continue
   236         r1 = rangesForSecondaryAlignments(ranges1, seqLen1)
   237         r2 = rangesForSecondaryAlignments(ranges2, seqLen2)
   238         b = list(croppedBlocks(list(blocks), r1, r2))
   239         if not b: continue
   240         aln = seqName1, seqName2, b
   241         alignments.append(aln)
   242         if not ranges1:
   243             coveredRange1 = b[0][0], b[-1][0] + b[-1][2]
   244             updateSeqs(coverDict1, seqRanges1, seqName1, r1, coveredRange1)
   245         if not ranges2:
   246             coveredRange2 = b[0][1], b[-1][1] + b[-1][2]
   247             updateSeqs(coverDict2, seqRanges2, seqName2, r2, coveredRange2)
   248     return alignments, seqRanges1, coverDict1, seqRanges2, coverDict2
   249 
   250 def twoValuesFromOption(text, separator):
   251     if separator in text:
   252         return text.split(separator)
   253     return text, text
   254 
   255 def mergedRanges(ranges):
   256     oldBeg, maxEnd = ranges[0]
   257     for beg, end in ranges:
   258         if beg > maxEnd:
   259             yield oldBeg, maxEnd
   260             oldBeg = beg
   261             maxEnd = end
   262         elif end > maxEnd:
   263             maxEnd = end
   264     yield oldBeg, maxEnd
   265 
   266 def mergedRangesPerSeq(coverDict):
   267     for k, v in coverDict.items():
   268         v.sort()
   269         yield k, list(mergedRanges(v))
   270 
   271 def coveredLength(mergedCoverDict):
   272     return sum(sum(e - b for b, e in v) for v in mergedCoverDict.values())
   273 
   274 def trimmed(seqRanges, coverDict, minAlignedBases, maxGapFrac, endPad, midPad):
   275     maxEndGapFrac, maxMidGapFrac = twoValuesFromOption(maxGapFrac, ",")
   276     maxEndGap = max(float(maxEndGapFrac) * minAlignedBases, endPad * 1.0)
   277     maxMidGap = max(float(maxMidGapFrac) * minAlignedBases, midPad * 2.0)
   278 
   279     for seqName, rangeBeg, rangeEnd in seqRanges:
   280         seqBlocks = coverDict[seqName]
   281         blocks = [i for i in seqBlocks if i[0] < rangeEnd and i[1] > rangeBeg]
   282         if blocks[0][0] - rangeBeg > maxEndGap:
   283             rangeBeg = blocks[0][0] - endPad
   284         for j, y in enumerate(blocks):
   285             if j:
   286                 x = blocks[j - 1]
   287                 if y[0] - x[1] > maxMidGap:
   288                     yield seqName, rangeBeg, x[1] + midPad
   289                     rangeBeg = y[0] - midPad
   290         if rangeEnd - blocks[-1][1] > maxEndGap:
   291             rangeEnd = blocks[-1][1] + endPad
   292         yield seqName, rangeBeg, rangeEnd
   293 
   294 def rangesWithStrandInfo(seqRanges, strandOpt, alignments, seqIndex):
   295     if strandOpt == "1":
   296         forwardMinusReverse = collections.defaultdict(int)
   297         for i in alignments:
   298             blocks = i[2]
   299             beg1, beg2, size = blocks[0]
   300             numOfAlignedLetterPairs = sum(i[2] for i in blocks)
   301             if (beg1 < 0) != (beg2 < 0):  # opposite-strand alignment
   302                 numOfAlignedLetterPairs *= -1
   303             forwardMinusReverse[i[seqIndex]] += numOfAlignedLetterPairs
   304     strandNum = 0
   305     for seqName, beg, end in seqRanges:
   306         if strandOpt == "1":
   307             strandNum = 1 if forwardMinusReverse[seqName] >= 0 else 2
   308         yield seqName, beg, end, strandNum
   309 
   310 def natural_sort_key(my_string):
   311     '''Return a sort key for "natural" ordering, e.g. chr9 < chr10.'''
   312     parts = re.split(r'(\d+)', my_string)
   313     parts[1::2] = map(int, parts[1::2])
   314     return parts
   315 
   316 def nameKey(oneSeqRanges):
   317     return natural_sort_key(oneSeqRanges[0][0])
   318 
   319 def sizeKey(oneSeqRanges):
   320     return sum(b - e for n, b, e, s in oneSeqRanges), nameKey(oneSeqRanges)
   321 
   322 def alignmentKey(seqNamesToLists, oneSeqRanges):
   323     seqName = oneSeqRanges[0][0]
   324     alignmentsOfThisSequence = seqNamesToLists[seqName]
   325     numOfAlignedLetterPairs = sum(i[3] for i in alignmentsOfThisSequence)
   326     toMiddle = numOfAlignedLetterPairs // 2
   327     for i in alignmentsOfThisSequence:
   328         toMiddle -= i[3]
   329         if toMiddle < 0:
   330             return i[1:3]  # sequence-rank and "position" of this alignment
   331 
   332 def rankAndFlipPerSeq(seqRanges):
   333     rangesGroupedBySeqName = itertools.groupby(seqRanges, itemgetter(0))
   334     for rank, group in enumerate(rangesGroupedBySeqName):
   335         seqName, ranges = group
   336         strandNum = next(ranges)[3]
   337         flip = 1 if strandNum < 2 else -1
   338         yield seqName, (rank, flip)
   339 
   340 def alignmentSortData(alignments, seqIndex, otherNamesToRanksAndFlips):
   341     otherIndex = 1 - seqIndex
   342     for i in alignments:
   343         blocks = i[2]
   344         otherRank, otherFlip = otherNamesToRanksAndFlips[i[otherIndex]]
   345         otherPos = otherFlip * abs(blocks[0][otherIndex] +
   346                                    blocks[-1][otherIndex] + blocks[-1][2])
   347         numOfAlignedLetterPairs = sum(i[2] for i in blocks)
   348         yield i[seqIndex], otherRank, otherPos, numOfAlignedLetterPairs
   349 
   350 def mySortedRanges(seqRanges, sortOpt, seqIndex, alignments, otherRanges):
   351     rangesGroupedBySeqName = itertools.groupby(seqRanges, itemgetter(0))
   352     g = [list(ranges) for seqName, ranges in rangesGroupedBySeqName]
   353     for i in g:
   354         if i[0][3] > 1:
   355             i.reverse()
   356     if sortOpt == "1":
   357         g.sort(key=nameKey)
   358     if sortOpt == "2":
   359         g.sort(key=sizeKey)
   360     if sortOpt == "3":
   361         otherNamesToRanksAndFlips = dict(rankAndFlipPerSeq(otherRanges))
   362         alns = sorted(alignmentSortData(alignments, seqIndex,
   363                                         otherNamesToRanksAndFlips))
   364         alnsGroupedBySeqName = itertools.groupby(alns, itemgetter(0))
   365         seqNamesToLists = dict((k, list(v)) for k, v in alnsGroupedBySeqName)
   366         g.sort(key=functools.partial(alignmentKey, seqNamesToLists))
   367     return [j for i in g for j in i]
   368 
   369 def allSortedRanges(opts, alignments, alignmentsB,
   370                     seqRanges1, seqRangesB1, seqRanges2, seqRangesB2):
   371     o1, oB1 = twoValuesFromOption(opts.strands1, ":")
   372     o2, oB2 = twoValuesFromOption(opts.strands2, ":")
   373     if o1 == "1" and o2 == "1":
   374         raise Exception("the strand options have circular dependency")
   375     seqRanges1 = list(rangesWithStrandInfo(seqRanges1, o1, alignments, 0))
   376     seqRanges2 = list(rangesWithStrandInfo(seqRanges2, o2, alignments, 1))
   377     seqRangesB1 = list(rangesWithStrandInfo(seqRangesB1, oB1, alignmentsB, 0))
   378     seqRangesB2 = list(rangesWithStrandInfo(seqRangesB2, oB2, alignmentsB, 1))
   379 
   380     o1, oB1 = twoValuesFromOption(opts.sort1, ":")
   381     o2, oB2 = twoValuesFromOption(opts.sort2, ":")
   382     if o1 == "3" and o2 == "3":
   383         raise Exception("the sort options have circular dependency")
   384     if o1 != "3":
   385         s1 = mySortedRanges(seqRanges1, o1, None, None, None)
   386     if o2 != "3":
   387         s2 = mySortedRanges(seqRanges2, o2, None, None, None)
   388     if o1 == "3":
   389         s1 = mySortedRanges(seqRanges1, o1, 0, alignments, s2)
   390     if o2 == "3":
   391         s2 = mySortedRanges(seqRanges2, o2, 1, alignments, s1)
   392     t1 = mySortedRanges(seqRangesB1, oB1, 0, alignmentsB, s2)
   393     t2 = mySortedRanges(seqRangesB2, oB2, 1, alignmentsB, s1)
   394     return s1 + t1, s2 + t2
   395 
   396 def sizesPerText(texts, font, textDraw):
   397     sizes = 0, 0
   398     for t in texts:
   399         if textDraw is not None:
   400             sizes = textDraw.textsize(t, font=font)
   401         yield t, sizes
   402 
   403 def prettyNum(n):
   404     t = str(n)
   405     groups = []
   406     while t:
   407         groups.append(t[-3:])
   408         t = t[:-3]
   409     return ",".join(reversed(groups))
   410 
   411 def sizeText(size):
   412     suffixes = "bp", "kb", "Mb", "Gb"
   413     for i, x in enumerate(suffixes):
   414         j = 10 ** (i * 3)
   415         if size < j * 10:
   416             return "%.2g" % (1.0 * size / j) + x
   417         if size < j * 1000 or i == len(suffixes) - 1:
   418             return "%.0f" % (1.0 * size / j) + x
   419 
   420 def labelText(seqRange, labelOpt):
   421     seqName, beg, end, strandNum = seqRange
   422     if labelOpt == 1:
   423         return seqName + ": " + sizeText(end - beg)
   424     if labelOpt == 2:
   425         return seqName + ":" + prettyNum(beg) + ": " + sizeText(end - beg)
   426     if labelOpt == 3:
   427         return seqName + ":" + prettyNum(beg) + "-" + prettyNum(end)
   428     return seqName
   429 
   430 def rangeLabels(seqRanges, labelOpt, font, textDraw, textRot):
   431     x = y = 0
   432     for r in seqRanges:
   433         text = labelText(r, labelOpt)
   434         if textDraw is not None:
   435             x, y = textDraw.textsize(text, font=font)
   436             if textRot:
   437                 x, y = y, x
   438         yield text, x, y, r[3]
   439 
   440 def dataFromRanges(sortedRanges, font, textDraw, labelOpt, textRot):
   441     for seqName, rangeBeg, rangeEnd, strandNum in sortedRanges:
   442         out = [seqName, str(rangeBeg), str(rangeEnd)]
   443         if strandNum > 0:
   444             out.append(".+-"[strandNum])
   445         logging.info("\t".join(out))
   446     logging.info("")
   447     rangeSizes = [e - b for n, b, e, s in sortedRanges]
   448     labs = list(rangeLabels(sortedRanges, labelOpt, font, textDraw, textRot))
   449     margin = max(i[2] for i in labs)
   450     # xxx the margin may be too big, because some labels may get omitted
   451     return rangeSizes, labs, margin
   452 
   453 def div_ceil(x, y):
   454     '''Return x / y rounded up.'''
   455     q, r = divmod(x, y)
   456     return q + (r != 0)
   457 
   458 def get_bp_per_pix(rangeSizes, pixTweenRanges, maxPixels):
   459     '''Get the minimum bp-per-pixel that fits in the size limit.'''
   460     logging.info("choosing bp per pixel...")
   461     numOfRanges = len(rangeSizes)
   462     maxPixelsInRanges = maxPixels - pixTweenRanges * (numOfRanges - 1)
   463     if maxPixelsInRanges < numOfRanges:
   464         raise Exception("can't fit the image: too many sequences?")
   465     negLimit = -maxPixelsInRanges
   466     negBpPerPix = sum(rangeSizes) // negLimit
   467     while True:
   468         if sum(i // negBpPerPix for i in rangeSizes) >= negLimit:
   469             return -negBpPerPix
   470         negBpPerPix -= 1
   471 
   472 def getRangePixBegs(rangePixLens, pixTweenRanges, margin):
   473     '''Get the start pixel for each range.'''
   474     rangePixBegs = []
   475     pix_tot = margin - pixTweenRanges
   476     for i in rangePixLens:
   477         pix_tot += pixTweenRanges
   478         rangePixBegs.append(pix_tot)
   479         pix_tot += i
   480     return rangePixBegs
   481 
   482 def pixelData(rangeSizes, bp_per_pix, pixTweenRanges, margin):
   483     '''Return pixel information about the ranges.'''
   484     rangePixLens = [div_ceil(i, bp_per_pix) for i in rangeSizes]
   485     rangePixBegs = getRangePixBegs(rangePixLens, pixTweenRanges, margin)
   486     tot_pix = rangePixBegs[-1] + rangePixLens[-1]
   487     return rangePixBegs, rangePixLens, tot_pix
   488 
   489 def drawLineForward(hits, width, bp_per_pix, beg1, beg2, size):
   490     while True:
   491         q1, r1 = divmod(beg1, bp_per_pix)
   492         q2, r2 = divmod(beg2, bp_per_pix)
   493         hits[q2 * width + q1] |= 1
   494         next_pix = min(bp_per_pix - r1, bp_per_pix - r2)
   495         if next_pix >= size: break
   496         beg1 += next_pix
   497         beg2 += next_pix
   498         size -= next_pix
   499 
   500 def drawLineReverse(hits, width, bp_per_pix, beg1, beg2, size):
   501     while True:
   502         q1, r1 = divmod(beg1, bp_per_pix)
   503         q2, r2 = divmod(beg2, bp_per_pix)
   504         hits[q2 * width + q1] |= 2
   505         next_pix = min(bp_per_pix - r1, r2 + 1)
   506         if next_pix >= size: break
   507         beg1 += next_pix
   508         beg2 -= next_pix
   509         size -= next_pix
   510 
   511 def strandAndOrigin(ranges, beg, size):
   512     isReverseStrand = (beg < 0)
   513     if isReverseStrand:
   514         beg = -(beg + size)
   515     for rangeBeg, rangeEnd, isReverseRange, origin in ranges:
   516         if rangeEnd > beg:  # assumes the ranges are sorted
   517             return (isReverseStrand != isReverseRange), origin
   518 
   519 def alignmentPixels(width, height, alignments, bp_per_pix,
   520                     rangeDict1, rangeDict2):
   521     hits = [0] * (width * height)  # the image data
   522     for seq1, seq2, blocks in alignments:
   523         beg1, beg2, size = blocks[0]
   524         isReverse1, ori1 = strandAndOrigin(rangeDict1[seq1], beg1, size)
   525         isReverse2, ori2 = strandAndOrigin(rangeDict2[seq2], beg2, size)
   526         for beg1, beg2, size in blocks:
   527             if isReverse1:
   528                 beg1 = -(beg1 + size)
   529                 beg2 = -(beg2 + size)
   530             if isReverse1 == isReverse2:
   531                 drawLineForward(hits, width, bp_per_pix,
   532                                 ori1 + beg1, ori2 + beg2, size)
   533             else:
   534                 drawLineReverse(hits, width, bp_per_pix,
   535                                 ori1 + beg1, ori2 - beg2 - 1, size)
   536     return hits
   537 
   538 def orientedBlocks(alignments, seqIndex):
   539     otherIndex = 1 - seqIndex
   540     for a in alignments:
   541         seq1, seq2, blocks = a
   542         for b in blocks:
   543             beg1, beg2, size = b
   544             if b[seqIndex] < 0:
   545                 b = -(beg1 + size), -(beg2 + size), size
   546             yield a[seqIndex], b[seqIndex], a[otherIndex], b[otherIndex], size
   547 
   548 def drawJoins(im, alignments, bpPerPix, seqIndex, rangeDict1, rangeDict2):
   549     blocks = orientedBlocks(alignments, seqIndex)
   550     oldSeq1 = ""
   551     for seq1, beg1, seq2, beg2, size in sorted(blocks):
   552         isReverse1, ori1 = strandAndOrigin(rangeDict1[seq1], beg1, size)
   553         isReverse2, ori2 = strandAndOrigin(rangeDict2[seq2], beg2, size)
   554         end1 = beg1 + size - 1
   555         end2 = beg2 + size - 1
   556         if isReverse1:
   557             beg1 = -(beg1 + 1)
   558             end1 = -(end1 + 1)
   559         if isReverse2:
   560             beg2 = -(beg2 + 1)
   561             end2 = -(end2 + 1)
   562         newPix1 = (ori1 + beg1) // bpPerPix
   563         newPix2 = (ori2 + beg2) // bpPerPix
   564         if seq1 == oldSeq1:
   565             lowerPix2 = min(oldPix2, newPix2)
   566             upperPix2 = max(oldPix2, newPix2)
   567             midPix1 = (oldPix1 + newPix1) // 2
   568             if isReverse1:
   569                 midPix1 = (oldPix1 + newPix1 + 1) // 2
   570                 oldPix1, newPix1 = newPix1, oldPix1
   571             if upperPix2 - lowerPix2 > 1 and oldPix1 <= newPix1 <= oldPix1 + 1:
   572                 if seqIndex == 0:
   573                     box = midPix1, lowerPix2, midPix1 + 1, upperPix2 + 1
   574                 else:
   575                     box = lowerPix2, midPix1, upperPix2 + 1, midPix1 + 1
   576                 im.paste("lightgray", box)
   577         oldPix1 = (ori1 + end1) // bpPerPix
   578         oldPix2 = (ori2 + end2) // bpPerPix
   579         oldSeq1 = seq1
   580 
   581 def expandedSeqDict(seqDict):
   582     '''Allow lookup by short sequence names, e.g. chr7 as well as hg19.chr7.'''
   583     newDict = seqDict.copy()
   584     for name, x in seqDict.items():
   585         if "." in name:
   586             base = name.split(".")[-1]
   587             if base in newDict:  # an ambiguous case was found:
   588                 return seqDict   # so give up completely
   589             newDict[base] = x
   590     return newDict
   591 
   592 def readBed(fileName, rangeDict):
   593     for line in myOpen(fileName):
   594         w = line.split()
   595         if not w: continue
   596         seqName = w[0]
   597         if seqName not in rangeDict: continue
   598         seqRanges = rangeDict[seqName]
   599         beg = int(w[1])
   600         end = int(w[2])
   601         if all(beg >= i[2] or end <= i[1] for i in seqRanges):
   602             continue
   603         itemName = w[3] if len(w) > 3 and w[3] != "." else ""
   604         layer = 900
   605         color = "#fbf"
   606         if len(w) > 4:
   607             if w[4] != ".":
   608                 layer = float(w[4])
   609             if len(w) > 5:
   610                 if len(w) > 8 and w[8].count(",") == 2:
   611                     color = "rgb(" + w[8] + ")"
   612                 else:
   613                     strand = w[5]
   614                     isRev = (seqRanges[0][3] > 1)
   615                     if strand == "+" and not isRev or strand == "-" and isRev:
   616                         color = "#ffe8e8"
   617                     if strand == "-" and not isRev or strand == "+" and isRev:
   618                         color = "#e8e8ff"
   619         yield layer, color, seqName, beg, end, itemName
   620 
   621 def commaSeparatedInts(text):
   622     return map(int, text.rstrip(",").split(","))
   623 
   624 def readGenePred(opts, fileName, rangeDict):
   625     for line in myOpen(fileName):
   626         fields = line.split()  # xxx tab ???
   627         if not fields: continue
   628         geneName = fields[12 if len(fields) > 12 else 0]  # XXX ???
   629         if fields[2] not in "+-":
   630             fields = fields[1:]
   631         seqName = fields[1]
   632         if seqName not in rangeDict: continue
   633         seqRanges = rangeDict[seqName]
   634         #strand = fields[2]
   635         cdsBeg = int(fields[5])
   636         cdsEnd = int(fields[6])
   637         exonBegs = commaSeparatedInts(fields[8])
   638         exonEnds = commaSeparatedInts(fields[9])
   639         for beg, end in zip(exonBegs, exonEnds):
   640             if all(beg >= i[2] or end <= i[1] for i in seqRanges):
   641                 continue
   642             yield 300, opts.exon_color, seqName, beg, end, geneName
   643             b = max(beg, cdsBeg)
   644             e = min(end, cdsEnd)
   645             if b < e: yield 400, opts.cds_color, seqName, b, e, ""
   646 
   647 def readRmsk(fileName, rangeDict):
   648     for line in myOpen(fileName):
   649         fields = line.split()
   650         if len(fields) == 17:  # rmsk.txt
   651             seqName = fields[5]
   652             if seqName not in rangeDict: continue  # do this ASAP for speed
   653             beg = int(fields[6])
   654             end = int(fields[7])
   655             strand = fields[9]
   656             repeatName = fields[10]
   657             repeatClass = fields[11]
   658         elif len(fields) == 15:  # .out
   659             seqName = fields[4]
   660             if seqName not in rangeDict: continue
   661             beg = int(fields[5]) - 1
   662             end = int(fields[6])
   663             strand = fields[8]
   664             repeatName = fields[9]
   665             repeatClass = fields[10].split("/")[0]
   666         else:
   667             continue
   668         seqRanges = rangeDict[seqName]
   669         if all(beg >= i[2] or end <= i[1] for i in seqRanges):
   670             continue
   671         if repeatClass in ("Low_complexity", "Simple_repeat", "Satellite"):
   672             yield 200, "#fbf", seqName, beg, end, repeatName
   673         elif (strand == "+") != (seqRanges[0][3] > 1):
   674             yield 100, "#ffe8e8", seqName, beg, end, repeatName
   675         else:
   676             yield 100, "#e8e8ff", seqName, beg, end, repeatName
   677 
   678 def isExtraFirstGapField(fields):
   679     return fields[4].isdigit()
   680 
   681 def readGaps(opts, fileName, rangeDict):
   682     '''Read locations of unsequenced gaps, from an agp or gap file.'''
   683     for line in myOpen(fileName):
   684         w = line.split()
   685         if not w or w[0][0] == "#": continue
   686         if isExtraFirstGapField(w): w = w[1:]
   687         if w[4] not in "NU": continue
   688         seqName = w[0]
   689         if seqName not in rangeDict: continue
   690         end = int(w[2])
   691         beg = end - int(w[5])  # zero-based coordinate
   692         if w[7] == "yes":
   693             yield 3000, opts.bridged_color, seqName, beg, end, ""
   694         else:
   695             yield 2000, opts.unbridged_color, seqName, beg, end, ""
   696 
   697 def bedBoxes(beds, rangeDict, edge, isTop, bpPerPix, textSizes):
   698     cover = [(edge, edge)]
   699     for layer, color, seqName, bedBeg, bedEnd, name in reversed(beds):
   700         for rangeBeg, rangeEnd, isReverseRange, origin in rangeDict[seqName]:
   701             beg = max(bedBeg, rangeBeg)
   702             end = min(bedEnd, rangeEnd)
   703             if beg >= end: continue
   704             if isReverseRange:
   705                 beg, end = -end, -beg
   706             if layer <= 1000:
   707                 # include partly-covered pixels
   708                 pixBeg = (origin + beg) // bpPerPix
   709                 pixEnd = div_ceil(origin + end, bpPerPix)
   710             else:
   711                 # exclude partly-covered pixels
   712                 pixBeg = div_ceil(origin + beg, bpPerPix)
   713                 pixEnd = (origin + end) // bpPerPix
   714                 if pixEnd <= pixBeg: continue
   715                 if bedEnd >= rangeEnd:  # include partly-covered end pixels
   716                     if isReverseRange:
   717                         pixBeg = (origin + beg) // bpPerPix
   718                     else:
   719                         pixEnd = div_ceil(origin + end, bpPerPix)
   720             textWidth, textHeight = textSizes[name]
   721             nameBeg = (pixBeg + pixEnd - textHeight) // 2
   722             nameEnd = nameBeg + textHeight
   723             if name and all(e <= nameBeg or b >= nameEnd for b, e in cover):
   724                 cover.append((nameBeg, nameEnd))
   725             else:
   726                 name = ""
   727             yield layer, color, isTop, pixBeg, pixEnd, name, nameBeg, textWidth
   728 
   729 def drawAnnotations(im, boxes, tMargin, bMarginBeg, lMargin, rMarginBeg):
   730     # xxx use partial transparency for different-color overlaps?
   731     for layer, color, isTop, beg, end, name, nameBeg, nameLen in boxes:
   732         if isTop:
   733             box = beg, tMargin, end, bMarginBeg
   734         else:
   735             box = lMargin, beg, rMarginBeg, end
   736         im.paste(color, box)
   737 
   738 def placedLabels(labels, rangePixBegs, rangePixLens, beg, end):
   739     '''Return axis labels with endpoint & sort-order information.'''
   740     maxWidth = end - beg
   741     for i, j, k in zip(labels, rangePixBegs, rangePixLens):
   742         text, textWidth, textHeight, strandNum = i
   743         if textWidth > maxWidth:
   744             continue
   745         labelBeg = j + (k - textWidth) // 2
   746         labelEnd = labelBeg + textWidth
   747         sortKey = textWidth - k
   748         if labelBeg < beg:
   749             sortKey += maxWidth * (beg - labelBeg)
   750             labelBeg = beg
   751             labelEnd = beg + textWidth
   752         if labelEnd > end:
   753             sortKey += maxWidth * (labelEnd - end)
   754             labelEnd = end
   755             labelBeg = end - textWidth
   756         yield sortKey, labelBeg, labelEnd, text, textHeight, strandNum
   757 
   758 def nonoverlappingLabels(labels, minPixTweenLabels):
   759     '''Get a subset of non-overlapping axis labels, greedily.'''
   760     out = []
   761     for i in labels:
   762         beg = i[1] - minPixTweenLabels
   763         end = i[2] + minPixTweenLabels
   764         if all(j[2] <= beg or j[1] >= end for j in out):
   765             out.append(i)
   766     return out
   767 
   768 def axisImage(labels, rangePixBegs, rangePixLens, textRot,
   769               textAln, font, image_mode, opts):
   770     '''Make an image of axis labels.'''
   771     beg = rangePixBegs[0]
   772     end = rangePixBegs[-1] + rangePixLens[-1]
   773     margin = max(i[2] for i in labels)
   774     labels = sorted(placedLabels(labels, rangePixBegs, rangePixLens, beg, end))
   775     minPixTweenLabels = 0 if textRot else opts.label_space
   776     labels = nonoverlappingLabels(labels, minPixTweenLabels)
   777     image_size = (margin, end) if textRot else (end, margin)
   778     im = Image.new(image_mode, image_size, opts.margin_color)
   779     draw = ImageDraw.Draw(im)
   780     for sortKey, labelBeg, labelEnd, text, textHeight, strandNum in labels:
   781         base = margin - textHeight if textAln else 0
   782         position = (base, labelBeg) if textRot else (labelBeg, base)
   783         fill = ("black", opts.forwardcolor, opts.reversecolor)[strandNum]
   784         draw.text(position, text, font=font, fill=fill)
   785     return im
   786 
   787 def rangesPerSeq(sortedRanges):
   788     for seqName, group in itertools.groupby(sortedRanges, itemgetter(0)):
   789         yield seqName, sorted(group)
   790 
   791 def rangesWithOrigins(sortedRanges, rangePixBegs, rangePixLens, bpPerPix):
   792     for i, j, k in zip(sortedRanges, rangePixBegs, rangePixLens):
   793         seqName, rangeBeg, rangeEnd, strandNum = i
   794         isReverseRange = (strandNum > 1)
   795         if isReverseRange:
   796             origin = bpPerPix * (j + k) + rangeBeg
   797         else:
   798             origin = bpPerPix * j - rangeBeg
   799         yield seqName, (rangeBeg, rangeEnd, isReverseRange, origin)
   800 
   801 def rangesAndOriginsPerSeq(sortedRanges, rangePixBegs, rangePixLens, bpPerPix):
   802     a = rangesWithOrigins(sortedRanges, rangePixBegs, rangePixLens, bpPerPix)
   803     for seqName, group in itertools.groupby(a, itemgetter(0)):
   804         yield seqName, sorted(i[1] for i in group)
   805 
   806 def getFont(opts):
   807     if opts.fontfile:
   808         return ImageFont.truetype(opts.fontfile, opts.fontsize)
   809     fileNames = []
   810     try:
   811         x = ["fc-match", "-f%{file}", "arial"]
   812         p = subprocess.Popen(x, stdout=subprocess.PIPE, stderr=subprocess.PIPE,
   813                              universal_newlines=True)
   814         out, err = p.communicate()
   815         fileNames.append(out)
   816     except OSError as e:
   817         logging.info("fc-match error: " + str(e))
   818     fileNames.append("/Library/Fonts/Arial.ttf")  # for Mac
   819     for i in fileNames:
   820         try:
   821             font = ImageFont.truetype(i, opts.fontsize)
   822             logging.info("font: " + i)
   823             return font
   824         except IOError as e:
   825             logging.info("font load error: " + str(e))
   826     return ImageFont.load_default()
   827 
   828 def sequenceSizesAndNames(seqRanges):
   829     for seqName, ranges in itertools.groupby(seqRanges, itemgetter(0)):
   830         size = sum(e - b for n, b, e in ranges)
   831         yield size, seqName
   832 
   833 def biggestSequences(seqRanges, maxNumOfSequences):
   834     s = sorted(sequenceSizesAndNames(seqRanges), reverse=True)
   835     if len(s) > maxNumOfSequences:
   836         logging.warning("too many sequences - discarding the smallest ones")
   837         s = s[:maxNumOfSequences]
   838     return set(i[1] for i in s)
   839 
   840 def remainingSequenceRanges(seqRanges, alignments, seqIndex):
   841     remainingSequences = set(i[seqIndex] for i in alignments)
   842     return [i for i in seqRanges if i[0] in remainingSequences]
   843 
   844 def readAnnotations(opts, font, textDraw, sortedRanges,
   845                     bedFile, repFile, geneFile, gapFile):
   846     rangeDict = expandedSeqDict(dict(rangesPerSeq(sortedRanges)))
   847     annots = sorted(itertools.chain(readBed(bedFile, rangeDict),
   848                                     readRmsk(repFile, rangeDict),
   849                                     readGenePred(opts, geneFile, rangeDict),
   850                                     readGaps(opts, gapFile, rangeDict)))
   851     names = set(i[5] for i in annots)
   852     textSizes = dict(sizesPerText(names, font, textDraw))
   853     return annots, textSizes
   854 
   855 def lastDotplot(opts, args):
   856     logLevel = logging.INFO if opts.verbose else logging.WARNING
   857     logging.basicConfig(format="%(filename)s: %(message)s", level=logLevel)
   858 
   859     font = getFont(opts)
   860     image_mode = 'RGB'
   861     forward_color = ImageColor.getcolor(opts.forwardcolor, image_mode)
   862     reverse_color = ImageColor.getcolor(opts.reversecolor, image_mode)
   863     zipped_colors = zip(forward_color, reverse_color)
   864     overlap_color = tuple([(i + j) // 2 for i, j in zipped_colors])
   865 
   866     maxGap1, maxGapB1 = twoValuesFromOption(opts.max_gap1, ":")
   867     maxGap2, maxGapB2 = twoValuesFromOption(opts.max_gap2, ":")
   868 
   869     logging.info("reading alignments...")
   870     alnData = readAlignments(args[0], opts)
   871     alignments, seqRanges1, coverDict1, seqRanges2, coverDict2 = alnData
   872     if not alignments: raise Exception("there are no alignments")
   873     logging.info("cutting...")
   874     coverDict1 = dict(mergedRangesPerSeq(coverDict1))
   875     coverDict2 = dict(mergedRangesPerSeq(coverDict2))
   876     minAlignedBases = min(coveredLength(coverDict1), coveredLength(coverDict2))
   877     pad = int(opts.pad * minAlignedBases)
   878     cutRanges1 = list(trimmed(seqRanges1, coverDict1, minAlignedBases,
   879                               maxGap1, pad, pad))
   880     cutRanges2 = list(trimmed(seqRanges2, coverDict2, minAlignedBases,
   881                               maxGap2, pad, pad))
   882 
   883     biggestSeqs1 = biggestSequences(cutRanges1, opts.maxseqs)
   884     cutRanges1 = [i for i in cutRanges1 if i[0] in biggestSeqs1]
   885     alignments = [i for i in alignments if i[0] in biggestSeqs1]
   886     cutRanges2 = remainingSequenceRanges(cutRanges2, alignments, 1)
   887 
   888     biggestSeqs2 = biggestSequences(cutRanges2, opts.maxseqs)
   889     cutRanges2 = [i for i in cutRanges2 if i[0] in biggestSeqs2]
   890     alignments = [i for i in alignments if i[1] in biggestSeqs2]
   891     cutRanges1 = remainingSequenceRanges(cutRanges1, alignments, 0)
   892 
   893     logging.info("reading secondary alignments...")
   894     alnDataB = readSecondaryAlignments(opts, cutRanges1, cutRanges2)
   895     alignmentsB, seqRangesB1, coverDictB1, seqRangesB2, coverDictB2 = alnDataB
   896     logging.info("cutting...")
   897     coverDictB1 = dict(mergedRangesPerSeq(coverDictB1))
   898     coverDictB2 = dict(mergedRangesPerSeq(coverDictB2))
   899     cutRangesB1 = trimmed(seqRangesB1, coverDictB1, minAlignedBases,
   900                           maxGapB1, 0, 0)
   901     cutRangesB2 = trimmed(seqRangesB2, coverDictB2, minAlignedBases,
   902                           maxGapB2, 0, 0)
   903 
   904     logging.info("sorting...")
   905     sortOut = allSortedRanges(opts, alignments, alignmentsB,
   906                               cutRanges1, cutRangesB1, cutRanges2, cutRangesB2)
   907     sortedRanges1, sortedRanges2 = sortOut
   908 
   909     textDraw = None
   910     if opts.fontsize:
   911         textDraw = ImageDraw.Draw(Image.new(image_mode, (1, 1)))
   912 
   913     textRot1 = "vertical".startswith(opts.rot1)
   914     i1 = dataFromRanges(sortedRanges1, font, textDraw, opts.labels1, textRot1)
   915     rangeSizes1, labelData1, tMargin = i1
   916 
   917     textRot2 = "horizontal".startswith(opts.rot2)
   918     i2 = dataFromRanges(sortedRanges2, font, textDraw, opts.labels2, textRot2)
   919     rangeSizes2, labelData2, lMargin = i2
   920 
   921     logging.info("reading annotations...")
   922 
   923     a1 = readAnnotations(opts, font, textDraw, sortedRanges1,
   924                          opts.bed1, opts.rmsk1, opts.genePred1, opts.gap1)
   925     annots1, annoTextSizes1 = a1
   926 
   927     a2 = readAnnotations(opts, font, textDraw, sortedRanges2,
   928                          opts.bed2, opts.rmsk2, opts.genePred2, opts.gap2)
   929     annots2, annoTextSizes2 = a2
   930 
   931     bMargin = rMargin = 0  # xxx
   932 
   933     maxPixels1 = opts.width  - lMargin - rMargin
   934     maxPixels2 = opts.height - tMargin - bMargin
   935     bpPerPix1 = get_bp_per_pix(rangeSizes1, opts.border_pixels, maxPixels1)
   936     bpPerPix2 = get_bp_per_pix(rangeSizes2, opts.border_pixels, maxPixels2)
   937     bpPerPix = max(bpPerPix1, bpPerPix2)
   938     logging.info("bp per pixel = " + str(bpPerPix))
   939 
   940     p1 = pixelData(rangeSizes1, bpPerPix, opts.border_pixels, lMargin)
   941     rangePixBegs1, rangePixLens1, rMarginBeg = p1
   942     width = rMarginBeg + rMargin
   943     rangeDict1 = dict(rangesAndOriginsPerSeq(sortedRanges1, rangePixBegs1,
   944                                              rangePixLens1, bpPerPix))
   945 
   946     p2 = pixelData(rangeSizes2, bpPerPix, opts.border_pixels, tMargin)
   947     rangePixBegs2, rangePixLens2, bMarginBeg = p2
   948     height = bMarginBeg + bMargin
   949     rangeDict2 = dict(rangesAndOriginsPerSeq(sortedRanges2, rangePixBegs2,
   950                                              rangePixLens2, bpPerPix))
   951 
   952     logging.info("width:  " + str(width))
   953     logging.info("height: " + str(height))
   954 
   955     logging.info("processing alignments...")
   956     allAlignments = alignments + alignmentsB
   957     hits = alignmentPixels(width, height, allAlignments, bpPerPix,
   958                            rangeDict1, rangeDict2)
   959 
   960     rangeDict1 = expandedSeqDict(rangeDict1)
   961     rangeDict2 = expandedSeqDict(rangeDict2)
   962 
   963     boxes1 = list(bedBoxes(annots1, rangeDict1, bMarginBeg, True, bpPerPix,
   964                            annoTextSizes1))
   965     boxes2 = list(bedBoxes(annots2, rangeDict2, rMarginBeg, False, bpPerPix,
   966                            annoTextSizes2))
   967     boxes = sorted(itertools.chain(boxes1, boxes2))
   968 
   969     logging.info("drawing...")
   970 
   971     image_size = width, height
   972     im = Image.new(image_mode, image_size, opts.background_color)
   973 
   974     drawAnnotations(im, boxes, tMargin, bMarginBeg, lMargin, rMarginBeg)
   975 
   976     joinA, joinB = twoValuesFromOption(opts.join, ":")
   977     if joinA in "13":
   978         drawJoins(im, alignments, bpPerPix, 0, rangeDict1, rangeDict2)
   979     if joinB in "13":
   980         drawJoins(im, alignmentsB, bpPerPix, 0, rangeDict1, rangeDict2)
   981     if joinA in "23":
   982         drawJoins(im, alignments, bpPerPix, 1, rangeDict2, rangeDict1)
   983     if joinB in "23":
   984         drawJoins(im, alignmentsB, bpPerPix, 1, rangeDict2, rangeDict1)
   985 
   986     for i in range(height):
   987         for j in range(width):
   988             store_value = hits[i * width + j]
   989             xy = j, i
   990             if   store_value == 1: im.putpixel(xy, forward_color)
   991             elif store_value == 2: im.putpixel(xy, reverse_color)
   992             elif store_value == 3: im.putpixel(xy, overlap_color)
   993 
   994     if opts.fontsize != 0:
   995         axis1 = axisImage(labelData1, rangePixBegs1, rangePixLens1,
   996                           textRot1, False, font, image_mode, opts)
   997         if textRot1:
   998             axis1 = axis1.transpose(Image.ROTATE_90)
   999         axis2 = axisImage(labelData2, rangePixBegs2, rangePixLens2,
  1000                           textRot2, textRot2, font, image_mode, opts)
  1001         if not textRot2:
  1002             axis2 = axis2.transpose(Image.ROTATE_270)
  1003         im.paste(axis1, (0, 0))
  1004         im.paste(axis2, (0, 0))
  1005 
  1006     for i in rangePixBegs1[1:]:
  1007         box = i - opts.border_pixels, tMargin, i, bMarginBeg
  1008         im.paste(opts.border_color, box)
  1009 
  1010     for i in rangePixBegs2[1:]:
  1011         box = lMargin, i - opts.border_pixels, rMarginBeg, i
  1012         im.paste(opts.border_color, box)
  1013 
  1014     im.save(args[1])
  1015 
  1016 if __name__ == "__main__":
  1017     usage = """%prog --help
  1018    or: %prog [options] maf-or-tab-alignments dotplot.png
  1019    or: %prog [options] maf-or-tab-alignments dotplot.gif
  1020    or: ..."""
  1021     description = "Draw a dotplot of pair-wise sequence alignments in MAF or tabular format."
  1022     op = optparse.OptionParser(usage=usage, description=description)
  1023     op.add_option("-v", "--verbose", action="count",
  1024                   help="show progress messages & data about the plot")
  1025     # Replace "width" & "height" with a single "length" option?
  1026     op.add_option("-x", "--width", metavar="INT", type="int", default=1000,
  1027                   help="maximum width in pixels (default: %default)")
  1028     op.add_option("-y", "--height", metavar="INT", type="int", default=1000,
  1029                   help="maximum height in pixels (default: %default)")
  1030     op.add_option("-m", "--maxseqs", type="int", default=100, metavar="M",
  1031                   help="maximum number of horizontal or vertical sequences "
  1032                   "(default=%default)")
  1033     op.add_option("-1", "--seq1", metavar="PATTERN", action="append",
  1034                   default=[],
  1035                   help="which sequences to show from the 1st genome")
  1036     op.add_option("-2", "--seq2", metavar="PATTERN", action="append",
  1037                   default=[],
  1038                   help="which sequences to show from the 2nd genome")
  1039     op.add_option("-c", "--forwardcolor", metavar="COLOR", default="red",
  1040                   help="color for forward alignments (default: %default)")
  1041     op.add_option("-r", "--reversecolor", metavar="COLOR", default="blue",
  1042                   help="color for reverse alignments (default: %default)")
  1043     op.add_option("--alignments", metavar="FILE", help="secondary alignments")
  1044     op.add_option("--sort1", default="1", metavar="N",
  1045                   help="genome1 sequence order: 0=input order, 1=name order, "
  1046                   "2=length order, 3=alignment order (default=%default)")
  1047     op.add_option("--sort2", default="1", metavar="N",
  1048                   help="genome2 sequence order: 0=input order, 1=name order, "
  1049                   "2=length order, 3=alignment order (default=%default)")
  1050     op.add_option("--strands1", default="0", metavar="N", help=
  1051                   "genome1 sequence orientation: 0=forward orientation, "
  1052                   "1=alignment orientation (default=%default)")
  1053     op.add_option("--strands2", default="0", metavar="N", help=
  1054                   "genome2 sequence orientation: 0=forward orientation, "
  1055                   "1=alignment orientation (default=%default)")
  1056     op.add_option("--max-gap1", metavar="FRAC", default="0.5,2", help=
  1057                   "maximum unaligned (end,mid) gap in genome1: "
  1058                   "fraction of aligned length (default=%default)")
  1059     op.add_option("--max-gap2", metavar="FRAC", default="0.5,2", help=
  1060                   "maximum unaligned (end,mid) gap in genome2: "
  1061                   "fraction of aligned length (default=%default)")
  1062     op.add_option("--pad", metavar="FRAC", type="float", default=0.04, help=
  1063                   "pad length when cutting unaligned gaps: "
  1064                   "fraction of aligned length (default=%default)")
  1065     op.add_option("-j", "--join", default="0", metavar="N", help=
  1066                   "join: 0=nothing, 1=alignments adjacent in genome1, "
  1067                   "2=alignments adjacent in genome2 (default=%default)")
  1068     op.add_option("--border-pixels", metavar="INT", type="int", default=1,
  1069                   help="number of pixels between sequences (default=%default)")
  1070     op.add_option("--border-color", metavar="COLOR", default="black",
  1071                   help="color for pixels between sequences (default=%default)")
  1072     # --break-color and/or --break-pixels for intra-sequence breaks?
  1073     op.add_option("--margin-color", metavar="COLOR", default="#dcdcdc",
  1074                   help="margin color")
  1075 
  1076     og = optparse.OptionGroup(op, "Text options")
  1077     og.add_option("-f", "--fontfile", metavar="FILE",
  1078                   help="TrueType or OpenType font file")
  1079     og.add_option("-s", "--fontsize", metavar="SIZE", type="int", default=14,
  1080                   help="TrueType or OpenType font size (default: %default)")
  1081     og.add_option("--labels1", type="int", default=0, metavar="N", help=
  1082                   "genome1 labels: 0=name, 1=name:length, "
  1083                   "2=name:start:length, 3=name:start-end (default=%default)")
  1084     og.add_option("--labels2", type="int", default=0, metavar="N", help=
  1085                   "genome2 labels: 0=name, 1=name:length, "
  1086                   "2=name:start:length, 3=name:start-end (default=%default)")
  1087     og.add_option("--rot1", metavar="ROT", default="h",
  1088                   help="text rotation for the 1st genome (default=%default)")
  1089     og.add_option("--rot2", metavar="ROT", default="v",
  1090                   help="text rotation for the 2nd genome (default=%default)")
  1091     op.add_option_group(og)
  1092 
  1093     og = optparse.OptionGroup(op, "Annotation options")
  1094     og.add_option("--bed1", metavar="FILE",
  1095                   help="read genome1 annotations from BED file")
  1096     og.add_option("--bed2", metavar="FILE",
  1097                   help="read genome2 annotations from BED file")
  1098     og.add_option("--rmsk1", metavar="FILE", help="read genome1 repeats from "
  1099                   "RepeatMasker .out or rmsk.txt file")
  1100     og.add_option("--rmsk2", metavar="FILE", help="read genome2 repeats from "
  1101                   "RepeatMasker .out or rmsk.txt file")
  1102     op.add_option_group(og)
  1103 
  1104     og = optparse.OptionGroup(op, "Gene options")
  1105     og.add_option("--genePred1", metavar="FILE",
  1106                   help="read genome1 genes from genePred file")
  1107     og.add_option("--genePred2", metavar="FILE",
  1108                   help="read genome2 genes from genePred file")
  1109     og.add_option("--exon-color", metavar="COLOR", default="PaleGreen",
  1110                   help="color for exons (default=%default)")
  1111     og.add_option("--cds-color", metavar="COLOR", default="LimeGreen",
  1112                   help="color for protein-coding regions (default=%default)")
  1113     op.add_option_group(og)
  1114 
  1115     og = optparse.OptionGroup(op, "Unsequenced gap options")
  1116     og.add_option("--gap1", metavar="FILE",
  1117                   help="read genome1 unsequenced gaps from agp or gap file")
  1118     og.add_option("--gap2", metavar="FILE",
  1119                   help="read genome2 unsequenced gaps from agp or gap file")
  1120     og.add_option("--bridged-color", metavar="COLOR", default="yellow",
  1121                   help="color for bridged gaps (default: %default)")
  1122     og.add_option("--unbridged-color", metavar="COLOR", default="orange",
  1123                   help="color for unbridged gaps (default: %default)")
  1124     op.add_option_group(og)
  1125     (opts, args) = op.parse_args()
  1126     if len(args) != 2: op.error("2 arguments needed")
  1127 
  1128     opts.background_color = "white"
  1129     opts.label_space = 5     # minimum number of pixels between axis labels
  1130 
  1131     try: lastDotplot(opts, args)
  1132     except KeyboardInterrupt: pass  # avoid silly error message
  1133     except Exception as e:
  1134         prog = os.path.basename(sys.argv[0])
  1135         sys.exit(prog + ": error: " + str(e))