scripts/last-dotplot
author Martin C. Frith
Tue Apr 04 12:15:35 2017 +0900 (2017-04-04)
changeset 846 1f46ab956351
parent 845 16060c00b129
child 847 3ca7aa4b27a8
permissions -rwxr-xr-x
Added last-dotplot options to show sequence lengths.
Martin@1
     1
#! /usr/bin/env python
Martin@1
     2
Martin@272
     3
# Read pair-wise alignments in MAF or LAST tabular format: write an
Martin@272
     4
# "Oxford grid", a.k.a. dotplot.
Martin@1
     5
Martin@1
     6
# TODO: Currently, pixels with zero aligned nt-pairs are white, and
Martin@1
     7
# pixels with one or more aligned nt-pairs are black.  This can look
Martin@1
     8
# too crowded for large genome alignments.  I tried shading each pixel
Martin@1
     9
# according to the number of aligned nt-pairs within it, but the
Martin@1
    10
# result is too faint.  How can this be done better?
Martin@1
    11
Martin@844
    12
import fnmatch, itertools, optparse, os, re, sys
Martin@475
    13
Martin@475
    14
# Try to make PIL/PILLOW work:
Martin@475
    15
try: from PIL import Image, ImageDraw, ImageFont, ImageColor
Martin@475
    16
except ImportError: import Image, ImageDraw, ImageFont, ImageColor
Martin@1
    17
Martin@844
    18
def myOpen(fileName):  # faster than fileinput
Martin@844
    19
    if fileName == "-":
Martin@844
    20
        return sys.stdin
Martin@844
    21
    return open(fileName)
Martin@844
    22
Martin@644
    23
def warn(message):
Martin@644
    24
    prog = os.path.basename(sys.argv[0])
Martin@644
    25
    sys.stderr.write(prog + ": " + message + "\n")
Martin@644
    26
Martin@840
    27
def croppedBlocks(blocks, range1, range2):
Martin@840
    28
    cropBeg1, cropEnd1 = range1
Martin@840
    29
    cropBeg2, cropEnd2 = range2
Martin@840
    30
    if blocks[0][0] < 0: cropBeg1, cropEnd1 = -cropEnd1, -cropBeg1
Martin@840
    31
    if blocks[0][1] < 0: cropBeg2, cropEnd2 = -cropEnd2, -cropBeg2
Martin@840
    32
    for beg1, beg2, size in blocks:
Martin@840
    33
        b1 = max(cropBeg1, beg1)
Martin@840
    34
        e1 = min(cropEnd1, beg1 + size)
Martin@840
    35
        if b1 >= e1: continue
Martin@840
    36
        offset = beg2 - beg1
Martin@840
    37
        b2 = max(cropBeg2, b1 + offset)
Martin@840
    38
        e2 = min(cropEnd2, e1 + offset)
Martin@840
    39
        if b2 >= e2: continue
Martin@840
    40
        yield b2 - offset, b2, e2 - b2
Martin@840
    41
Martin@482
    42
def tabBlocks(beg1, beg2, blocks):
Martin@482
    43
    '''Get the gapless blocks of an alignment, from LAST tabular format.'''
Martin@482
    44
    for i in blocks.split(","):
Martin@482
    45
        if ":" in i:
Martin@482
    46
            x, y = i.split(":")
Martin@482
    47
            beg1 += int(x)
Martin@482
    48
            beg2 += int(y)
Martin@482
    49
        else:
Martin@482
    50
            size = int(i)
Martin@482
    51
            yield beg1, beg2, size
Martin@482
    52
            beg1 += size
Martin@482
    53
            beg2 += size
Martin@272
    54
Martin@482
    55
def mafBlocks(beg1, beg2, seq1, seq2):
Martin@482
    56
    '''Get the gapless blocks of an alignment, from MAF format.'''
Martin@482
    57
    size = 0
Martin@482
    58
    for x, y in itertools.izip(seq1, seq2):
Martin@482
    59
        if x == "-":
Martin@482
    60
            if size:
Martin@482
    61
                yield beg1, beg2, size
Martin@482
    62
                beg1 += size
Martin@482
    63
                beg2 += size
Martin@482
    64
                size = 0
Martin@482
    65
            beg2 += 1
Martin@482
    66
        elif y == "-":
Martin@482
    67
            if size:
Martin@482
    68
                yield beg1, beg2, size
Martin@482
    69
                beg1 += size
Martin@482
    70
                beg2 += size
Martin@482
    71
                size = 0
Martin@482
    72
            beg1 += 1
Martin@272
    73
        else:
Martin@482
    74
            size += 1
Martin@482
    75
    if size: yield beg1, beg2, size
Martin@272
    76
Martin@482
    77
def alignmentInput(lines):
Martin@482
    78
    '''Get alignments and sequence lengths, from MAF or tabular format.'''
Martin@482
    79
    mafCount = 0
Martin@272
    80
    for line in lines:
Martin@272
    81
        w = line.split()
Martin@272
    82
        if line[0].isdigit():  # tabular format
Martin@482
    83
            chr1, beg1, seqlen1 = w[1], int(w[2]), int(w[5])
Martin@482
    84
            if w[4] == "-": beg1 -= seqlen1
Martin@482
    85
            chr2, beg2, seqlen2 = w[6], int(w[7]), int(w[10])
Martin@482
    86
            if w[9] == "-": beg2 -= seqlen2
Martin@839
    87
            blocks = list(tabBlocks(beg1, beg2, w[11]))
Martin@482
    88
            yield chr1, seqlen1, chr2, seqlen2, blocks
Martin@272
    89
        elif line[0] == "s":  # MAF format
Martin@482
    90
            if mafCount == 0:
Martin@482
    91
                chr1, beg1, seqlen1, seq1 = w[1], int(w[2]), int(w[5]), w[6]
Martin@482
    92
                if w[4] == "-": beg1 -= seqlen1
Martin@482
    93
                mafCount = 1
Martin@482
    94
            else:
Martin@482
    95
                chr2, beg2, seqlen2, seq2 = w[1], int(w[2]), int(w[5]), w[6]
Martin@482
    96
                if w[4] == "-": beg2 -= seqlen2
Martin@839
    97
                blocks = list(mafBlocks(beg1, beg2, seq1, seq2))
Martin@482
    98
                yield chr1, seqlen1, chr2, seqlen2, blocks
Martin@482
    99
                mafCount = 0
Martin@272
   100
Martin@840
   101
def seqRangeFromText(text):
Martin@840
   102
    if ":" in text:
Martin@840
   103
        pattern, interval = text.rsplit(":", 1)
Martin@840
   104
        if "-" in interval:
Martin@840
   105
            beg, end = interval.rsplit("-", 1)
Martin@840
   106
            return pattern, int(beg), int(end)  # beg may be negative
Martin@840
   107
    return text, 0, sys.maxsize
Martin@840
   108
Martin@840
   109
def rangeFromSeqName(seqRanges, name, seqLen):
Martin@840
   110
    if not seqRanges: return 0, seqLen
Martin@651
   111
    base = name.split(".")[-1]  # allow for names like hg19.chr7
Martin@840
   112
    for pat, beg, end in seqRanges:
Martin@840
   113
        if fnmatch.fnmatchcase(name, pat) or fnmatch.fnmatchcase(base, pat):
Martin@840
   114
            return max(beg, 0), min(end, seqLen)
Martin@844
   115
    return None
Martin@651
   116
Martin@840
   117
def updateSeqLimits(isTrim, seqLimits, seqName, seqRange, blocks, index):
Martin@839
   118
    if isTrim:
Martin@839
   119
        beg = blocks[0][index]
Martin@839
   120
        end = blocks[-1][index] + blocks[-1][2]
Martin@839
   121
        if beg < 0: beg, end = -end, -beg
Martin@839
   122
        if seqName in seqLimits:
Martin@839
   123
            b, e = seqLimits[seqName]
Martin@839
   124
            seqLimits[seqName] = min(b, beg), max(e, end)
Martin@839
   125
        else:
Martin@839
   126
            seqLimits[seqName] = beg, end
Martin@839
   127
    else:
Martin@840
   128
        seqLimits[seqName] = seqRange
Martin@839
   129
Martin@651
   130
def readAlignments(fileName, opts):
Martin@839
   131
    '''Get alignments and sequence limits, from MAF or tabular format.'''
Martin@840
   132
    seqRanges1 = map(seqRangeFromText, opts.seq1)
Martin@840
   133
    seqRanges2 = map(seqRangeFromText, opts.seq2)
Martin@840
   134
Martin@482
   135
    alignments = []
Martin@839
   136
    seqLimits1 = {}
Martin@839
   137
    seqLimits2 = {}
Martin@844
   138
    lines = myOpen(fileName)
Martin@838
   139
    for seqName1, seqLen1, seqName2, seqLen2, blocks in alignmentInput(lines):
Martin@840
   140
        range1 = rangeFromSeqName(seqRanges1, seqName1, seqLen1)
Martin@844
   141
        if not range1: continue
Martin@840
   142
        range2 = rangeFromSeqName(seqRanges2, seqName2, seqLen2)
Martin@844
   143
        if not range2: continue
Martin@840
   144
        b = list(croppedBlocks(blocks, range1, range2))
Martin@840
   145
        if not b: continue
Martin@840
   146
        aln = seqName1, seqName2, b
Martin@482
   147
        alignments.append(aln)
Martin@840
   148
        updateSeqLimits(opts.trim1, seqLimits1, seqName1, range1, b, 0)
Martin@840
   149
        updateSeqLimits(opts.trim2, seqLimits2, seqName2, range2, b, 1)
Martin@839
   150
    return alignments, seqLimits1, seqLimits2
Martin@1
   151
Martin@1
   152
def natural_sort_key(my_string):
Martin@1
   153
    '''Return a sort key for "natural" ordering, e.g. chr9 < chr10.'''
Martin@1
   154
    parts = re.split(r'(\d+)', my_string)
Martin@1
   155
    parts[1::2] = map(int, parts[1::2])
Martin@1
   156
    return parts
Martin@1
   157
Martin@647
   158
def get_text_sizes(my_strings, font, fontsize, image_mode):
Martin@1
   159
    '''Get widths & heights, in pixels, of some strings.'''
Martin@647
   160
    if fontsize == 0: return [(0, 0) for i in my_strings]
Martin@1
   161
    image_size = 1, 1
Martin@134
   162
    im = Image.new(image_mode, image_size)
Martin@1
   163
    draw = ImageDraw.Draw(im)
Martin@1
   164
    return [draw.textsize(i, font=font) for i in my_strings]
Martin@1
   165
Martin@846
   166
def sizeText(size):
Martin@846
   167
    suffixes = "bp", "kb", "Mb", "Gb"
Martin@846
   168
    for i, x in enumerate(suffixes):
Martin@846
   169
        j = 10 ** (i * 3)
Martin@846
   170
        if size < j * 10:
Martin@846
   171
            return "%.2g" % (1.0 * size / j) + x
Martin@846
   172
        if size < j * 1000 or i == len(suffixes) - 1:
Martin@846
   173
            return "%.0f" % (1.0 * size / j) + x
Martin@846
   174
Martin@846
   175
def seqNameAndSizeText(seqName, seqSize):
Martin@846
   176
    return seqName + ": " + sizeText(seqSize)
Martin@846
   177
Martin@846
   178
def getSeqInfo(seqLimits, font, fontsize, image_mode, isShowSize):
Martin@1
   179
    '''Return miscellaneous information about the sequences.'''
Martin@839
   180
    seqNames = seqLimits.keys()
Martin@837
   181
    seqNames.sort(key=natural_sort_key)
Martin@839
   182
    seq_sizes = [seqLimits[i][1] - seqLimits[i][0] for i in seqNames]
Martin@846
   183
    if isShowSize: seqLabels = map(seqNameAndSizeText, seqNames, seq_sizes)
Martin@846
   184
    else:          seqLabels = seqNames
Martin@846
   185
    labelSizes = get_text_sizes(seqLabels, font, fontsize, image_mode)
Martin@846
   186
    margin = max(zip(*labelSizes)[1])  # maximum text height
Martin@846
   187
    return seqNames, seq_sizes, seqLabels, labelSizes, margin
Martin@1
   188
Martin@1
   189
def div_ceil(x, y):
Martin@1
   190
    '''Return x / y rounded up.'''
Martin@1
   191
    q, r = divmod(x, y)
Martin@1
   192
    return q + (r != 0)
Martin@1
   193
Martin@1
   194
def tot_seq_pix(seq_sizes, bp_per_pix):
Martin@1
   195
    '''Return the total pixels needed for sequences of the given sizes.'''
Martin@28
   196
    return sum([div_ceil(i, bp_per_pix) for i in seq_sizes])
Martin@1
   197
Martin@645
   198
def get_bp_per_pix(seq_sizes, pix_tween_seqs, pix_limit):
Martin@1
   199
    '''Get the minimum bp-per-pixel that fits in the size limit.'''
Martin@1
   200
    seq_num = len(seq_sizes)
Martin@1
   201
    seq_pix_limit = pix_limit - pix_tween_seqs * (seq_num - 1)
Martin@1
   202
    if seq_pix_limit < seq_num:
Martin@649
   203
        raise Exception("can't fit the image: too many sequences?")
Martin@51
   204
    lower_bound = div_ceil(sum(seq_sizes), seq_pix_limit)
Martin@1
   205
    for bp_per_pix in itertools.count(lower_bound):  # slow linear search
Martin@1
   206
        if tot_seq_pix(seq_sizes, bp_per_pix) <= seq_pix_limit: break
Martin@1
   207
    return bp_per_pix
Martin@1
   208
Martin@1
   209
def get_seq_starts(seq_pix, pix_tween_seqs, margin):
Martin@1
   210
    '''Get the start pixel for each sequence.'''
Martin@1
   211
    seq_starts = []
Martin@1
   212
    pix_tot = margin - pix_tween_seqs
Martin@1
   213
    for i in seq_pix:
Martin@1
   214
        pix_tot += pix_tween_seqs
Martin@1
   215
        seq_starts.append(pix_tot)
Martin@1
   216
        pix_tot += i
Martin@1
   217
    return seq_starts
Martin@1
   218
Martin@645
   219
def get_pix_info(seq_sizes, bp_per_pix, pix_tween_seqs, margin):
Martin@1
   220
    '''Return pixel information about the sequences.'''
Martin@1
   221
    seq_pix = [div_ceil(i, bp_per_pix) for i in seq_sizes]
Martin@1
   222
    seq_starts = get_seq_starts(seq_pix, pix_tween_seqs, margin)
Martin@1
   223
    tot_pix = seq_starts[-1] + seq_pix[-1]
Martin@1
   224
    return seq_pix, seq_starts, tot_pix
Martin@1
   225
Martin@835
   226
def drawLineForward(hits, width, bp_per_pix, beg1, beg2, size):
Martin@639
   227
    while True:
Martin@639
   228
        q1, r1 = divmod(beg1, bp_per_pix)
Martin@639
   229
        q2, r2 = divmod(beg2, bp_per_pix)
Martin@835
   230
        hits[q2 * width + q1] |= 1
Martin@639
   231
        next_pix = min(bp_per_pix - r1, bp_per_pix - r2)
Martin@639
   232
        if next_pix >= size: break
Martin@639
   233
        beg1 += next_pix
Martin@639
   234
        beg2 += next_pix
Martin@639
   235
        size -= next_pix
Martin@639
   236
Martin@835
   237
def drawLineReverse(hits, width, bp_per_pix, beg1, beg2, size):
Martin@639
   238
    beg2 = -1 - beg2
Martin@639
   239
    while True:
Martin@639
   240
        q1, r1 = divmod(beg1, bp_per_pix)
Martin@639
   241
        q2, r2 = divmod(beg2, bp_per_pix)
Martin@835
   242
        hits[q2 * width + q1] |= 2
Martin@639
   243
        next_pix = min(bp_per_pix - r1, r2 + 1)
Martin@639
   244
        if next_pix >= size: break
Martin@639
   245
        beg1 += next_pix
Martin@639
   246
        beg2 -= next_pix
Martin@639
   247
        size -= next_pix
Martin@639
   248
Martin@836
   249
def alignmentPixels(width, height, alignments, bp_per_pix, origins1, origins2):
Martin@640
   250
    hits = [0] * (width * height)  # the image data
Martin@640
   251
    for seq1, seq2, blocks in alignments:
Martin@836
   252
        ori1 = origins1[seq1]
Martin@836
   253
        ori2 = origins2[seq2]
Martin@640
   254
        for beg1, beg2, size in blocks:
Martin@640
   255
            if beg1 < 0:
Martin@640
   256
                beg1 = -(beg1 + size)
Martin@640
   257
                beg2 = -(beg2 + size)
Martin@640
   258
            if beg2 >= 0:
Martin@835
   259
                drawLineForward(hits, width, bp_per_pix,
Martin@835
   260
                                beg1 + ori1, beg2 + ori2, size)
Martin@640
   261
            else:
Martin@835
   262
                drawLineReverse(hits, width, bp_per_pix,
Martin@835
   263
                                beg1 + ori1, beg2 - ori2, size)
Martin@640
   264
    return hits
Martin@1
   265
Martin@650
   266
def expandedSeqDict(seqDict):
Martin@650
   267
    '''Allow lookup by short sequence names, e.g. chr7 as well as hg19.chr7.'''
Martin@650
   268
    newDict = {}
Martin@650
   269
    for name, x in seqDict.items():
Martin@650
   270
        base = name.split(".")[-1]
Martin@650
   271
        newDict[name] = x
Martin@650
   272
        newDict[base] = x
Martin@650
   273
    return newDict
Martin@650
   274
Martin@845
   275
def readBed(fileName, seqLimits):
Martin@845
   276
    if not fileName: return
Martin@845
   277
    for line in myOpen(fileName):
Martin@845
   278
        w = line.split()
Martin@845
   279
        if not w or w[0][0] == "#": continue
Martin@845
   280
        seqName = w[0]
Martin@845
   281
        if seqName not in seqLimits: continue
Martin@845
   282
        cropBeg, cropEnd = seqLimits[seqName]
Martin@845
   283
        beg = int(w[1])
Martin@845
   284
        end = int(w[2])
Martin@845
   285
        b = max(beg, cropBeg)
Martin@845
   286
        e = min(end, cropEnd)
Martin@845
   287
        if b >= e: continue
Martin@845
   288
        if len(w) > 8:
Martin@845
   289
            color = tuple(map(int, w[8].split(",")))
Martin@845
   290
        elif len(w) > 5:
Martin@845
   291
            if   w[5] == "+": color = 255, 244, 244
Martin@845
   292
            elif w[5] == "-": color = 244, 244, 255
Martin@845
   293
            else:             color = 255, 228, 255
Martin@845
   294
        else:
Martin@845
   295
            color = 255, 228, 255
Martin@845
   296
        yield seqName, b, e, color
Martin@845
   297
Martin@650
   298
def isExtraFirstGapField(fields):
Martin@650
   299
    return fields[4].isdigit()
Martin@650
   300
Martin@839
   301
def readGaps(fileName, seqLimits):
Martin@650
   302
    '''Read locations of unsequenced gaps, from an agp or gap file.'''
Martin@650
   303
    if not fileName: return
Martin@844
   304
    for line in myOpen(fileName):
Martin@650
   305
        w = line.split()
Martin@650
   306
        if not w or w[0][0] == "#": continue
Martin@650
   307
        if isExtraFirstGapField(w): w = w[1:]
Martin@650
   308
        if w[4] not in "NU": continue
Martin@650
   309
        seqName = w[0]
Martin@839
   310
        if seqName not in seqLimits: continue
Martin@839
   311
        cropBeg, cropEnd = seqLimits[seqName]
Martin@650
   312
        end = int(w[2])
Martin@650
   313
        beg = end - int(w[5])  # zero-based coordinate
Martin@839
   314
        b = max(beg, cropBeg)
Martin@839
   315
        e = min(end, cropEnd)
Martin@839
   316
        if b >= e: continue
Martin@650
   317
        bridgedText = w[7]
Martin@839
   318
        yield seqName, b, e, bridgedText
Martin@650
   319
Martin@845
   320
def drawAnnotations(im, beds, origins, margin, limit, isTop, bp_per_pix):
Martin@845
   321
    # XXX no consideration of different-color overlaps
Martin@845
   322
    for seqName, beg, end, color in beds:
Martin@845
   323
        ori = origins[seqName]
Martin@845
   324
        b = (ori + beg) // bp_per_pix
Martin@845
   325
        e = div_ceil(ori + end, bp_per_pix)
Martin@845
   326
        if isTop: box = b, margin, e, limit
Martin@845
   327
        else:     box = margin, b, limit, e
Martin@845
   328
        im.paste(color, box)
Martin@845
   329
Martin@836
   330
def drawUnsequencedGaps(im, gaps, origins, margin, limit, isTop, bridgedText,
Martin@650
   331
                        bp_per_pix, color):
Martin@650
   332
    '''Draw rectangles representing unsequenced gaps.'''
Martin@650
   333
    for seqName, beg, end, b in gaps:
Martin@650
   334
        if b != bridgedText: continue
Martin@836
   335
        ori = origins[seqName]
Martin@836
   336
        b = div_ceil(ori + beg, bp_per_pix)  # use fully-covered pixels only
Martin@836
   337
        e = (ori + end) // bp_per_pix
Martin@650
   338
        if e <= b: continue
Martin@836
   339
        if isTop: box = b, margin, e, limit
Martin@836
   340
        else:     box = margin, b, limit, e
Martin@650
   341
        im.paste(color, box)
Martin@650
   342
Martin@1
   343
def make_label(text, text_size, range_start, range_size):
Martin@1
   344
    '''Return an axis label with endpoint & sort-order information.'''
Martin@1
   345
    text_width  = text_size[0]
Martin@1
   346
    label_start = range_start + (range_size - text_width) // 2
Martin@1
   347
    label_end   = label_start + text_width
Martin@1
   348
    sort_key    = text_width - range_size
Martin@1
   349
    return sort_key, label_start, label_end, text
Martin@1
   350
Martin@645
   351
def get_nonoverlapping_labels(labels, label_space):
Martin@1
   352
    '''Get a subset of non-overlapping axis labels, greedily.'''
Martin@1
   353
    nonoverlapping_labels = []
Martin@1
   354
    for i in labels:
Martin@28
   355
        if True not in [i[1] < j[2] + label_space and j[1] < i[2] + label_space
Martin@28
   356
                        for j in nonoverlapping_labels]:
Martin@1
   357
            nonoverlapping_labels.append(i)
Martin@1
   358
    return nonoverlapping_labels
Martin@1
   359
Martin@837
   360
def get_axis_image(seqNames, name_sizes, seq_starts, seq_pix,
Martin@647
   361
                   font, image_mode, opts):
Martin@1
   362
    '''Make an image of axis labels.'''
Martin@1
   363
    min_pos = seq_starts[0]
Martin@1
   364
    max_pos = seq_starts[-1] + seq_pix[-1]
Martin@28
   365
    height = max(zip(*name_sizes)[1])
Martin@1
   366
    labels = [make_label(i, j, k, l) for i, j, k, l in
Martin@837
   367
              zip(seqNames, name_sizes, seq_starts, seq_pix)]
Martin@1
   368
    labels = [i for i in labels if i[1] >= min_pos and i[2] <= max_pos]
Martin@1
   369
    labels.sort()
Martin@646
   370
    labels = get_nonoverlapping_labels(labels, opts.label_space)
Martin@1
   371
    image_size = max_pos, height
Martin@646
   372
    im = Image.new(image_mode, image_size, opts.border_shade)
Martin@1
   373
    draw = ImageDraw.Draw(im)
Martin@1
   374
    for i in labels:
Martin@1
   375
        position = i[1], 0
Martin@646
   376
        draw.text(position, i[3], font=font, fill=opts.text_color)
Martin@1
   377
    return im
Martin@1
   378
Martin@839
   379
def seqOrigins(seqNames, seq_starts, seqLimits, bp_per_pix):
Martin@837
   380
    for i, j in zip(seqNames, seq_starts):
Martin@839
   381
        yield i, bp_per_pix * j - seqLimits[i][0]
Martin@836
   382
Martin@648
   383
def lastDotplot(opts, args):
Martin@643
   384
    if opts.fontfile:  font = ImageFont.truetype(opts.fontfile, opts.fontsize)
Martin@643
   385
    else:              font = ImageFont.load_default()
Martin@641
   386
Martin@643
   387
    image_mode = 'RGB'
Martin@643
   388
    forward_color = ImageColor.getcolor(opts.forwardcolor, image_mode)
Martin@643
   389
    reverse_color = ImageColor.getcolor(opts.reversecolor, image_mode)
Martin@643
   390
    zipped_colors = zip(forward_color, reverse_color)
Martin@643
   391
    overlap_color = tuple([(i + j) // 2 for i, j in zipped_colors])
Martin@641
   392
Martin@644
   393
    warn("reading alignments...")
Martin@839
   394
    alignments, seqLimits1, seqLimits2 = readAlignments(args[0], opts)
Martin@644
   395
    warn("done")
Martin@641
   396
Martin@649
   397
    if not alignments: raise Exception("there are no alignments")
Martin@641
   398
Martin@846
   399
    i1 = getSeqInfo(seqLimits1, font, opts.fontsize, image_mode, opts.lengths1)
Martin@846
   400
    seqNames1, seq_sizes1, seqLabels1, labelSizes1, margin1 = i1
Martin@846
   401
Martin@846
   402
    i2 = getSeqInfo(seqLimits2, font, opts.fontsize, image_mode, opts.lengths2)
Martin@846
   403
    seqNames2, seq_sizes2, seqLabels2, labelSizes2, margin2 = i2
Martin@641
   404
Martin@644
   405
    warn("choosing bp per pixel...")
Martin@645
   406
    pix_limit1 = opts.width  - margin1
Martin@645
   407
    pix_limit2 = opts.height - margin2
Martin@646
   408
    bp_per_pix1 = get_bp_per_pix(seq_sizes1, opts.pix_tween_seqs, pix_limit1)
Martin@646
   409
    bp_per_pix2 = get_bp_per_pix(seq_sizes2, opts.pix_tween_seqs, pix_limit2)
Martin@643
   410
    bp_per_pix = max(bp_per_pix1, bp_per_pix2)
Martin@644
   411
    warn("bp per pixel = " + str(bp_per_pix))
Martin@641
   412
Martin@645
   413
    seq_pix1, seq_starts1, width  = get_pix_info(seq_sizes1, bp_per_pix,
Martin@646
   414
                                                 opts.pix_tween_seqs, margin1)
Martin@645
   415
    seq_pix2, seq_starts2, height = get_pix_info(seq_sizes2, bp_per_pix,
Martin@646
   416
                                                 opts.pix_tween_seqs, margin2)
Martin@839
   417
Martin@839
   418
    origins1 = dict(seqOrigins(seqNames1, seq_starts1, seqLimits1, bp_per_pix))
Martin@839
   419
    origins2 = dict(seqOrigins(seqNames2, seq_starts2, seqLimits2, bp_per_pix))
Martin@641
   420
Martin@644
   421
    warn("processing alignments...")
Martin@643
   422
    hits = alignmentPixels(width, height, alignments, bp_per_pix,
Martin@836
   423
                           origins1, origins2)
Martin@644
   424
    warn("done")
Martin@641
   425
Martin@643
   426
    image_size = width, height
Martin@646
   427
    im = Image.new(image_mode, image_size, opts.background_color)
Martin@134
   428
Martin@845
   429
    seqLimits1 = expandedSeqDict(seqLimits1)
Martin@845
   430
    seqLimits2 = expandedSeqDict(seqLimits2)
Martin@836
   431
    origins1 = expandedSeqDict(origins1)
Martin@836
   432
    origins2 = expandedSeqDict(origins2)
Martin@845
   433
Martin@845
   434
    beds1 = list(readBed(opts.bed1, seqLimits1))
Martin@845
   435
    beds2 = list(readBed(opts.bed2, seqLimits2))
Martin@845
   436
    drawAnnotations(im, beds1, origins1, margin2, height, True, bp_per_pix)
Martin@845
   437
    drawAnnotations(im, beds2, origins2, margin1, width, False, bp_per_pix)
Martin@845
   438
Martin@839
   439
    gaps1 = list(readGaps(opts.gap1, seqLimits1))
Martin@839
   440
    gaps2 = list(readGaps(opts.gap2, seqLimits2))
Martin@650
   441
    # draw bridged gaps first, then unbridged gaps on top:
Martin@836
   442
    drawUnsequencedGaps(im, gaps1, origins1, margin2, height, True, "yes",
Martin@650
   443
                        bp_per_pix, opts.bridged_color)
Martin@836
   444
    drawUnsequencedGaps(im, gaps2, origins2, margin1, width, False, "yes",
Martin@650
   445
                        bp_per_pix, opts.bridged_color)
Martin@836
   446
    drawUnsequencedGaps(im, gaps1, origins1, margin2, height, True, "no",
Martin@650
   447
                        bp_per_pix, opts.unbridged_color)
Martin@836
   448
    drawUnsequencedGaps(im, gaps2, origins2, margin1, width, False, "no",
Martin@650
   449
                        bp_per_pix, opts.unbridged_color)
Martin@650
   450
Martin@643
   451
    for i in range(height):
Martin@643
   452
        for j in range(width):
Martin@643
   453
            store_value = hits[i * width + j]
Martin@643
   454
            xy = j, i
Martin@643
   455
            if   store_value == 1: im.putpixel(xy, forward_color)
Martin@643
   456
            elif store_value == 2: im.putpixel(xy, reverse_color)
Martin@643
   457
            elif store_value == 3: im.putpixel(xy, overlap_color)
Martin@95
   458
Martin@643
   459
    if opts.fontsize != 0:
Martin@846
   460
        axis1 = get_axis_image(seqLabels1, labelSizes1, seq_starts1, seq_pix1,
Martin@647
   461
                               font, image_mode, opts)
Martin@846
   462
        axis2 = get_axis_image(seqLabels2, labelSizes2, seq_starts2, seq_pix2,
Martin@647
   463
                               font, image_mode, opts)
Martin@834
   464
        axis2 = axis2.transpose(Image.ROTATE_270)  # !!! bug hotspot
Martin@643
   465
        im.paste(axis1, (0, 0))
Martin@643
   466
        im.paste(axis2, (0, 0))
Martin@1
   467
Martin@643
   468
    for i in seq_starts1[1:]:
Martin@646
   469
        box = i - opts.pix_tween_seqs, margin2, i, height
Martin@646
   470
        im.paste(opts.border_shade, box)
Martin@1
   471
Martin@643
   472
    for i in seq_starts2[1:]:
Martin@646
   473
        box = margin1, i - opts.pix_tween_seqs, width, i
Martin@646
   474
        im.paste(opts.border_shade, box)
Martin@1
   475
Martin@643
   476
    im.save(args[1])
Martin@648
   477
Martin@648
   478
if __name__ == "__main__":
Martin@649
   479
    usage = """%prog --help
Martin@649
   480
   or: %prog [options] maf-or-tab-alignments dotplot.png
Martin@649
   481
   or: %prog [options] maf-or-tab-alignments dotplot.gif
Martin@649
   482
   or: ..."""
Martin@649
   483
    description = "Draw a dotplot of pair-wise sequence alignments in MAF or tabular format."
Martin@649
   484
    op = optparse.OptionParser(usage=usage, description=description)
Martin@651
   485
    op.add_option("-1", "--seq1", metavar="PATTERN", action="append",
Martin@840
   486
                  default=[],
Martin@651
   487
                  help="which sequences to show from the 1st genome")
Martin@651
   488
    op.add_option("-2", "--seq2", metavar="PATTERN", action="append",
Martin@840
   489
                  default=[],
Martin@651
   490
                  help="which sequences to show from the 2nd genome")
Martin@648
   491
    # Replace "width" & "height" with a single "length" option?
Martin@648
   492
    op.add_option("-x", "--width", type="int", default=1000,
Martin@648
   493
                  help="maximum width in pixels (default: %default)")
Martin@648
   494
    op.add_option("-y", "--height", type="int", default=1000,
Martin@648
   495
                  help="maximum height in pixels (default: %default)")
Martin@649
   496
    op.add_option("-f", "--fontfile", metavar="FILE",
Martin@648
   497
                  help="TrueType or OpenType font file")
Martin@649
   498
    op.add_option("-s", "--fontsize", metavar="SIZE", type="int", default=11,
Martin@648
   499
                  help="TrueType or OpenType font size (default: %default)")
Martin@649
   500
    op.add_option("-c", "--forwardcolor", metavar="COLOR", default="red",
Martin@649
   501
                  help="color for forward alignments (default: %default)")
Martin@649
   502
    op.add_option("-r", "--reversecolor", metavar="COLOR", default="blue",
Martin@649
   503
                  help="color for reverse alignments (default: %default)")
Martin@839
   504
    op.add_option("--trim1", action="store_true",
Martin@839
   505
                  help="trim unaligned sequence flanks from the 1st genome")
Martin@839
   506
    op.add_option("--trim2", action="store_true",
Martin@839
   507
                  help="trim unaligned sequence flanks from the 2nd genome")
Martin@846
   508
    op.add_option("--lengths1", action="store_true",
Martin@846
   509
                  help="show sequence lengths for the 1st genome")
Martin@846
   510
    op.add_option("--lengths2", action="store_true",
Martin@846
   511
                  help="show sequence lengths for the 2nd genome")
Martin@845
   512
    op.add_option("--bed1", metavar="FILE",
Martin@845
   513
                  help="read genome1 annotations from bed file")
Martin@845
   514
    op.add_option("--bed2", metavar="FILE",
Martin@845
   515
                  help="read genome2 annotations from bed file")
Martin@846
   516
Martin@650
   517
    og = optparse.OptionGroup(op, "Unsequenced gap options")
Martin@650
   518
    og.add_option("--gap1", metavar="FILE",
Martin@650
   519
                  help="read genome1 unsequenced gaps from agp or gap file")
Martin@650
   520
    og.add_option("--gap2", metavar="FILE",
Martin@650
   521
                  help="read genome2 unsequenced gaps from agp or gap file")
Martin@650
   522
    og.add_option("--bridged-color", metavar="COLOR", default="yellow",
Martin@650
   523
                  help="color for bridged gaps (default: %default)")
Martin@650
   524
    og.add_option("--unbridged-color", metavar="COLOR", default="pink",
Martin@650
   525
                  help="color for unbridged gaps (default: %default)")
Martin@650
   526
    op.add_option_group(og)
Martin@648
   527
    (opts, args) = op.parse_args()
Martin@648
   528
    if len(args) != 2: op.error("2 arguments needed")
Martin@648
   529
Martin@648
   530
    opts.text_color = "black"
Martin@648
   531
    opts.background_color = "white"
Martin@648
   532
    opts.pix_tween_seqs = 2  # number of border pixels between sequences
Martin@648
   533
    opts.border_shade = 239, 239, 239  # the shade of grey for border pixels
Martin@648
   534
    opts.label_space = 5     # minimum number of pixels between axis labels
Martin@648
   535
Martin@649
   536
    try: lastDotplot(opts, args)
Martin@649
   537
    except KeyboardInterrupt: pass  # avoid silly error message
Martin@649
   538
    except Exception, e:
Martin@649
   539
        prog = os.path.basename(sys.argv[0])
Martin@649
   540
        sys.exit(prog + ": error: " + str(e))