lib.py 48.3 KB
Newer Older
jhoogenboom's avatar
jhoogenboom committed
1
2
#!/usr/bin/env python

3
import re, sys, argparse, random
jhoogenboom's avatar
jhoogenboom committed
4
#import numpy as np  # Imported only when calling nnls()
jhoogenboom's avatar
jhoogenboom committed
5
6

from ConfigParser import RawConfigParser, MissingSectionHeaderError
7
from StringIO import StringIO
jhoogenboom's avatar
jhoogenboom committed
8
9
10
11

# Patterns that match entire sequences.
PAT_SEQ_RAW = re.compile("^[ACGT]*$")
PAT_SEQ_TSSV = re.compile("^(?:[ACGT]+\(\d+\))*$")
12
13
PAT_SEQ_ALLELENAME = re.compile(  # First line: n_ACT[m] or alias.
    "^(?:(?:(?:CE)?\d+(?:\.\d+)?_(?:[ACGT]+\[\d+\])+)|((?!_).+?))"
jhoogenboom's avatar
jhoogenboom committed
14
15
16
    "(?:_[-+]\d+(?:\.1)?(?P<a>(?:(?<=\.1)-)|(?<!\.1)[ACGT]+)>"  # _+3A>
        "(?!(?P=a))(?:[ACTG]+|-))*$")  # Portion of variants after '>'.

17
# Patterns that match blocks of TSSV-style sequences and allele names.
jhoogenboom's avatar
jhoogenboom committed
18
19
PAT_TSSV_BLOCK = re.compile("([ACGT]+)\((\d+)\)")
PAT_ALLELENAME_BLOCK = re.compile("([ACGT]+)\[(\d+)\]")
20
PAT_ALIAS = re.compile("^(?!_).+$")
jhoogenboom's avatar
jhoogenboom committed
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35

# Pattern that matches a single prefix/suffix variant.
PAT_VARIANT = re.compile(
    "^([-+]\d+)(?:\.1)?"  # Position number
    "(?P<a>(?:(?<=\.1)-)|(?<!\.1)[ACGT]+)>"  # From part
    "(?!(?P=a))([ACTG]+|-)$")  # To part

# Patterns that match (parts of) an STR definition.
PAT_STR_DEF = re.compile(
    "^(?:[ACGT]+\s+\d+\s+\d+(?:\s+[ACGT]+\s+\d+\s+\d+)*)?$")
PAT_STR_DEF_BLOCK = re.compile("([ACGT]+)\s+(\d+)\s+(\d+)")

# Pattern to split a comma-, semicolon-, or space-separated list.
PAT_SPLIT = re.compile("[,; ]+")

jhoogenboom's avatar
jhoogenboom committed
36
37
# Default regular expression to capture sample tags in file names.
# This is the default of the -e command line option.
jhoogenboom's avatar
jhoogenboom committed
38
DEF_TAG_EXPR = "^(.*?)(?:\.[^.]+)?$"
jhoogenboom's avatar
jhoogenboom committed
39
40
41
42
43

# Default formatting template to write sample tags.
# This is the default of the -f command line option.
DEF_TAG_FORMAT = "\\1"

44
45
46
47
48
# Default formatting template to construct output file names for batch
# processing.  \1 and \2 refer to sample tag and tool name.
# This is the default for the -o command line option with batch support.
DEF_OUTFILE_FORMAT = "\\1-\\2.out"

jhoogenboom's avatar
jhoogenboom committed
49
50
51
52
53
54
# IUPAC Table of complementary bases.
COMPL = {"A": "T", "T": "A", "U": "A", "G": "C", "C": "G", "R": "Y", "Y": "R",
         "K": "M", "M": "K", "B": "V", "V": "B", "D": "H", "H": "D",
         "a": "t", "t": "a", "u": "a", "g": "c", "c": "g", "r": "y", "y": "r",
         "k": "m", "m": "k", "b": "v", "v": "b", "d": "h", "h": "d"}

jhoogenboom's avatar
jhoogenboom committed
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224

def call_variants(template, sequence, reverse_indices=False, cache=True,
                  debug=False):
    """
    Perform a global alignment of sequence to template and return a
    list of variants detected.  All variants are given as substitutions
    in the form posX>Y, where the first base in the template is pos=1.
    Set reverse_indices to True to count from right to left instead.
    Insertions and deletions are pos.1->Y and posX>-, respectively.

    By default, the results of this function are cached.  Set cache to
    False to suppress caching the result and reduce memory usage.

    Setting debug to True will cause the alignment matrices to be
    printed to sys.stdout.  Be aware that they can be quite large.
    """
    # Saving the results in a cache to avoid repeating alignments.
    try:
        return call_variants.cache[template, sequence, reverse_indices]
    except KeyError:
        pass

    row_offset = len(template) + 1
    matrix_match = [0] * row_offset * (len(sequence)+1)
    matrix_gap1 = [-sys.maxint-1] * row_offset * (len(sequence)+1)
    matrix_gap2 = [-sys.maxint-1] * row_offset * (len(sequence)+1)

    MATCH_SCORE = 1
    MISMATCH_SCORE = -1
    GAP_OPEN_SCORE = -10
    GAP_EXTEND_SCORE = -1

    for i in range(len(matrix_match)):
        x = i % row_offset
        y = i / row_offset

        # Initialisation of first row and column.
        if x == 0 or y == 0:
            if x != 0:
                # Top row.
                matrix_gap1[i] = GAP_OPEN_SCORE + GAP_EXTEND_SCORE * (x - 1)
                matrix_match[i] = matrix_gap1[i]
            elif y != 0:
                # Left column.
                matrix_gap2[i] = GAP_OPEN_SCORE + GAP_EXTEND_SCORE * (y - 1)
                matrix_match[i] = matrix_gap2[i]
            continue

        matrix_gap1[i] = max(
            matrix_match[i-1] + GAP_OPEN_SCORE,
            matrix_gap1[i-1] + GAP_EXTEND_SCORE)
        matrix_gap2[i] = max(
            matrix_match[i-row_offset] + GAP_OPEN_SCORE,
            matrix_gap2[i-row_offset] + GAP_EXTEND_SCORE)

        if template[x-1] == sequence[y-1]:
            match = MATCH_SCORE
        else:
            match = MISMATCH_SCORE

        matrix_match[i] = max(
            matrix_match[i-1-row_offset] + match,
            matrix_gap1[i],
            matrix_gap2[i])

    if debug:
        print("GAP1")
        for i in range(0, len(matrix_gap1), row_offset):
            print(("%5i" * row_offset) % tuple(matrix_gap1[i:i+row_offset]))
        print("GAP2")
        for i in range(0, len(matrix_gap2), row_offset):
            print(("%5i" * row_offset) % tuple(matrix_gap2[i:i+row_offset]))
        print("Match")
        for i in range(0, len(matrix_match), row_offset):
            print(("%5i" * row_offset) % tuple(matrix_match[i:i+row_offset]))


    # Backtracking.
    variants = []
    variant_template = 0
    variant_sequence = 0
    i = len(matrix_match) - 1
    while i >= 0:
        x = i % row_offset
        y = i / row_offset

        if matrix_gap1[i] == matrix_match[i]:
            # Go horizontally.  Deletion.
            variant_template += 1
            i -= 1
            continue

        if matrix_gap2[i] == matrix_match[i]:
            # Go vertically.  Insertion.
            variant_sequence += 1
            i -= row_offset
            continue

        # Only backtracking diagonally if a gap is out of the question.
        # Go diagonally.  Either match or mismatch.
        if i == 0 or template[x - 1] == sequence[y - 1]:
            # Match.  Flush variants.
            if variant_template or variant_sequence:
                if variant_template == 0:
                    # Insertions: "-131.1->C" instead of "-130->C".
                    variants.append("%+i.1->%s" % (
                        x - int(reverse_indices) * row_offset,
                        sequence[y:y+variant_sequence]))
                else:
                    variants.append("%+i%s>%s" % (
                        (x + 1) - int(reverse_indices) * row_offset,
                        template[x:x+variant_template],
                        sequence[y:y+variant_sequence] or "-"))
                variant_template = 0
                variant_sequence = 0
        else:
            # Start/extend mismatch.
            variant_template += 1
            variant_sequence += 1
        i -= 1 + row_offset

    # If reverse_indices=False, we need to reverse the output instead.
    if not reverse_indices:
        variants.reverse()

    # Store the result in the cache.
    if cache:
        call_variants.cache[template, sequence, reverse_indices] = variants
    return variants
#call_variants
call_variants.cache = {}


def mutate_sequence(sequence, variants):
    """Apply the given variants to the given sequence."""
    if not sequence and len(variants) > 1:
        raise ValueError("With an empty sequence, only a single variant is "
                         "possible: an insertion '+0.1->x' or '-1.1->x'.")
    sequence = list(sequence)
    for variant in variants:
        vm = PAT_VARIANT.match(variant)
        if vm is None:
            raise ValueError("Unrecognised variant '%s'" % variant)
        pos = int(vm.group(1))
        old = vm.group(2)
        new = vm.group(3)
        if old == "-":
            old = ""
        if new == "-":
            new = ""
        if pos < 0:
            pos += len(sequence) + 1
        if pos == 0 and len(sequence) == 0:
            # Insertion into empty sequence.
            return new
        if old != "".join(sequence[pos-1:pos+len(old)-1]):
            raise ValueError("Incorrect original sequence in variant '%s'; "
                             "should be '%s'!" % (variant,
                               "".join(sequence[pos-1:pos+len(old)-1]) or "-"))
        sequence[pos-1:pos+len(old)-1] = [""] * len(old)
        if pos:
            sequence[pos-1] += new
        else:
            # Insertion at the beginning of the sequence
            sequence[0] = new + sequence[0]
    return "".join(sequence)
#mutate_sequence


def parse_library(handle):
225
226
227
    if handle == sys.stdin:
        # Can't seek on pipes, so read it into a buffer first.
        handle = StringIO(sys.stdin.read())
jhoogenboom's avatar
jhoogenboom committed
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
    try:
        return parse_library_ini(handle)
    except MissingSectionHeaderError:
        # Not an ini file.
        pass
    handle.seek(0)
    return parse_library_tsv(handle)
#parse_library


def parse_library_tsv(handle):
    """
    Parse a TSSV library file (tab-separated values format).

    The provided file should contain at least four columns: marker name,
    left flanking sequence, right flanking sequence, and STR definition.

    Return a nested dict with top-level keys "flanks" and "regex".
    """
    library = {
      "flanks": {},
249
250
      "regex": {},
      "regex_middle": {}
jhoogenboom's avatar
jhoogenboom committed
251
252
    }
    for line in handle:
jhoogenboom's avatar
jhoogenboom committed
253
        line = [x.strip() for x in line.rstrip("\r\n").split("\t")]
jhoogenboom's avatar
jhoogenboom committed
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
        if line == [""]:
            continue
        if len(line) < 4:
            raise ValueError(
                "Invalid library file: encountered line with %i columns, "
                "need at least 4" % len(line))
        marker = line[0]
        if PAT_SEQ_RAW.match(line[1]) is None:
            raise ValueError("Flanking sequence '%s' of marker %s is invalid" %
                             (line[1], marker))
        if PAT_SEQ_RAW.match(line[2]) is None:
            raise ValueError("Flanking sequence '%s' of marker %s is invalid" %
                             (line[2], marker))
        if PAT_STR_DEF.match(line[3]) is None:
            raise ValueError("STR definition '%s' of marker %s is invalid" %
                             (line[3], marker))
        library["flanks"][marker] = line[1:3]
jhoogenboom's avatar
jhoogenboom committed
271
272
        library["regex_middle"][marker] = re.compile("".join(
            "(%s){%s,%s}" % x for x in PAT_STR_DEF_BLOCK.findall(line[3])))
273
274
        library["regex"][marker] = re.compile(
            "".join(["^", library["regex_middle"][marker].pattern, "$"]))
jhoogenboom's avatar
jhoogenboom committed
275
276
277
278
279
280
281
282
283
284
    return library
#parse_library_tsv


def parse_library_ini(handle):
    library = {
      "flanks": {},
      "prefix": {},
      "suffix": {},
      "regex": {},
285
      "regex_middle": {},
286
287
      "length_adjust": {},
      "block_length": {},
jhoogenboom's avatar
jhoogenboom committed
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
      "aliases": {}
    }
    markers = set()

    ini = RawConfigParser()
    ini.optionxform = str
    ini.readfp(handle)
    for section in ini.sections():
        for marker in ini.options(section):
            value = ini.get(section, marker)
            if section == "flanks":
                values = PAT_SPLIT.split(value)
                if len(values) != 2:
                    raise ValueError(
                        "For marker %s, %i flanking sequences were given,"
                        "need exactly 2" % (marker, len(values)))
                for value in values:
                    if PAT_SEQ_RAW.match(value) is None:
                        raise ValueError(
                            "Flanking sequence '%s' of marker %s is invalid" %
                            (value, marker))
                library["flanks"][marker] = values
                markers.add(marker)
            elif section == "prefix":
                values = PAT_SPLIT.split(value)
                for value in values:
                    if PAT_SEQ_RAW.match(value) is None:
                        raise ValueError(
                            "Prefix sequence '%s' of marker %s is invalid" %
                            (value, marker))
318
                library["prefix"][marker] = values
jhoogenboom's avatar
jhoogenboom committed
319
320
321
322
323
324
325
326
                markers.add(marker)
            elif section == "suffix":
                values = PAT_SPLIT.split(value)
                for value in values:
                    if PAT_SEQ_RAW.match(value) is None:
                        raise ValueError(
                            "Suffix sequence '%s' of marker %s is invalid" %
                            (value, marker))
327
                library["suffix"][marker] = values
jhoogenboom's avatar
jhoogenboom committed
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
                markers.add(marker)
            elif section == "length_adjust":
                try:
                    value = int(value)
                except:
                    raise ValueError(
                        "Length adjustment '%s' of marker %s is not a valid "
                        "integer" % (value, marker))
                library["length_adjust"][marker] = value
                markers.add(marker)
            elif section == "block_length":
                try:
                    value = int(value)
                except:
                    raise ValueError(
                        "Block length '%s' of marker %s is not a valid integer"
                        % (value, marker))
                library["block_length"][marker] = value
                markers.add(marker)
            elif section == "aliases":
                values = PAT_SPLIT.split(value)
                if len(values) != 3:
                    raise ValueError("Alias %s does not have 3 values, but %i"
                                     % (marker, len(values)))
                if PAT_SEQ_RAW.match(values[1]) is None:
353
354
355
356
357
358
359
                    raise ValueError(
                        "Alias sequence '%s' of alias %s is invalid" %
                        (values[1], marker))
                if PAT_ALIAS.match(values[2]) is None:
                    raise ValueError(
                        "Allele name '%s' of alias %s is invalid" %
                        (values[2], marker))
jhoogenboom's avatar
jhoogenboom committed
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
                library["aliases"][marker] = {
                    "marker": values[0],
                    "sequence": values[1],
                    "name": values[2]
                }
                markers.add(marker)
            elif section == "repeat":
                if PAT_STR_DEF.match(value) is None:
                    raise ValueError(
                        "STR definition '%s' of marker %s is invalid" %
                        (value, marker))
                library["regex"][marker] = value
                markers.add(marker)

    # Compile regular expressions.
    # NOTE: The libconvert tool expects "(seq){num,num}" blocks ONLY!
    # TODO: Should a single prefix/suffix be required (i.e., seq{1,1})?
    #       Then also update libconvert when converting to TSSV format.
    for marker in markers:
        parts = []
380
        partsm = []
jhoogenboom's avatar
jhoogenboom committed
381
        if marker in library["prefix"]:
jhoogenboom's avatar
jhoogenboom committed
382
            parts += ("(%s){0,1}" % x for x in library["prefix"][marker])
jhoogenboom's avatar
jhoogenboom committed
383
384
        if marker in library["aliases"]:
            parts.append("(%s){0,1}" % library["aliases"][marker]["sequence"])
385
            partsm.append("(%s){0,1}" % library["aliases"][marker]["sequence"])
jhoogenboom's avatar
jhoogenboom committed
386
        elif marker in library["regex"]:
jhoogenboom's avatar
jhoogenboom committed
387
388
            partsm = ("(%s){%s,%s}" % x for x in
                      PAT_STR_DEF_BLOCK.findall(library["regex"][marker]))
389
            parts += partsm
jhoogenboom's avatar
jhoogenboom committed
390
        if marker in library["suffix"]:
jhoogenboom's avatar
jhoogenboom committed
391
            parts += ("(%s){0,1}" % x for x in library["suffix"][marker])
jhoogenboom's avatar
jhoogenboom committed
392
393
394
        if parts:
            library["regex"][marker] = re.compile(
                "".join(["^"] + parts + ["$"]))
395
396
        if partsm:
            library["regex_middle"][marker] = re.compile("".join(partsm))
jhoogenboom's avatar
jhoogenboom committed
397
398
    return library
#parse_library_ini
jhoogenboom's avatar
jhoogenboom committed
399
400


401
def load_profiles(profilefile, library=None):
jhoogenboom's avatar
jhoogenboom committed
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
    if profilefile == sys.stdin:
        # Can't seek on pipes, so read it into a buffer first.
        profilefile = StringIO(sys.stdin.read())
    headline = profilefile.readline().rstrip("\r\n").split("\t")
    profilefile.seek(0)
    try:
        get_column_ids(headline, "marker", "allele", "sequence", "fmean",
            "rmean")
    except ValueError:
        try:
            int(headline[1])
        except:
            raise ValueError(
                "Invalid background noise profiles file: unable to determine "
                "file format!")
        return load_profiles_crosstab(profilefile, library)
    return load_profiles_columnar(profilefile, library)
#load_profiles


def load_profiles_columnar(profilefile, library=None):
    column_names = profilefile.readline().rstrip("\r\n").split("\t")
    colid_marker, colid_allele, colid_sequence, colid_fmean, colid_rmean = \
        get_column_ids(column_names, "marker", "allele", "sequence", "fmean",
            "rmean")

    profiles = {}
    for line in profilefile:
        line = line.rstrip("\r\n").split("\t")
        if line == [""]:
            continue
        marker = line[colid_marker]
        if marker not in profiles:
            profiles[marker] = {
                "m": set(),  # To be replaced by its length below.
                "n": set(),  # To be replaced by its length below.
                "seqs": [],
                "forward": {},  # To be replaced by a list below.
                "reverse": {}  # To be replaced by a list below.
                }
        allele = ensure_sequence_format(line[colid_allele], "raw",
            library=library, marker=marker)
        sequence = ensure_sequence_format(line[colid_sequence], "raw",
            library=library, marker=marker)
        if (allele, sequence) in profiles[marker]["forward"]:
            raise ValueError(
                "Invalid background noise profiles file: encountered "
                "multiple values for marker '%s' allele '%s' sequence '%s'" %
                (marker, allele, sequence))
        profiles[marker]["forward"][allele,sequence] = float(line[colid_fmean])
        profiles[marker]["reverse"][allele,sequence] = float(line[colid_rmean])
        profiles[marker]["m"].update((allele, sequence))
        profiles[marker]["n"].add(allele)

    # Check completeness and reorder true alleles.
    for marker in profiles:
        profiles[marker]["seqs"] = list(profiles[marker]["n"]) + \
            list(profiles[marker]["m"]-profiles[marker]["n"])
        profiles[marker]["n"] = len(profiles[marker]["n"])
        profiles[marker]["m"] = len(profiles[marker]["m"])
        newprofiles = {"forward": [], "reverse": []}
        for i in range(profiles[marker]["n"]):
            allele = profiles[marker]["seqs"][i]
            for direction in newprofiles:
                newprofiles[direction].append([0] * profiles[marker]["m"])
            for j in range(profiles[marker]["m"]):
                sequence = profiles[marker]["seqs"][j]
                if (allele, sequence) in profiles[marker]["forward"]:
                    for direction in newprofiles:
                        newprofiles[direction][i][j] = \
                            profiles[marker][direction][allele, sequence]
        profiles[marker]["forward"] = newprofiles["forward"]
        profiles[marker]["reverse"] = newprofiles["reverse"]

    return profiles
#load_profiles_columnar


def load_profiles_crosstab(profilefile, library=None):
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
    profiles = {}

    # Read the profile file without assuming it is sorted.
    for line in profilefile:
        line = line.rstrip("\r\n").split("\t")
        if line == [""]:
            continue
        if len(line) < 3:
            raise ValueError(
                "Invalid background noise profiles file: encountered line "
                "with %i columns, need at least 3" % len(line))
        marker = line[0]
        try:
            num = int(line[1])
        except ValueError:
            raise ValueError(
                "Invalid background noise profiles file: encountered "
                "non-number '%s' in the second column" % line[1])
        values = line[2:]
        if marker not in profiles:
            profiles[marker] = {
                "m": len(values),
                "n": abs(num),
                "seqs": [],
jhoogenboom's avatar
jhoogenboom committed
505
506
                "forward": {},  # To be replaced by a list below.
                "reverse": {}  # To be replaced by a list below.
507
508
509
510
511
512
513
514
515
                }
        elif len(values) != profiles[marker]["m"]:
            raise ValueError(
                "Invalid background noise profiles file: profiles of "
                "marker '%s'% have inconsistent lengths" % marker)
        profiles[marker]["n"] = max(abs(num), profiles[marker]["n"])
        if num == 0:
            if profiles[marker]["seqs"]:
                raise ValueError(
jhoogenboom's avatar
jhoogenboom committed
516
                    "Invalid background noise profiles file: encountered "
517
                    "multiple header lines for marker '%s'" % marker)
jhoogenboom's avatar
jhoogenboom committed
518
519
            values = [ensure_sequence_format(seq, "raw", library=library,
                        marker=marker) for seq in values]
520
521
522
523
524
            profiles[marker]["seqs"] = values
        else:
            direction = "forward" if num > 0 else "reverse"
            if abs(num) in profiles[marker][direction]:
                raise ValueError(
jhoogenboom's avatar
jhoogenboom committed
525
                    "Invalid background noise profiles file: encountered "
526
527
528
529
530
531
532
533
534
535
                    "multiple %s profiles for marker '%s' allele %i" %
                    (direction, marker, abs(num)))
            values = map(float, values)
            profiles[marker][direction][abs(num)] = values

    # Check completeness and reorder true alleles.
    for marker in profiles:
        newprofiles = {"forward": [], "reverse": []}
        if not profiles[marker]["seqs"]:
            raise ValueError(
jhoogenboom's avatar
jhoogenboom committed
536
                "Invalid background noise profiles file: missing header line "
537
538
539
540
541
                "for marker '%s'" % marker)
        for i in range(1, profiles[marker]["n"] + 1):
            for direction in newprofiles:
                if i not in profiles[marker][direction]:
                    raise ValueError(
jhoogenboom's avatar
jhoogenboom committed
542
                        "Invalid background noise profiles file: missing %s "
543
544
545
546
547
548
549
                        "profile for marker '%s' allele %i" %
                        (direction, marker, i))
                newprofiles[direction].append(profiles[marker][direction][i])
        profiles[marker]["forward"] = newprofiles["forward"]
        profiles[marker]["reverse"] = newprofiles["reverse"]

    return profiles
jhoogenboom's avatar
jhoogenboom committed
550
#load_profiles_crosstab
551
552


553
554
555
556
557
558
559
560
def regex_longest_match(pattern, subject):
    """Return the longest match of the pattern in the subject string."""
    return reduce(lambda x, y:
        y if x is None or x.end()-x.start() < y.end()-y.start() else x,
        pattern.finditer(subject), None)
#regex_longest_match


jhoogenboom's avatar
jhoogenboom committed
561
562
563
def detect_sequence_format(seq):
    """Return format of seq.  One of 'raw', 'tssv', or 'allelename'."""
    if not seq:
jhoogenboom's avatar
jhoogenboom committed
564
        raise ValueError("Empty sequence")
jhoogenboom's avatar
jhoogenboom committed
565
566
567
568
569
570
    if PAT_SEQ_RAW.match(seq):
        return 'raw'
    if PAT_SEQ_TSSV.match(seq):
        return 'tssv'
    if PAT_SEQ_ALLELENAME.match(seq):
        return 'allelename'
jhoogenboom's avatar
jhoogenboom committed
571
    raise ValueError("Unrecognised sequence format")
jhoogenboom's avatar
jhoogenboom committed
572
573
574
#detect_sequence_format


jhoogenboom's avatar
jhoogenboom committed
575
def convert_sequence_tssv_raw(seq):
jhoogenboom's avatar
jhoogenboom committed
576
577
    return "".join(block[0] * int(block[1])
                   for block in PAT_TSSV_BLOCK.findall(seq))
jhoogenboom's avatar
jhoogenboom committed
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
#convert_sequence_tssv_raw


def convert_sequence_raw_tssv(seq, library, marker, return_alias=False):
    # Try to match this marker's pattern, or any of its aliases.
    match = None
    if "aliases" in library:
        for alias in library["aliases"]:
            if (library["aliases"][alias]["marker"] == marker and
                    alias in library["regex"]):
                match = library["regex"][alias].match(seq)
                if match is not None:
                    marker = alias
                    break
    if match is None and marker in library["regex"]:
        match = library["regex"][marker].match(seq)
594
    if match is not None:
jhoogenboom's avatar
jhoogenboom committed
595
596
        parts = ((match.group(i), match.end(i))
                 for i in range(1, match.lastindex+1) if match.group(i))
jhoogenboom's avatar
jhoogenboom committed
597

598
    # Use heuristics if the sequence does not match the pattern.
jhoogenboom's avatar
jhoogenboom committed
599
    else:
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624

        # Find explictily provided prefix and/or suffix if present.
        pre_suf = ["", ""]
        if "prefix" in library and marker in library["prefix"]:
            for prefix in library["prefix"][marker]:
                if seq.startswith(prefix):
                    pre_suf[0] = prefix
                    seq = seq[len(prefix):]
                    break
        if "suffix" in library and marker in library["suffix"]:
            for suffix in library["suffix"][marker]:
                if seq.endswith(suffix):
                    pre_suf[1] = suffix
                    seq = seq[:-len(suffix)]
                    break

        # Find longest match of middle pattern.
        middle = [(seq, len(pre_suf[0])+len(seq))]
        if marker in library["regex_middle"]:
            match = regex_longest_match(library["regex_middle"][marker], seq)
            if match is not None and match.end()-match.start():
                # TODO: If I did not have a prefix yet, but the
                # canonical prefix ends with the first few blocks that
                # we obtained in the match, move these blocks into the
                # prefix.  Also do this with the suffix.
jhoogenboom's avatar
jhoogenboom committed
625
626
                middle = ((match.group(i), match.end(i)+len(pre_suf[0]))
                    for i in range(1, match.lastindex+1) if match.group(i))
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
                pre_suf[0] += seq[:match.start()]
                pre_suf[1] = seq[match.end():] + pre_suf[1]
                seq = match.group()

        # Now construct parts.
        parts = []
        if pre_suf[0]:
            parts.append((pre_suf[0], len(pre_suf[0])))
        parts += middle
        if pre_suf[1]:
            parts.append((pre_suf[1], sum(map(len,pre_suf))+len(seq)))

    seq = reduce(
        lambda a, b: (a[0] + "%s(%i)" % (b[0], (b[1]-a[1])/len(b[0])), b[1]),
        reduce(
            lambda x, y: x[:-1] + [y] if x[-1][0] == y[0] else x + [y],
            parts,
            [("", 0)]))[0]
jhoogenboom's avatar
jhoogenboom committed
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
    return (seq, marker) if return_alias else seq
#convert_sequence_raw_tssv


def convert_sequence_allelename_tssv(seq, library, marker):
    # Check whether there is an alias for this sequence.
    if "aliases" in library:
        for alias in library["aliases"]:
            if library["aliases"][alias]["marker"] == marker and (
                    seq == library["aliases"][alias]["name"] or
                    seq.startswith(library["aliases"][alias]["name"] + "_")):
                marker = alias
                seq = "".join([
                    "0_",
                    library["aliases"][alias]["sequence"] + "[1]",
                    seq[len(library["aliases"][alias]["name"]):]])
                break

663
664
665
666
667
    # It should NOT be an alias now.
    match = PAT_SEQ_ALLELENAME.match(seq)
    if match is None or match.group(1) is not None:
        raise ValueError("Invalid allele name '%s' encountered!" % seq)

jhoogenboom's avatar
jhoogenboom committed
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
    allele = seq.split("_")

    # Get and mutate prefix and suffix.
    if "prefix" in library and marker in library["prefix"]:
        prefix = library["prefix"][marker][0]
    else:
        prefix = ""
    if "suffix" in library and marker in library["suffix"]:
        suffix = library["suffix"][marker][0]
    else:
        suffix = ""
    variants = [[], []]
    for variant in allele[2:]:
        if variant[0] == "-":
            if not prefix:
                raise ValueError("Encountered prefix variant '%s', but marker "
                                 "'%s' has no prefix!" % (variant, marker))
            variants[0].append(variant)
        elif variant[0] == "+":
            if not suffix:
                raise ValueError("Encountered suffix variant '%s', but marker "
                                 "'%s' has no suffix!" % (variant, marker))
            variants[1].append(variant)
        else:
            raise ValueError("Unrecognised variant '%s'" % variant)
    if variants[0]:
        prefix = mutate_sequence(prefix, variants[0])
    if variants[1]:
        suffix = mutate_sequence(suffix, variants[1])

    blocks = []
    if prefix:
        blocks.append((prefix, 1))
    for block in PAT_ALLELENAME_BLOCK.findall(allele[1]):
        blocks.append((block[0], int(block[1])))
    if suffix:
        blocks.append((suffix, 1))
jhoogenboom's avatar
jhoogenboom committed
705
    return "".join("%s(%i)" % block for block in blocks)
jhoogenboom's avatar
jhoogenboom committed
706
707
708
709
710
711
#convert_sequence_allelename_tssv


def convert_sequence_raw_allelename(seq, library, marker):
    # We actually convert raw->allelename via TSSV format.
    seq, alias = convert_sequence_raw_tssv(seq, library, marker, True)
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
    blocks = PAT_TSSV_BLOCK.findall(seq)

    # Find prefix and suffix.
    prefix = suffix = this_prefix = this_suffix = ""
    remaining_blocks = len(blocks)
    if "prefix" in library and marker in library["prefix"]:
        prefix = library["prefix"][marker][0]
        remaining_blocks -= 1
    if "suffix" in library and marker in library["suffix"]:
        suffix = library["suffix"][marker][0]
        remaining_blocks -= 1
    if remaining_blocks > 0 and prefix and blocks[0][1] == "1":
        this_prefix = blocks[0][0]
        blocks = blocks[1:]
    if remaining_blocks > 0 and suffix and blocks[-1][1] == "1":
        this_suffix = blocks[-1][0]
        blocks = blocks[:-1]
jhoogenboom's avatar
jhoogenboom committed
729
730
731
732

    # Generate prefix/suffix variants.
    length = 0
    variants = []
733
734
735
736
737
738
    if prefix != this_prefix:
        variants += call_variants(prefix, this_prefix, True)
        length += len(this_prefix) - len(prefix)
    if suffix != this_suffix:
        variants += call_variants(suffix, this_suffix, False)
        length += len(this_suffix) - len(suffix)
jhoogenboom's avatar
jhoogenboom committed
739
740
741
742
743
744
745
746
747

    # We are ready to return the allele name of aliases.
    if alias != marker:
        return "_".join([library["aliases"][marker]["name"]] + variants)

    # Compute CE allele number for the other alleles.
    # TODO: perhaps produce a more intelligent name if there is exactly
    #       one alias with the same length
    blocknames = []
748
749
    blocksize = library.get("block_length", {}).get(marker, 4)
    length -= library.get("length_adjust", {}).get(marker, 0)
jhoogenboom's avatar
jhoogenboom committed
750
751
752
753
754
755
756
757
758
759
760
761
762
763
    for block in blocks:
        blocknames.append("%s[%s]" % (block[0], block[1]))
        length += len(block[0]) * int(block[1])

    allelename = "CE" + str(length / blocksize)
    if length % blocksize:
        allelename += "." + str(length % blocksize)
    return "_".join([allelename, "".join(blocknames)] + variants)
#convert_sequence_raw_allelename


def ensure_sequence_format(seq, to_format, from_format=None, library=None,
                           marker=None):
    """Convert seq to 'raw', 'tssv', or 'allelename' format."""
jhoogenboom's avatar
jhoogenboom committed
764
    known_formats = ("raw", "tssv", "allelename")
jhoogenboom's avatar
jhoogenboom committed
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
    if to_format not in known_formats:
        raise ValueError("Unknown format '%s', choose from %s" %
                         (to_format, known_formats))
    if from_format is None:
        from_format = detect_sequence_format(seq) if seq else "raw"
    elif from_format not in known_formats:
        raise ValueError("Unknown format '%s', choose from %s" %
                         (from_format, known_formats))

    # No conversion needed?
    if to_format == from_format:
        return seq

    # From TSSV to raw sequence is easy.
    # We'll need a library and marker name for anything else.
    if (library is None or marker is None) and (from_format != "tssv" or
            to_format != "raw"):
        raise ValueError("Sequence needs to be converted from %s to %s, this "
                          "conversion requires a library file" %
                          (from_format, to_format))

    # Perform conversions.
    if from_format == "allelename":
        seq = convert_sequence_allelename_tssv(seq, library, marker)
    if to_format == "tssv":
        if from_format == "raw":
            return convert_sequence_raw_tssv(seq, library, marker)
        return seq
    if from_format != "raw":
        seq = convert_sequence_tssv_raw(seq)
    if to_format == "raw":
        return seq
    return convert_sequence_raw_allelename(seq, library, marker)
#ensure_sequence_format


jhoogenboom's avatar
jhoogenboom committed
801
802
def reverse_complement(sequence):
    """Return the reverse complement of the given DNA sequence."""
jhoogenboom's avatar
jhoogenboom committed
803
    return "".join(COMPL[x] if x in COMPL else x for x in reversed(sequence))
jhoogenboom's avatar
jhoogenboom committed
804
805
806
#reverse_complement


jhoogenboom's avatar
jhoogenboom committed
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
def nnls(A, C, B=None, max_iter=200, min_change=0.0001, debug=False):
    """
    Solve for B in A * B = C in the least squares sense, s.t. B >= 0.

    Hint: call nnls(B.T, C.T).T to solve for A.

    Algorithm has converged if the sum of squared error has decreased
    by less than a factor of min_change in one iteration.  If debug is
    True, print the sum of squared error to sys.stdout after each
    iteration.

    This code was partially adopted from nimfa.methods.factorization.bd.
    """
    import numpy as np
    if B is None:
        B = np.matrix(np.zeros([A.shape[1], C.shape[1]]))
    E = A.T * A
    F = A.T * C
    prev_score = cur_score = sys.float_info.max
    for i in range(max_iter):
        for n in range(B.shape[0]):
            nn = list(range(n)) + list(range(n + 1, B.shape[0]))
            tmp = (F[n, :] - E[n, nn] * B[nn, :]) / E[n, n]
            tmp[np.isnan(tmp)] = 0
            tmp[tmp < 0] = 0
            B[n, :] = tmp
        prev_score = cur_score
        cur_score = np.square(C - A * B).sum()
        score_change = (prev_score-cur_score)/prev_score

        if debug:
            if i:
                print("%4i %15.6f %15.6f %6.2f" % (i, cur_score,
                    prev_score-cur_score, 100*score_change))
            else:
                print("%4i %15.6f" % (i, cur_score))

        if not cur_score or score_change < min_change:
            # We have converged.
            break

    return B
#nnls


jhoogenboom's avatar
jhoogenboom committed
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
def adjust_stats(value, stats=None):
    """
    Adjust the given stats in place with the given observed value and
    return the adjusted stats as well.  If no stats dict is given,
    create a new stats dict with the following initial values:
    {"n": 1, "min": value, "max": value, "mean": value, "m2": 0.0,
     "variance": 0.0}
    """
    value += 0.0
    if not stats:
        return {"n": 1, "min": value, "max": value, "mean": value, "m2": 0.0,
                "variance": 0.0}
    stats["n"] += 1
    delta = value - stats["mean"]
    stats["mean"] += delta / stats["n"]
    stats["m2"] += delta * (value - stats["mean"])
    try:
        stats["variance"] = stats["m2"] / (stats["n"] - 1)
        stats["min"] = min(stats["min"], value)
        stats["max"] = max(stats["max"], value)
    except ZeroDivisionError:
        stats["variance"] = 0
        stats["min"] = value
        stats["max"] = value
    return stats
#adjust_stats


jhoogenboom's avatar
jhoogenboom committed
880
881
882
883
884
885
886
887
888
889
890
def get_repeat_pattern(seq):
    """Return compiled regular expression that finds repeats of seq."""
    return re.compile("".join(             # For AGAT, one obtains:
        ["(?:" * (len(seq)-1)] +           # (?:(?:(?:
        ["%s)?" % x for x in seq[1:]] +    # G)?A)?T)?
        ["(?:", seq, ")+"] +               # (?AGAT)+
        ["(?:%s" % x for x in seq[:-1]] +  # (?:A(?:G(?:A
        [")?" * (len(seq)-1)]))            # )?)?)?
#get_repeat_pattern


891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
def read_sample_data_file(infile, data, annotation_column=None, seqformat=None,
                          library=None, default_marker=None):
    """Add data from infile to data dict as [marker, allele]=reads."""
    # Get column numbers.
    column_names = infile.readline().rstrip("\r\n").split("\t")
    colid_allele, colid_forward, colid_reverse = \
        get_column_ids(column_names, "allele", "forward", "reverse")

    # Get marker name column if it exists.
    try:
        colid_name = get_column_ids(column_names, "name")
    except:
        colid_name = None

    # Also try to get annotation column if we have one.
    if annotation_column is not None:
        try:
            colid_annotation = get_column_ids(column_names, annotation_column)
        except:
            annotation_column = None

    found_alleles = []
    for line in infile:
        line = line.rstrip("\r\n").split("\t")
        marker = line[colid_name] if colid_name is not None else default_marker
        allele = line[colid_allele] if seqformat is None \
jhoogenboom's avatar
jhoogenboom committed
917
918
            else ensure_sequence_format(line[colid_allele], seqformat,
                                        library=library, marker=marker)
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
        if (annotation_column is not None and
                line[colid_annotation].startswith("ALLELE")):
            found_alleles.append(marker, allele)
        data[marker, allele] = map(int,
            (line[colid_forward], line[colid_reverse]))

    return found_alleles
#read_sample_data_file


def reduce_read_counts(data, limit_reads):
    sum_reads = 0
    for markerallele in data:
        sum_reads += sum(data[markerallele])
    if sum_reads <= limit_reads:
        return

    remove = sorted(random.sample(xrange(sum_reads), sum_reads - limit_reads))
    i = 0
    seen = 0
    while i < len(remove) and seen > remove[i]:
        # Skip the reads filtered out above.
        i += 1
    for markerallele in data:
        for direction in (0, 1):
            seen += data[markerallele][direction]
            while i < len(remove) and seen > remove[i]:
                data[markerallele][direction] -= 1
                i += 1
#reduce_read_counts


def get_sample_data(tags_to_files, callback, allelelist=None,
                    annotation_column=None, seqformat=None, library=None,
                    marker=None, homozygotes=False, limit_reads=sys.maxint,
                    drop_samples=0):
    if drop_samples:
        sample_tags = tags_to_files.keys()
        for tag in random.sample(xrange(len(sample_tags)),
                                 int(len(sample_tags) * drop_samples)):
            del tags_to_files[sample_tags[tag]]

    for tag in tags_to_files:
        data = {}
        alleles = set()
        for infile in tags_to_files[tag]:
            alleles.update(read_sample_data_file(infile, data,
                annotation_column, seqformat, library, marker))
        if limit_reads < sys.maxint:
            reduce_read_counts(data, limit_reads)
        if allelelist is not None:
            if tag not in allelelist:
                allelelist[tag] = {}
            for markerx, allele in alleles:
                if markerx not in allelelist[tag]:
                    allelelist[tag][markerx] = set()
                allelelist[tag][markerx].add(allele)
            if marker:
                if marker in allelelist[tag]:
                    allelelist[tag] = {marker: allelelist[tag][marker]}
                else:
                    allelelist[tag] = {}
            if homozygotes:
jhoogenboom's avatar
jhoogenboom committed
982
983
984
                for markerx in allelelist[tag].keys():
                    if len(allelelist[tag][markerx]) > 1:
                        del allelelist[tag][markerx]
985
986
987
988
        callback(tag, data)
#get_sample_data


jhoogenboom's avatar
jhoogenboom committed
989
990
991
992
993
994
995
def get_column_ids(column_names, *names):
    """Find all names in column_names and return their indices."""
    result = []
    for name in names:
        try:
            result.append(column_names.index(name))
        except ValueError:
jhoogenboom's avatar
jhoogenboom committed
996
            raise ValueError("Column not found in input file: %s" % name)
jhoogenboom's avatar
jhoogenboom committed
997
998
999
1000
1001
1002
    if len(result) == 1:
        return result[0]
    return tuple(result)
#get_column_ids


jhoogenboom's avatar
jhoogenboom committed
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
def parse_allelelist(allelelist, convert=None, library=None):
    """Read allele list from open file handle."""
    column_names = allelelist.readline().rstrip("\r\n").split("\t")
    colid_sample, colid_marker, colid_allele = get_column_ids(column_names,
        "sample", "marker", "allele")
    alleles = {}
    for line in allelelist:
        line = line.rstrip("\r\n").split("\t")
        sample = line[colid_sample]
        marker = line[colid_marker]
        allele = line[colid_allele]
        if convert is not None:
            allele = ensure_sequence_format(allele, convert, library=library,
                                            marker=marker)
        if sample not in alleles:
            alleles[sample] = {}
        if marker not in alleles[sample]:
            alleles[sample][marker] = set()
        alleles[sample][marker].add(allele)
    return alleles
#parse_allelelist


jhoogenboom's avatar
jhoogenboom committed
1026
1027
1028
1029
1030
1031
1032
1033
1034
def pos_int_arg(value):
    """Convert str to int, raise ArgumentTypeError if not positive."""
    if not value.isdigit() or not int(value):
        raise argparse.ArgumentTypeError(
            "invalid positive int value: '%s'" % value)
    return int(value)
#pos_int_arg


jhoogenboom's avatar
jhoogenboom committed
1035
def add_allele_detection_args(parser):
jhoogenboom's avatar
jhoogenboom committed
1036
1037
    group = parser.add_argument_group("allele detection options")
    group.add_argument('-a', '--allelelist', metavar="ALLELEFILE",
jhoogenboom's avatar
jhoogenboom committed
1038
        type=argparse.FileType('r'),
jhoogenboom's avatar
jhoogenboom committed
1039
1040
1041
        help="file containing a list of the true alleles of each sample "
             "(e.g., obtained from allelefinder)")
    group.add_argument('-c', '--annotation-column', metavar="COLNAME",
jhoogenboom's avatar
jhoogenboom committed
1042
1043
1044
1045
1046
        help="name of a column in the sample files, which contains a value "
             "beginning with 'ALLELE' for the true alleles of the sample")
#add_allele_detection_args


1047
1048
def add_random_subsampling_args(parser):
    group = parser.add_argument_group("random subsampling options (advanced)")
jhoogenboom's avatar
jhoogenboom committed
1049
    group.add_argument('-Q', '--limit-reads', metavar="N", type=pos_int_arg,
1050
1051
1052
1053
1054
1055
        default=sys.maxint,
        help="simulate lower sequencing depth by randomly dropping reads down "
             "to this maximum total number of reads for each sample")
    group.add_argument('-x', '--drop-samples', metavar="N", type=float,
        default=0, help="randomly drop this fraction of input samples")
#add_random_subsampling_args
jhoogenboom's avatar
jhoogenboom committed
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075


def add_sequence_format_args(parser, default_format=None, force=False):
    group = parser.add_argument_group("sequence format options")
    if force:
        group.set_defaults(sequence_format=default_format)
    else:
        group.add_argument('-F', '--sequence-format', metavar="FORMAT",
            choices=("raw", "tssv", "allelename"),
            default=default_format,
            help="convert sequences to the specified format: one of "
                 "%(choices)s (default: " + (
                 "no conversion" if default_format is None else default_format)
                 + ")")
    group.add_argument('-l', '--library', metavar="LIBRARY",
        type=argparse.FileType('r'),
        help="library file for sequence format conversion")
#add_sequence_format_args


1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
def add_input_output_args(parser, single_in=False, batch_support=False,
                          report_out=False):
    """Add arguments for opening sample files to the given parser."""
    # Input file options group.
    if not single_in:
        parser.add_argument('infiles', nargs='*', metavar="FILE",
            default=[sys.stdin], type=argparse.FileType('r'),
            help="the sample data file(s) to process (default: read from "
                 "stdin)")
    elif not batch_support:
        parser.add_argument('infile', nargs='?', metavar="IN",
            default=sys.stdin, type=argparse.FileType('r'),
            help="the sample data file to process (default: read from stdin)")
    else:
        mutex = parser.add_argument_group(
                    "input file options").add_mutually_exclusive_group()
        mutex.add_argument('infile', nargs='?', metavar="IN",
            default=sys.stdin, type=argparse.FileType('r'),
            help="single sample data file to process (default: read from "
                 "stdin)")
        mutex.add_argument("-i", "--input", dest="infiles", nargs="+",
            metavar="IN", type=argparse.FileType('r'),
            help="multiple sample data files to process (use with "
                 "-o/--output)")

    # Output file options group.
    group = parser.add_argument_group("output file options")
    if batch_support and single_in:
        mutex = group.add_mutually_exclusive_group()
        mutex.add_argument('outfile', nargs='?', metavar="OUT",
            default=sys.stdout,
            help="the file to write the output to (default: write to stdout)")
        mutex.add_argument('-o', '--output', dest="outfiles", nargs="+",
            metavar="OUT",
            help="list of names of output files to match with input files "
                 "specified with -i/--input, or a format string to construct "
                 "file names from sample tags; e.g., the default value is "
                 "'\\1-%s.out', which expands to 'sampletag-%s.out'" %
                    ((parser.prog.rsplit(" ", 1)[-1],)*2))
    elif batch_support:
        group.add_argument('-o', '--output', dest="outfiles", nargs="+",
            metavar="OUT",
            default=[sys.stdout],
            help="a single file name to write all output to (default: write "
                 "to stdout) OR a list of names of output files to match with "
                 "input files OR a format string to construct file names from "
                 "sample tags; e.g., the value '\\1-%s.out' expands to "
                 "'sampletag-%s.out'" % ((parser.prog.rsplit(" ", 1)[-1],)*2))
    else:
        group.add_argument('-o', '--output', dest="outfile", metavar="FILE",
            type=argparse.FileType('w'),
            default=sys.stdout,
            help="file to write output to (default: write to stdout)")
    if report_out:
jhoogenboom's avatar
jhoogenboom committed
1130
        group.add_argument('-R', '--report', metavar="FILE",
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
            type=argparse.FileType('w'),
            default=sys.stderr,
            help="file to write a report to (default: write to stderr)")

    # Sample tag parsing options group.
    group = parser.add_argument_group("sample tag parsing options")
    group.add_argument('-e', '--tag-expr', metavar="REGEX", type=re.compile,
        default=DEF_TAG_EXPR,
        help="regular expression that captures (using one or more capturing "
             "groups) the sample tags from the file names; by default, the "
             "entire file name except for its extension (if any) is captured")
    group.add_argument('-f', '--tag-format', metavar="EXPR",
        default=DEF_TAG_FORMAT,
        help="format of the sample tags produced; a capturing group reference "
             "like '\\n' refers to the n-th capturing group in the regular "
             "expression specified with -e/--tag-expr (the default of '\\1' "
             "simply uses the first capturing group); with a single sample, "
             "you can enter the sample tag here explicitly")
#add_input_output_args
jhoogenboom's avatar
jhoogenboom committed
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160


def get_tag(filename, tag_expr, tag_format):
    """Return formatted sample tag from filename using regex."""
    try:
        return tag_expr.search(filename).expand(tag_format)
    except:
        return filename
#get_tag


1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
def get_input_output_files(args, single=False, batch_support=False):
    if single and not batch_support:
        # One infile, one outfile.  Return 2-tuple (infile, outfile).
        if args.infile.isatty():
            return False  # No input specified.
        return args.infile, args.outfile


    if not single and not batch_support:
        # N infiles, one outfile.  Return 2-tuple ({tag: infiles}, out).
        infiles = args.infiles if "infiles" in args \
                  and args.infiles is not None else [args.infile]
        if len(infiles) == 1 and infiles[0].isatty():
            return False  # No input specified.

        tags_to_files = {}
        for infile in infiles:
            tag = get_tag(infile.name, args.tag_expr, args.tag_format)
            try:
                tags_to_files[tag].append(infile)
            except KeyError:
                tags_to_files[tag] = [infile]
        return tags_to_files, args.outfile


    if single and batch_support:
        # N infiles, N outfiles.  Return generator of (tag, [ins], out).
        # Each yielded tuple should cause a separate run of the tool.
        infiles = args.infiles if "infiles" in args \
                  and args.infiles is not None else [args.infile]
        if len(infiles) == 1 and infiles[0].isatty():
            return False  # No input specified.

        outfiles = args.outfiles if "outfiles" in args \
                   and args.outfiles is not None else [args.outfile]
        if len(outfiles) > 1 and len(outfiles) != len(infiles):
            raise ValueError(
                "Number of input files (%i) is not equal to number of output "
                "files (%i)." % (len(infiles), len(outfiles)))

        tags = [get_tag(infile.name, args.tag_expr, args.tag_format)
                for infile in infiles]

        if len(outfiles) == 1:
            outfile = outfiles[0]
            if outfile == sys.stdout and len(set(tags)) == 1:
                # Write output of single sample to stdout.
                return ((tag, infiles, outfile) for tag in set(tags))

            # Write output of each sample to its own outfile.
            if outfile == sys.stdout:
                outfile = DEF_OUTFILE_FORMAT
            return ((tag,
                    [infiles[i] for i in range(len(tags)) if tags[i]==tag],
                    open(outfile.replace("\\1", tag).replace("\\2",
                         args.tool), "w")) for tag in set(tags))

        # Link each output file to each input file.
        # Treating files with the same sample tag as separate samples.
        return ((tags[i], [infiles[i]], open(outfiles[i], 'w'))
                for i in range(len(tags)))

    if not single and batch_support:
        # N infiles, one or N outfiles.
        # If one outfile, return ({tag: [infiles]}, outfile).
        # If N outfiles, return generator of (tag, [infiles], outfile).
        raise NotImplementedError("Multi-input with optional multi-output not supported yet.")
#get_input_output_files
jhoogenboom's avatar
jhoogenboom committed
1229
1230


jhoogenboom's avatar
jhoogenboom committed
1231
1232
1233
1234
1235
def print_db(text, debug):
    """Print text if debug is True."""
    if debug:
        print(text)
#print_db