Mercurial > repos > crs4 > sspace

comparison sspace.py @ 0:b1be94418db9 draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Uploaded
author crs4
date Tue, 17 Sep 2013 09:41:18 -0400
parents
children
comparison
equal deleted inserted replaced
-1:000000000000 0:b1be94418db9
1 # -*- coding: utf-8 -*-
2 """
3 SSPACE wrapper
4 """
5
6 import logging
7 import optparse
8 import os
9 import shutil
10 import subprocess
11 import tempfile
12
13
14 def which(name, flags=os.X_OK):
15 """ Search PATH for executable files with the given name. """
16 result = []
17 exts = filter(None, os.environ.get('PATHEXT', '').split(os.pathsep))
18 path = os.environ.get('PATH', None)
19 if path is None:
20 return []
21 for p in os.environ.get('PATH', '').split(os.pathsep):
22 p = os.path.join(p, str(name))
23 if os.access(p, flags):
24 result.append(p)
25 for e in exts:
26 pext = p + e
27 if os.access(pext, flags):
28 result.append(pext)
29 return result
30
31
32 LOG_FORMAT = '%(asctime)s|%(levelname)-8s|%(message)s'
33 LOG_DATEFMT = '%Y-%m-%d %H:%M:%S'
34 LOG_LEVELS = ['DEBUG', 'INFO', 'WARNING', 'ERROR', 'CRITICAL']
35
36
37 def __main__():
38 """ main function """
39 parser = optparse.OptionParser()
40 parser.add_option('-c', dest='contigs', help='contigs mandatory (-s)')
41 parser.add_option('--r1', dest='reads1', help='')
42 parser.add_option('--r2', dest='reads2', help='')
43 parser.add_option('-i', dest='insert', type='int', help='')
44 parser.add_option('-e', dest='error', type='float', help='')
45 parser.add_option('-o', dest='orientation', choices=['FF', 'FR', 'RF', 'RR'], help='')
46 parser.add_option('-x', action='store_true', dest='extension', help='whether to extend the contigs (-x)')
47 parser.add_option('--minoverlap', dest='minoverlap', type='int', help='minoverlap (-m)')
48 parser.add_option('--numofreads', dest='numofreads', type='int', help='num of reads to call an extension (-o)')
49 parser.add_option('-t', dest='max_trim', type='int', help='maximum number of bases to trim on the contig end when all possibilities have been exhausted for an extension (-t)')
50 parser.add_option('-u', dest='unpaired', help='unpaired reads (-u, optional)')
51 parser.add_option('-r', dest='min_base_ratio', type='float', help='minimum base ratio used to accept a overhang consensus base (-r)')
52 parser.add_option('--minlink', dest='minlink', type='int', help='min link (-k)')
53 parser.add_option('--maxratio', dest='maxratio', type='float', help='max ratio (-a)')
54 parser.add_option('--contigoverlap', dest='contigoverlap', type='int', help='contigoverlap (-n)')
55 parser.add_option('--mincontig', dest='mincontig', type='int', help='mincontig (-z)')
56 parser.add_option('-T', dest='n_threads', type='int', help='number of threads to use in Bowtie (-T)')
57 parser.add_option('-p', dest='prefix', default='sspace_pre', help='prefix (-b)')
58 parser.add_option('--lib', dest='libraryname', default='galx', help='libraryfile')
59 parser.add_option('--fe', dest='finalevidence', help='')
60 parser.add_option('--fs', dest='finalscaffolds', help='')
61 parser.add_option('--lg', dest='logfile', help='')
62 parser.add_option('--summ', dest='summaryfile', help='')
63 parser.add_option('--loglevel', choices=LOG_LEVELS, help='logging level (default: INFO)', default='DEBUG')
64 (options, args) = parser.parse_args()
65 if len(args) > 0:
66 parser.error('Wrong number of arguments')
67
68 log_level = getattr(logging, options.loglevel)
69 kwargs = {'format' : LOG_FORMAT,
70 'datefmt' : LOG_DATEFMT,
71 'level' : log_level}
72 if options.logfile:
73 kwargs['filename'] = options.logfile
74 logging.basicConfig(**kwargs)
75 logger = logging.getLogger('SSPACE scaffold assembly')
76
77 logger.debug('Creating temp dir')
78 contigs = options.contigs
79 reads1 = options.reads1
80 reads2 = options.reads2
81 insert = options.insert
82 error = options.error
83 orientation = options.orientation
84 extension = '-x 1' if options.extension else ''
85 minoverlap = "-m %d" % options.minoverlap if options.minoverlap is not None else ''
86 numofreads = "-o %d" % options.numofreads if options.numofreads is not None else ''
87 max_trim = "-t %d" % options.max_trim if options.max_trim is not None else ''
88 unpaired = "-u %s" % options.unpaired if options.unpaired else ''
89 min_base_ratio = "-r %s" % options.min_base_ratio if options.min_base_ratio is not None else ''
90 minlink = "-k %d" % options.minlink if options.minlink is not None else ''
91 maxratio = "-a %s" % options.maxratio if options.maxratio is not None else ''
92 contigoverlap = "-n %d" % options.contigoverlap if options.contigoverlap is not None else ''
93 mincontig = "-z %d" % options.mincontig if options.mincontig is not None else ''
94 n_threads = "-T %d" % options.n_threads if options.n_threads is not None else ''
95 prefix = options.prefix
96 libraryname = options.libraryname
97 finalevidence = options.finalevidence
98 finalscaffolds = options.finalscaffolds
99 summaryfile = options.summaryfile
100 exe_name = "SSPACE_Basic_v2.0.pl"
101
102 # Set library
103 outline = 'lib1 %s %s %d %s %s\n' % (reads1, reads2, insert, error, orientation)
104 with open(libraryname, 'w') as out:
105 out.write(outline)
106
107 exe = which(exe_name, os.R_OK)
108 command = "perl %s -l %s -s %s %s %s %s %s %s %s %s %s %s %s %s -b %s" % (exe.pop(), libraryname, contigs, extension, minoverlap, numofreads, max_trim, unpaired, min_base_ratio, minlink, maxratio, contigoverlap, mincontig, n_threads, prefix)
109 wd = tempfile.mkdtemp()
110 try:
111 os.chdir(wd)
112 logger.info("SSPACE running")
113 logger.debug("executing %s" % command)
114 subprocess.check_call(args=command, shell=True)
115
116 # Clean environment
117 with open("%s.logfile.txt" % os.path.join(wd, prefix), 'rb') as sspace_log_file:
118 logger.info("\n".join(["Log from SSPACE", sspace_log_file.read()]))
119 logger.info("Moving result files")
120 shutil.move("%s.final.evidence" % os.path.join(wd, prefix), finalevidence)
121 shutil.move("%s.final.scaffolds.fasta" % os.path.join(wd, prefix), finalscaffolds)
122 shutil.move("%s.summaryfile.txt" % os.path.join(wd, prefix), summaryfile)
123 logger.debug("Removing temp dir: %s" % wd)
124 finally:
125 shutil.rmtree(wd)
126
127
128 if __name__ == "__main__":
129 __main__()