Mercurial > repos > urgi-team > mdust

comparison mdust_wrapper.py @ 0:cc6db1ee2d48 draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Uploaded
author urgi-team
date Mon, 05 Oct 2015 11:05:21 -0400
parents
children
comparison
equal deleted inserted replaced
-1:000000000000 0:cc6db1ee2d48
1 #!/usr/bin/env python
2
3 import subprocess
4 import tempfile
5 import sys
6 import os
7 import re
8 from optparse import OptionParser
9
10
11 class MdustWrapper(object):
12
13 def __init__(self):
14 self._options = None
15
16 def stop_err(self, msg):
17 sys.stderr.write("%s\n" % msg)
18 sys.exit()
19
20 def setAttributesFromCmdLine(self):
21 description = "mdust_wrapper"
22 description += "\nWrapper for mdust\n"
23 description += "example: mdust_wrapper.py -i seq.fasta -v 27\n"
24 parser = OptionParser(description = description, version = "0.1")
25 parser.add_option("-i", "--input", dest = "FastaFile", action = "store", type = "string", help = "Input Fasta File name [compulsory] [format: Fasta]", default = "")
26 parser.add_option("-o", "--output", dest = "outFile", action = "store", type = "string", help = "output File name [compulsory] [format: fasta,tab or bed]", default = "")
27 parser.add_option("-v", "--cutoff", dest = "cutoff", action = "store", type = "int", help = "cutoff", default = 28)
28 parser.add_option("-w", "--wsize", dest = "wsize", action = "store", type = "int", help = "window size", default = 3)
29 parser.add_option("-m", "--maskingletter", dest = "maskingletter", action = "store", type = "string", help = "masking letter", default = "N")
30 parser.add_option("-f", "--format", dest = "format", action = "store", type = "string", help = "format", default = "default")
31 options = parser.parse_args()[0]
32 self._setAttributesFromOptions(options)
33
34
35 def _setAttributesFromOptions(self, options):
36 self._options = options
37
38 if self._options.FastaFile == "":
39 raise Exception("Missing input file, please provide fasta file with -i file !")
40 if self._options.outFile == "":
41 raise Exception("Missing output file, please provide output file with -o file !")
42
43
44 def run(self):
45
46 prg = "mdust"
47 args = ""
48 args += " %s" % self._options.FastaFile
49 args += " -v %d" % self._options.cutoff
50 args += " -w %d" % self._options.wsize
51 args += " -m %s" % self._options.maskingletter
52 if self._options.format == "tab" or self._options.format == "bed":
53 args += " -c "
54 cmd = "%s %s" %(prg, args)
55
56 try:
57 tmp_err = tempfile.NamedTemporaryFile().name
58 tmp_out = "outfile"
59 tmp_stderr = open( tmp_err, 'wb' )
60 tmp_stdout = open( tmp_out, 'wb' )
61 proc = subprocess.Popen( args=cmd, shell=True, cwd=".", stdout=tmp_stdout, stderr=tmp_stderr )
62 returncode = proc.wait()
63 tmp_stderr.close()
64 # get stderr, allowing for case where it's very large
65 tmp_stderr = open( tmp_err, 'rb' )
66 tmp_stdout = open( tmp_out, 'rb' )
67
68 stderr = ''
69 stdout = ''
70 buffsize = 1048576
71 try:
72 while True:
73 stdout += tmp_stdout.read( buffsize )
74 if not stdout or len( stdout ) % buffsize != 0:
75 break
76 except OverflowError:
77 pass
78 tmp_stdout.close()
79
80 try:
81 while True:
82 stderr += tmp_stderr.read( buffsize )
83 if not stderr or len( stderr ) % buffsize != 0:
84 break
85 except OverflowError:
86 pass
87 tmp_stderr.close()
88 if stderr:
89 raise Exception, stderr
90 except Exception, e:
91 self.stop_err( 'Error with mdust :\n' + str( e ) )
92
93 if self._options.format == 'bed':
94 with open(tmp_out,"r") as fin:
95 with open(self._options.outFile, "w") as fout:
96 lineNumber = 0
97 for line in fin:
98 lineNumber += 1
99 m = re.search(r"^(\S+)\t(\d+)\t(\d+)\t(\d+)$", line)
100 if m is not None:
101 fout.write("%s\t%d\t%d\n" % (m.group(1), int(m.group(3))-1, int(m.group(4))))
102 if m is None:
103 raise Exception("\nLine %d '%s' does not has a mdust format." % (lineNumber, line))
104 else:
105 os.rename(tmp_out,self._options.outFile)
106
107
108
109
110 if __name__ == "__main__":
111 iWrapper = MdustWrapper()
112 iWrapper.setAttributesFromCmdLine()
113 iWrapper.run()