Mercurial > repos > bgruening > bismark

annotate bismark_deduplicate/bismark_deduplicate_wrapper.py @ 7:fcadce4d9a06 draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit b'e6ee273f75fff61d1e419283fa8088528cf59470\n'
author bgruening
date Sat, 06 May 2017 13:18:09 -0400
parents
children
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
7
fcadce4d9a06 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit b'e6ee273f75fff61d1e419283fa8088528cf59470\n'
bgruening
parents:
diff changeset
1 #!/usr/bin/python
fcadce4d9a06 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit b'e6ee273f75fff61d1e419283fa8088528cf59470\n'
bgruening
parents:
diff changeset
2
fcadce4d9a06 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit b'e6ee273f75fff61d1e419283fa8088528cf59470\n'
bgruening
parents:
diff changeset
3 import argparse
fcadce4d9a06 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit b'e6ee273f75fff61d1e419283fa8088528cf59470\n'
bgruening
parents:
diff changeset
4 import os
fcadce4d9a06 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit b'e6ee273f75fff61d1e419283fa8088528cf59470\n'
bgruening
parents:
diff changeset
5 import re
fcadce4d9a06 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit b'e6ee273f75fff61d1e419283fa8088528cf59470\n'
bgruening
parents:
diff changeset
6 import shutil
fcadce4d9a06 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit b'e6ee273f75fff61d1e419283fa8088528cf59470\n'
bgruening
parents:
diff changeset
7 import subprocess
fcadce4d9a06 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit b'e6ee273f75fff61d1e419283fa8088528cf59470\n'
bgruening
parents:
diff changeset
8 import sys
fcadce4d9a06 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit b'e6ee273f75fff61d1e419283fa8088528cf59470\n'
bgruening
parents:
diff changeset
9 import tempfile
fcadce4d9a06 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit b'e6ee273f75fff61d1e419283fa8088528cf59470\n'
bgruening
parents:
diff changeset
10 import logging
fcadce4d9a06 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit b'e6ee273f75fff61d1e419283fa8088528cf59470\n'
bgruening
parents:
diff changeset
11 from glob import glob
fcadce4d9a06 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit b'e6ee273f75fff61d1e419283fa8088528cf59470\n'
bgruening
parents:
diff changeset
12
fcadce4d9a06 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit b'e6ee273f75fff61d1e419283fa8088528cf59470\n'
bgruening
parents:
diff changeset
13 def cleanup_before_exit(tmp_dir):
fcadce4d9a06 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit b'e6ee273f75fff61d1e419283fa8088528cf59470\n'
bgruening
parents:
diff changeset
14 if tmp_dir and os.path.exists(tmp_dir):
fcadce4d9a06 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit b'e6ee273f75fff61d1e419283fa8088528cf59470\n'
bgruening
parents:
diff changeset
15 shutil.rmtree(tmp_dir)
fcadce4d9a06 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit b'e6ee273f75fff61d1e419283fa8088528cf59470\n'
bgruening
parents:
diff changeset
16
fcadce4d9a06 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit b'e6ee273f75fff61d1e419283fa8088528cf59470\n'
bgruening
parents:
diff changeset
17 def get_arg():
fcadce4d9a06 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit b'e6ee273f75fff61d1e419283fa8088528cf59470\n'
bgruening
parents:
diff changeset
18 parser = argparse.ArgumentParser()
fcadce4d9a06 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit b'e6ee273f75fff61d1e419283fa8088528cf59470\n'
bgruening
parents:
diff changeset
19 parser.add_argument('--tool_dir', dest='tool_dir', action='store', nargs=1, metavar='tool_dir', type=str)
fcadce4d9a06 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit b'e6ee273f75fff61d1e419283fa8088528cf59470\n'
bgruening
parents:
diff changeset
20 parser.add_argument('-p', action='store_true')
fcadce4d9a06 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit b'e6ee273f75fff61d1e419283fa8088528cf59470\n'
bgruening
parents:
diff changeset
21 parser.add_argument('-s', action='store_true')
fcadce4d9a06 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit b'e6ee273f75fff61d1e419283fa8088528cf59470\n'
bgruening
parents:
diff changeset
22 parser.add_argument('--input', dest='input', action='store', nargs=1, metavar='input', type=str)
fcadce4d9a06 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit b'e6ee273f75fff61d1e419283fa8088528cf59470\n'
bgruening
parents:
diff changeset
23 parser.add_argument('--output_report', dest='output_report', action='store', nargs=1, metavar='output_report', type=str)
fcadce4d9a06 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit b'e6ee273f75fff61d1e419283fa8088528cf59470\n'
bgruening
parents:
diff changeset
24 parser.add_argument('--output_bam', dest='output_bam', action='store', nargs=1, metavar='output_report', type=str)
fcadce4d9a06 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit b'e6ee273f75fff61d1e419283fa8088528cf59470\n'
bgruening
parents:
diff changeset
25 parser.add_argument('--log_report', dest='log_report', action='store', nargs=1, metavar='log_filename', type=str)
fcadce4d9a06 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit b'e6ee273f75fff61d1e419283fa8088528cf59470\n'
bgruening
parents:
diff changeset
26 args = parser.parse_args()
fcadce4d9a06 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit b'e6ee273f75fff61d1e419283fa8088528cf59470\n'
bgruening
parents:
diff changeset
27 return args
fcadce4d9a06 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit b'e6ee273f75fff61d1e419283fa8088528cf59470\n'
bgruening
parents:
diff changeset
28
fcadce4d9a06 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit b'e6ee273f75fff61d1e419283fa8088528cf59470\n'
bgruening
parents:
diff changeset
29 def __main__():
fcadce4d9a06 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit b'e6ee273f75fff61d1e419283fa8088528cf59470\n'
bgruening
parents:
diff changeset
30 args = get_arg()
fcadce4d9a06 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit b'e6ee273f75fff61d1e419283fa8088528cf59470\n'
bgruening
parents:
diff changeset
31
fcadce4d9a06 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit b'e6ee273f75fff61d1e419283fa8088528cf59470\n'
bgruening
parents:
diff changeset
32 tmp_dir = tempfile.mkdtemp(prefix='tmp', suffix='')
fcadce4d9a06 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit b'e6ee273f75fff61d1e419283fa8088528cf59470\n'
bgruening
parents:
diff changeset
33 os.chdir(tmp_dir)
fcadce4d9a06 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit b'e6ee273f75fff61d1e419283fa8088528cf59470\n'
bgruening
parents:
diff changeset
34
fcadce4d9a06 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit b'e6ee273f75fff61d1e419283fa8088528cf59470\n'
bgruening
parents:
diff changeset
35 if args.log_report:
fcadce4d9a06 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit b'e6ee273f75fff61d1e419283fa8088528cf59470\n'
bgruening
parents:
diff changeset
36 logging.basicConfig(level=logging.INFO, filename=args.log_report[0], filemode="a+", format='%(message)s')
fcadce4d9a06 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit b'e6ee273f75fff61d1e419283fa8088528cf59470\n'
bgruening
parents:
diff changeset
37 else:
fcadce4d9a06 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit b'e6ee273f75fff61d1e419283fa8088528cf59470\n'
bgruening
parents:
diff changeset
38 logging.basicConfig(level=logging.INFO, filename=os.path.join(tmp_dir, 'log_report.txt'), filemode="a+", format='%(message)s')
fcadce4d9a06 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit b'e6ee273f75fff61d1e419283fa8088528cf59470\n'
bgruening
parents:
diff changeset
39
fcadce4d9a06 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit b'e6ee273f75fff61d1e419283fa8088528cf59470\n'
bgruening
parents:
diff changeset
40 default_reads_name = 'submitted_reads.bam'
fcadce4d9a06 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit b'e6ee273f75fff61d1e419283fa8088528cf59470\n'
bgruening
parents:
diff changeset
41 os.symlink(args.input[0], default_reads_name)
fcadce4d9a06 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit b'e6ee273f75fff61d1e419283fa8088528cf59470\n'
bgruening
parents:
diff changeset
42
fcadce4d9a06 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit b'e6ee273f75fff61d1e419283fa8088528cf59470\n'
bgruening
parents:
diff changeset
43 if args.p is True:
fcadce4d9a06 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit b'e6ee273f75fff61d1e419283fa8088528cf59470\n'
bgruening
parents:
diff changeset
44 sPaired = '-p'
fcadce4d9a06 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit b'e6ee273f75fff61d1e419283fa8088528cf59470\n'
bgruening
parents:
diff changeset
45 if args.s is True:
fcadce4d9a06 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit b'e6ee273f75fff61d1e419283fa8088528cf59470\n'
bgruening
parents:
diff changeset
46 sPaired = '-s'
fcadce4d9a06 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit b'e6ee273f75fff61d1e419283fa8088528cf59470\n'
bgruening
parents:
diff changeset
47
fcadce4d9a06 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit b'e6ee273f75fff61d1e419283fa8088528cf59470\n'
bgruening
parents:
diff changeset
48 cmd = 'perl %s %s duplicated_reads.bam --bam' % (os.path.join(args.tool_dir[0], 'deduplicate_bismark'), sPaired)
fcadce4d9a06 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit b'e6ee273f75fff61d1e419283fa8088528cf59470\n'
bgruening
parents:
diff changeset
49 logging.info('COMMAND LINE:\n\n%s' % cmd)
fcadce4d9a06 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit b'e6ee273f75fff61d1e419283fa8088528cf59470\n'
bgruening
parents:
diff changeset
50
fcadce4d9a06 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit b'e6ee273f75fff61d1e419283fa8088528cf59470\n'
bgruening
parents:
diff changeset
51 proc = subprocess.Popen(['perl', os.path.join(args.tool_dir[0], 'deduplicate_bismark'), sPaired, default_reads_name, '--bam'], stdout=subprocess.PIPE, stderr=subprocess.STDOUT)
fcadce4d9a06 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit b'e6ee273f75fff61d1e419283fa8088528cf59470\n'
bgruening
parents:
diff changeset
52 proc_out, proc_err = proc.communicate()
fcadce4d9a06 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit b'e6ee273f75fff61d1e419283fa8088528cf59470\n'
bgruening
parents:
diff changeset
53
fcadce4d9a06 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit b'e6ee273f75fff61d1e419283fa8088528cf59470\n'
bgruening
parents:
diff changeset
54 logging.info("__________________________________________________________________\n")
fcadce4d9a06 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit b'e6ee273f75fff61d1e419283fa8088528cf59470\n'
bgruening
parents:
diff changeset
55 logging.info("BISMARK DEDUPLICATE STDOUT:\n\n%s" % proc_out)
fcadce4d9a06 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit b'e6ee273f75fff61d1e419283fa8088528cf59470\n'
bgruening
parents:
diff changeset
56 if proc_err:
fcadce4d9a06 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit b'e6ee273f75fff61d1e419283fa8088528cf59470\n'
bgruening
parents:
diff changeset
57 logging.critical("__________________________________________________________________\n")
fcadce4d9a06 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit b'e6ee273f75fff61d1e419283fa8088528cf59470\n'
bgruening
parents:
diff changeset
58 logging.critical("BISMARK DEDUPLICATE WARNING:\n\n%s" % proc_err)
fcadce4d9a06 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit b'e6ee273f75fff61d1e419283fa8088528cf59470\n'
bgruening
parents:
diff changeset
59 sys.exit("Dedpulicate Bismark crashed with the folowing error message:\n%s" % proc_err)
fcadce4d9a06 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit b'e6ee273f75fff61d1e419283fa8088528cf59470\n'
bgruening
parents:
diff changeset
60
fcadce4d9a06 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit b'e6ee273f75fff61d1e419283fa8088528cf59470\n'
bgruening
parents:
diff changeset
61 shutil.move( glob('*deduplicated.bam')[0], args.output_bam[0] )
fcadce4d9a06 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit b'e6ee273f75fff61d1e419283fa8088528cf59470\n'
bgruening
parents:
diff changeset
62 shutil.move( glob('*deduplication_report.txt')[0], args.output_report[0])
fcadce4d9a06 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit b'e6ee273f75fff61d1e419283fa8088528cf59470\n'
bgruening
parents:
diff changeset
63
fcadce4d9a06 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit b'e6ee273f75fff61d1e419283fa8088528cf59470\n'
bgruening
parents:
diff changeset
64 cleanup_before_exit(tmp_dir)
fcadce4d9a06 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit b'e6ee273f75fff61d1e419283fa8088528cf59470\n'
bgruening
parents:
diff changeset
65
fcadce4d9a06 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit b'e6ee273f75fff61d1e419283fa8088528cf59470\n'
bgruening
parents:
diff changeset
66 if __name__=="__main__": __main__()