Mercurial > repos > sanbi-uwc > gatk2

comparison gatk2_wrapper.py @ 0:6e985e2e0802 draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
planemo upload for repository https://github.com/SANBI-SA/tools-sanbi-uwc/tree/master/tools/gatk2 commit f53888603dbab77d16c36cceffe0f2060052d204
author sanbi-uwc
date Wed, 25 Jan 2017 08:00:49 -0500
parents
children
comparison
equal deleted inserted replaced
-1:000000000000 0:6e985e2e0802
1 #!/usr/bin/env python
2 #David Hoover, based on gatk by Dan Blankenberg
3
4 """
5 A wrapper script for running the GenomeAnalysisTK.jar commands.
6 """
7
8 import sys, optparse, os, tempfile, subprocess, shutil
9 from binascii import unhexlify
10
11 GALAXY_EXT_TO_GATK_EXT = { 'gatk_interval':'intervals', 'bam_index':'bam.bai', 'gatk_dbsnp':'dbSNP', 'picard_interval_list':'interval_list' } #items not listed here will use the galaxy extension as-is
12 GALAXY_EXT_TO_GATK_FILE_TYPE = GALAXY_EXT_TO_GATK_EXT #for now, these are the same, but could be different if needed
13 DEFAULT_GATK_PREFIX = "gatk_file"
14 CHUNK_SIZE = 2**20 #1mb
15
16
17 def cleanup_before_exit( tmp_dir ):
18 if tmp_dir and os.path.exists( tmp_dir ):
19 shutil.rmtree( tmp_dir )
20
21
22 def gatk_filename_from_galaxy( galaxy_filename, galaxy_ext, target_dir = None, prefix = None ):
23 suffix = GALAXY_EXT_TO_GATK_EXT.get( galaxy_ext, galaxy_ext )
24 if prefix is None:
25 prefix = DEFAULT_GATK_PREFIX
26 if target_dir is None:
27 target_dir = os.getcwd()
28 gatk_filename = os.path.join( target_dir, "%s.%s" % ( prefix, suffix ) )
29 os.symlink( galaxy_filename, gatk_filename )
30 return gatk_filename
31
32
33 def gatk_filetype_argument_substitution( argument, galaxy_ext ):
34 return argument % dict( file_type = GALAXY_EXT_TO_GATK_FILE_TYPE.get( galaxy_ext, galaxy_ext ) )
35
36
37 def open_file_from_option( filename, mode = 'rb' ):
38 if filename:
39 return open( filename, mode = mode )
40 return None
41
42
43 def html_report_from_directory( html_out, dir ):
44 html_out.write( '<html>\n<head>\n<title>Galaxy - GATK Output</title>\n</head>\n<body>\n<p/>\n<ul>\n' )
45 for fname in sorted( os.listdir( dir ) ):
46 html_out.write( '<li><a href="%s">%s</a></li>\n' % ( fname, fname ) )
47 html_out.write( '</ul>\n</body>\n</html>\n' )
48
49
50 def index_bam_files( bam_filenames ):
51 for bam_filename in bam_filenames:
52 bam_index_filename = "%s.bai" % bam_filename
53 if not os.path.exists( bam_index_filename ):
54 #need to index this bam file
55 stderr_name = tempfile.NamedTemporaryFile( prefix = "bam_index_stderr" ).name
56 command = 'samtools index %s %s' % ( bam_filename, bam_index_filename )
57 try:
58 subprocess.check_call( args=command, shell=True, stderr=open( stderr_name, 'wb' ) )
59 except:
60 for line in open( stderr_name ):
61 print >> sys.stderr, line
62 raise Exception( "Error indexing BAM file" )
63 finally:
64 os.unlink( stderr_name )
65
66 def __main__():
67 #Parse Command Line
68 parser = optparse.OptionParser()
69 parser.add_option( '-p', '--pass_through', dest='pass_through_options', action='append', type="string", help='These options are passed through directly to GATK, without any modification.' )
70 parser.add_option( '-o', '--pass_through_options', dest='pass_through_options_encoded', action='append', type="string", help='These options are passed through directly to GATK, with decoding from binascii.unhexlify.' )
71 parser.add_option( '-d', '--dataset', dest='datasets', action='append', type="string", nargs=4, help='"-argument" "original_filename" "galaxy_filetype" "name_prefix"' )
72 parser.add_option( '', '--max_jvm_heap', dest='max_jvm_heap', action='store', type="string", default=None, help='If specified, the maximum java virtual machine heap size will be set to the provide value.' )
73 parser.add_option( '', '--max_jvm_heap_fraction', dest='max_jvm_heap_fraction', action='store', type="int", default=None, help='If specified, the maximum java virtual machine heap size will be set to the provide value as a fraction of total physical memory.' )
74 parser.add_option( '', '--stdout', dest='stdout', action='store', type="string", default=None, help='If specified, the output of stdout will be written to this file.' )
75 parser.add_option( '', '--stderr', dest='stderr', action='store', type="string", default=None, help='If specified, the output of stderr will be written to this file.' )
76 parser.add_option( '', '--html_report_from_directory', dest='html_report_from_directory', action='append', type="string", nargs=2, help='"Target HTML File" "Directory"')
77 parser.add_option( '-e', '--phone_home', dest='phone_home', action='store', type="string", default='STANDARD', help='What kind of GATK run report should we generate(NO_ET|STANDARD|STDOUT)' )
78 parser.add_option( '-K', '--gatk_key', dest='gatk_key', action='store', type="string", default=None, help='What kind of GATK run report should we generate(NO_ET|STANDARD|STDOUT)' )
79 (options, args) = parser.parse_args()
80
81 if options.pass_through_options:
82 cmd = ' '.join( options.pass_through_options )
83 else:
84 cmd = ''
85 if options.pass_through_options_encoded:
86 cmd = '%s %s' % ( cmd, ' '.join( map( unhexlify, options.pass_through_options_encoded ) ) )
87 if options.max_jvm_heap is not None:
88 cmd = cmd.replace( 'java ', 'java -Xmx%s ' % ( options.max_jvm_heap ), 1 )
89 elif options.max_jvm_heap_fraction is not None:
90 cmd = cmd.replace( 'java ', 'java -XX:DefaultMaxRAMFraction=%s -XX:+UseParallelGC ' % ( options.max_jvm_heap_fraction ), 1 )
91 bam_filenames = []
92 tmp_dir = tempfile.mkdtemp( prefix='tmp-gatk-' )
93 try:
94 if options.datasets:
95 for ( dataset_arg, filename, galaxy_ext, prefix ) in options.datasets:
96 gatk_filename = gatk_filename_from_galaxy( filename, galaxy_ext, target_dir = tmp_dir, prefix = prefix )
97 if dataset_arg:
98 cmd = '%s %s "%s"' % ( cmd, gatk_filetype_argument_substitution( dataset_arg, galaxy_ext ), gatk_filename )
99 if galaxy_ext == "bam":
100 bam_filenames.append( gatk_filename )
101 if galaxy_ext == 'fasta':
102 subprocess.check_call( 'samtools faidx "%s"' % gatk_filename, shell=True )
103 subprocess.check_call( 'java -jar %s R=%s O=%s QUIET=true' % ( os.path.join(os.environ['JAVA_JAR_PATH'], 'CreateSequenceDictionary.jar'), gatk_filename, os.path.splitext(gatk_filename)[0] + '.dict' ), shell=True )
104 index_bam_files( bam_filenames )
105 #set up stdout and stderr output options
106 stdout = open_file_from_option( options.stdout, mode = 'wb' )
107 stderr = open_file_from_option( options.stderr, mode = 'wb' )
108 #if no stderr file is specified, we'll use our own
109 if stderr is None:
110 stderr = tempfile.NamedTemporaryFile( prefix="gatk-stderr-", dir=tmp_dir )
111
112 proc = subprocess.Popen( args=cmd, stdout=stdout, stderr=stderr, shell=True, cwd=tmp_dir )
113 return_code = proc.wait()
114
115 if return_code:
116 stderr_target = sys.stderr
117 else:
118 stderr_target = sys.stdout
119 stderr.flush()
120 stderr.seek(0)
121 while True:
122 chunk = stderr.read( CHUNK_SIZE )
123 if chunk:
124 stderr_target.write( chunk )
125 else:
126 break
127 stderr.close()
128 finally:
129 cleanup_before_exit( tmp_dir )
130
131 #generate html reports
132 if options.html_report_from_directory:
133 for ( html_filename, html_dir ) in options.html_report_from_directory:
134 html_report_from_directory( open( html_filename, 'wb' ), html_dir )
135
136
137 if __name__ == "__main__":
138 __main__()