Mercurial > repos > lz_hust > gatktools

comparison gatk2_wrapper.py @ 15:01ff8dd37d4d draft default tip

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Uploaded
author lz_hust
date Sat, 01 Jun 2019 07:20:41 -0400
parents
children
comparison
equal deleted inserted replaced
14:68426930d59c 15:01ff8dd37d4d
1 #!/usr/bin/env python
2 # David Hoover, based on gatk by Dan Blankenberg
3 """
4 A wrapper script for running the GenomeAnalysisTK.jar commands.
5 """
6
7 import optparse
8 import os
9 import shutil
10 import subprocess
11 import sys
12 import tempfile
13 from binascii import unhexlify
14
15 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
16 GALAXY_EXT_TO_GATK_FILE_TYPE = GALAXY_EXT_TO_GATK_EXT # for now, these are the same, but could be different if needed
17 DEFAULT_GATK_PREFIX = "gatk_file"
18 CHUNK_SIZE = 2**20 # 1mb
19
20
21 def cleanup_before_exit( tmp_dir ):
22 if tmp_dir and os.path.exists( tmp_dir ):
23 shutil.rmtree( tmp_dir )
24
25
26 def gatk_filename_from_galaxy( galaxy_filename, galaxy_ext, target_dir=None, prefix=None ):
27 suffix = GALAXY_EXT_TO_GATK_EXT.get( galaxy_ext, galaxy_ext )
28 if prefix is None:
29 prefix = DEFAULT_GATK_PREFIX
30 if target_dir is None:
31 target_dir = os.getcwd()
32 gatk_filename = os.path.join( target_dir, "%s.%s" % ( prefix, suffix ) )
33 os.symlink( galaxy_filename, gatk_filename )
34 return gatk_filename
35
36
37 def gatk_filetype_argument_substitution( argument, galaxy_ext ):
38 return argument % dict( file_type=GALAXY_EXT_TO_GATK_FILE_TYPE.get( galaxy_ext, galaxy_ext ) )
39
40
41 def open_file_from_option( filename, mode='rb' ):
42 if filename:
43 return open( filename, mode=mode )
44 return None
45
46
47 def html_report_from_directory( html_out, dir ):
48 html_out.write( '<html>\n<head>\n<title>Galaxy - GATK Output</title>\n</head>\n<body>\n<p/>\n<ul>\n' )
49 for fname in sorted( os.listdir( dir ) ):
50 html_out.write( '<li><a href="%s">%s</a></li>\n' % ( fname, fname ) )
51 html_out.write( '</ul>\n</body>\n</html>\n' )
52
53
54 def index_bam_files( bam_filenames ):
55 for bam_filename in bam_filenames:
56 bam_index_filename = "%s.bai" % bam_filename
57 if not os.path.exists( bam_index_filename ):
58 # need to index this bam file
59 stderr_name = tempfile.NamedTemporaryFile( prefix="bam_index_stderr" ).name
60 command = 'samtools index %s %s' % ( bam_filename, bam_index_filename )
61 try:
62 subprocess.check_call( args=command, shell=True, stderr=open( stderr_name, 'wb' ) )
63 except:
64 for line in open( stderr_name ):
65 print >> sys.stderr, line
66 raise Exception( "Error indexing BAM file" )
67 finally:
68 os.unlink( stderr_name )
69
70
71 def __main__():
72 # Parse Command Line
73 parser = optparse.OptionParser()
74 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.' )
75 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.' )
76 parser.add_option( '-d', '--dataset', dest='datasets', action='append', type="string", nargs=4, help='"-argument" "original_filename" "galaxy_filetype" "name_prefix"' )
77 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.' )
78 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.' )
79 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.' )
80 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.' )
81 parser.add_option( '', '--html_report_from_directory', dest='html_report_from_directory', action='append', type="string", nargs=2, help='"Target HTML File" "Directory"')
82 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)' )
83 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)' )
84 (options, args) = parser.parse_args()
85
86 if options.pass_through_options:
87 cmd = ' '.join( options.pass_through_options )
88 else:
89 cmd = ''
90 if options.pass_through_options_encoded:
91 cmd = '%s %s' % ( cmd, ' '.join( map( unhexlify, options.pass_through_options_encoded ) ) )
92 if options.max_jvm_heap is not None:
93 cmd = cmd.replace( 'java ', 'java -Xmx%s ' % ( options.max_jvm_heap ), 1 )
94 elif options.max_jvm_heap_fraction is not None:
95 cmd = cmd.replace( 'java ', 'java -XX:DefaultMaxRAMFraction=%s -XX:+UseParallelGC ' % ( options.max_jvm_heap_fraction ), 1 )
96 bam_filenames = []
97 tmp_dir = tempfile.mkdtemp( prefix='tmp-gatk-' )
98 try:
99 if options.datasets:
100 for ( dataset_arg, filename, galaxy_ext, prefix ) in options.datasets:
101 gatk_filename = gatk_filename_from_galaxy( filename, galaxy_ext, target_dir=tmp_dir, prefix=prefix )
102 if dataset_arg:
103 cmd = '%s %s "%s"' % ( cmd, gatk_filetype_argument_substitution( dataset_arg, galaxy_ext ), gatk_filename )
104 if galaxy_ext == "bam":
105 bam_filenames.append( gatk_filename )
106 if galaxy_ext == 'fasta':
107 subprocess.check_call( 'samtools faidx "%s"' % gatk_filename, shell=True )
108 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 )
109 index_bam_files( bam_filenames )
110 # set up stdout and stderr output options
111 stdout = open_file_from_option( options.stdout, mode='wb' )
112 stderr = open_file_from_option( options.stderr, mode='wb' )
113 # if no stderr file is specified, we'll use our own
114 if stderr is None:
115 stderr = tempfile.NamedTemporaryFile( prefix="gatk-stderr-", dir=tmp_dir )
116
117 proc = subprocess.Popen( args=cmd, stdout=stdout, stderr=stderr, shell=True, cwd=tmp_dir )
118 return_code = proc.wait()
119
120 if return_code:
121 stderr_target = sys.stderr
122 else:
123 stderr_target = sys.stdout
124 stderr.flush()
125 stderr.seek(0)
126 while True:
127 chunk = stderr.read( CHUNK_SIZE )
128 if chunk:
129 stderr_target.write( chunk )
130 else:
131 break
132 stderr.close()
133 finally:
134 cleanup_before_exit( tmp_dir )
135
136 # generate html reports
137 if options.html_report_from_directory:
138 for ( html_filename, html_dir ) in options.html_report_from_directory:
139 html_report_from_directory( open( html_filename, 'wb' ), html_dir )
140
141
142 if __name__ == "__main__":
143 __main__()