Mercurial > repos > jjohnson > snpeff

comparison data_manager/data_manager_snpEff_download.py @ 0:e1d9f6a0ad53

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Uploaded
author jjohnson
date Thu, 04 Jul 2013 10:43:55 -0400
parents
children 6ad9205c1307
comparison
equal deleted inserted replaced
-1:000000000000 0:e1d9f6a0ad53
1 #!/usr/bin/env python
2
3 import sys
4 import os
5 import re
6 import tempfile
7 import subprocess
8 import fileinput
9 import shutil
10 import optparse
11 import urllib2
12 from ftplib import FTP
13 import tarfile
14
15 from galaxy.util.json import from_json_string, to_json_string
16
17 def stop_err(msg):
18 sys.stderr.write(msg)
19 sys.exit(1)
20
21 """
22 # Download human database 'hg19'
23 java -jar snpEff.jar download -v hg19
24
25 <command>java -jar \$JAVA_JAR_PATH/snpEff.jar download -c \$JAVA_JAR_PATH/snpEff.config $genomeVersion > $logfile </command>
26
27 snpEffectPredictor.bin
28 regulation_HeLa-S3.bin
29 regulation_pattern = 'regulation_(.+).bin'
30
31
32 """
33 def download_database(data_manager_dict, target_directory, jar_path,config,genome_version,organism):
34 ## get data_dir from config
35 ##---
36 ## Databases are stored here
37 ## E.g.: Information for 'hg19' is stored in data_dir/hg19/
38 ##
39 ## Note: Since version 2.1 you can use tilde ('~') as first character to refer to your home directory
40 ##---
41 #data_dir = ~/snpEff/data/
42 data_dir = None
43 try:
44 fh = open(config)
45 for i,line in enumerate(fh):
46 if line.strip().startswith('data_dir'):
47 (k,v) = line.split('=')
48 data_dir = os.path.expanduser(v.strip())
49 break
50 except Exception, e:
51 stop_err( 'Error parsing %s %s\n' % (config,str( e )) )
52 else:
53 fh.close()
54 (snpEff_dir,snpEff_jar) = os.path.split(jar_path)
55 args = [ 'java','-jar' ]
56 args.append( jar_path )
57 args.append( 'download' )
58 # args.append( '-c' )
59 # args.append( 'config' )
60 args.append( '-v' )
61 args.append( genome_version )
62 proc = subprocess.Popen( args=args, shell=False, cwd=snpEff_dir )
63 return_code = proc.wait()
64 if return_code:
65 sys.exit( return_code )
66 ## search data_dir/genome_version for files
67 regulation_pattern = 'regulation_(.+).bin'
68 genome_path = os.path.join(data_dir,genome_version)
69 if os.path.isdir(genome_path):
70 for root, dirs, files in os.walk(genome_path):
71 for fname in files:
72 if fname.startswith('snpEffectPredictor'):
73 # if snpEffectPredictor.bin download succeeded
74 name = genome_version + (' : ' + organism if organism else '')
75 data_table_entry = dict(value=genome_version, name=name)
76 _add_data_table_entry( data_manager_dict, 'snpeff_genomedb', data_table_entry )
77 else:
78 m = re.match(regulation_pattern,fname)
79 if m:
80 name = m.groups()[0]
81 data_table_entry = dict(value=genome_version, name=name)
82 _add_data_table_entry( data_manager_dict, 'snpeff_regulationdb', data_table_entry )
83 return data_manager_dict
84
85 def _add_data_table_entry( data_manager_dict, data_table, data_table_entry ):
86 data_manager_dict['data_tables'] = data_manager_dict.get( 'data_tables', {} )
87 data_manager_dict['data_tables'][data_table] = data_manager_dict['data_tables'].get( data_table, [] )
88 data_manager_dict['data_tables'][data_table].append( data_table_entry )
89 return data_manager_dict
90
91 def main():
92 #Parse Command Line
93 parser = optparse.OptionParser()
94 parser.add_option( '-j', '--jar_path', dest='jar_path', action='store', type="string", default=None, help='snpEff.jar path' )
95 parser.add_option( '-c', '--config', dest='config', action='store', type="string", default=None, help='snpEff.config path' )
96 parser.add_option( '-g', '--genome_version', dest='genome_version', action='store', type="string", default=None, help='genome_version' )
97 parser.add_option( '-o', '--organism', dest='organism', action='store', type="string", default=None, help='organism name' )
98 (options, args) = parser.parse_args()
99
100 filename = args[0]
101
102 params = from_json_string( open( filename ).read() )
103 target_directory = params[ 'output_data' ][0]['extra_files_path']
104 os.mkdir( target_directory )
105 data_manager_dict = {}
106
107
108 #Create SnpEff Reference Data
109 download_database(data_manager_dict, target_directory, options.jar_path,options.config,options.genome_version,options.organism)
110
111 #save info to json file
112 open( filename, 'wb' ).write( to_json_string( data_manager_dict ) )
113
114 if __name__ == "__main__": main()
115