Mercurial > repos > jjohnson > defuse

comparison datamanager_create_reference.py @ 11:b22f8634ff84 draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
planemo upload for repository https://github.com/jj-umn/galaxytools/tree/master/defuse commit 23b94b5747c6956360cd2eca0a07a669929ea141-dirty
author jjohnson
date Sun, 17 Jan 2016 14:11:06 -0500
parents
children
comparison
equal deleted inserted replaced
10:f65857c1b92e 11:b22f8634ff84
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
18 def stop_err(msg):
19 sys.stderr.write(msg)
20 sys.exit(1)
21
22 def get_config_dict(config,dataset_directory=None):
23 keys = ['dataset_directory','ensembl_organism','ensembl_prefix','ensembl_version','ensembl_genome_version','ucsc_genome_version','ncbi_organism','ncbi_prefix','chromosomes','mt_chromosome','gene_sources','ig_gene_sources','rrna_gene_sources']
24 pat = '^([^=]+?)\s*=\s*(.*)$'
25 config_dict = {}
26 try:
27 fh = open(config)
28 for i,l in enumerate(fh):
29 line = l.strip()
30 if line.startswith('#'):
31 continue
32 m = re.match(pat,line)
33 if m and len(m.groups()) == 2:
34 (k,v) = m.groups()
35 if k in keys:
36 config_dict[k] = v
37 except Exception, e:
38 stop_err( 'Error parsing %s %s\n' % (config,str( e )) )
39 else:
40 fh.close()
41 if dataset_directory:
42 config_dict['dataset_directory'] = dataset_directory
43 return config_dict
44
45 def run_defuse_script(data_manager_dict, params, target_directory, dbkey, description, config, script):
46 if not os.path.isdir(target_directory):
47 os.makedirs(target_directory)
48 ## Name the config consistently with data_manager_conf.xml
49 # copy the config file to the target_directory
50 # when DataManager moves files to there tool-data location, the config will get moved as well,
51 # and the value_translation in data_manager_conf.xml will tell us the new location
52 # defuse.xml will use the path to this config file to set the dataset_directory
53 config_name = '%s.config' % dbkey
54 defuse_config = os.path.join( target_directory, config_name)
55 shutil.copyfile(config,defuse_config)
56 cmd = "/bin/bash %s %s" % (script,target_directory)
57 # Run
58 try:
59 tmp_out = tempfile.NamedTemporaryFile().name
60 tmp_stdout = open( tmp_out, 'wb' )
61 tmp_err = tempfile.NamedTemporaryFile().name
62 tmp_stderr = open( tmp_err, 'wb' )
63 proc = subprocess.Popen( args=cmd, shell=True, cwd=".", stdout=tmp_stdout, stderr=tmp_stderr )
64 returncode = proc.wait()
65 tmp_stderr.close()
66 # get stderr, allowing for case where it's very large
67 tmp_stderr = open( tmp_err, 'rb' )
68 stderr = ''
69 buffsize = 1048576
70 try:
71 while True:
72 stderr += tmp_stderr.read( buffsize )
73 if not stderr or len( stderr ) % buffsize != 0:
74 break
75 except OverflowError:
76 pass
77 tmp_stdout.close()
78 tmp_stderr.close()
79 if returncode != 0:
80 raise Exception, stderr
81
82 # TODO: look for errors in program output.
83 except Exception, e:
84 stop_err( 'Error creating defuse reference:\n' + str( e ) )
85 config_dict = get_config_dict(config, dataset_directory=target_directory)
86 data_table_entry = dict(value=dbkey, dbkey=dbkey, name=description, path=config_name)
87 _add_data_table_entry( data_manager_dict, data_table_entry )
88 def _add_data_table_entry( data_manager_dict, data_table_entry ):
89 data_manager_dict['data_tables'] = data_manager_dict.get( 'data_tables', {} )
90 data_manager_dict['data_tables']['defuse_reference'] = data_manager_dict['data_tables'].get( 'defuse_reference', [] )
91 data_manager_dict['data_tables']['defuse_reference'].append( data_table_entry )
92 return data_manager_dict
93
94 def main():
95 #Parse Command Line
96 parser = optparse.OptionParser()
97 parser.add_option( '-k', '--dbkey', dest='dbkey', action='store', type="string", default=None, help='dbkey' )
98 parser.add_option( '-d', '--description', dest='description', action='store', type="string", default=None, help='description' )
99 parser.add_option( '-c', '--defuse_config', dest='defuse_config', action='store', type="string", default=None, help='defuse_config' )
100 parser.add_option( '-s', '--defuse_script', dest='defuse_script', action='store', type="string", default=None, help='defuse_script' )
101 (options, args) = parser.parse_args()
102
103 filename = args[0]
104
105 params = from_json_string( open( filename ).read() )
106 target_directory = params[ 'output_data' ][0]['extra_files_path']
107 os.mkdir( target_directory )
108 data_manager_dict = {}
109
110
111 #Create Defuse Reference Data
112 run_defuse_script( data_manager_dict, params, target_directory, options.dbkey, options.description,options.defuse_config,options.defuse_script)
113
114 #save info to json file
115 open( filename, 'wb' ).write( to_json_string( data_manager_dict ) )
116
117 if __name__ == "__main__": main()
118