Mercurial > repos > iuc > data_manager_star_index_builder

comparison data_manager/rnastar_index_builder.py @ 0:e4b87a00b1df draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/data_managers/data_manager_star_index_builder commit 3265247e909410db2a6d6087a2c0d3a9885c120c-dirty
author iuc
date Wed, 23 Nov 2016 17:55:57 -0500
parents
children
comparison
equal deleted inserted replaced
-1:000000000000 0:e4b87a00b1df
1 #!/usr/bin/env python
2 # Dan Blankenberg
3 # adapted from Dan's BWA one for rna star
4 # ross lazarus sept 2014
5 # fixed some stupid bugs January 2015
6 import json
7 import optparse
8 import os
9 import subprocess
10 import sys
11 import tempfile
12
13 CHUNK_SIZE = 2**20
14 ONE_GB = 2**30
15
16 DEFAULT_DATA_TABLE_NAME = "rnastar_index"
17
18
19 def get_id_name( params, dbkey, fasta_description=None):
20 # TODO: ensure sequence_id is unique and does not already appear in location file
21 sequence_id = params['param_dict']['sequence_id']
22 if not sequence_id:
23 sequence_id = dbkey
24
25 sequence_name = params['param_dict']['sequence_name']
26 if not sequence_name:
27 sequence_name = fasta_description
28 if not sequence_name:
29 sequence_name = dbkey
30 return sequence_id, sequence_name
31
32
33 def _add_data_table_entry( data_manager_dict, data_table_name, data_table_entry ):
34 data_manager_dict['data_tables'] = data_manager_dict.get( 'data_tables', {} )
35 data_manager_dict['data_tables'][ data_table_name ] = data_manager_dict['data_tables'].get( data_table_name, [] )
36 data_manager_dict['data_tables'][ data_table_name ].append( data_table_entry )
37 return data_manager_dict
38
39
40 def build_rnastar_index(data_manager_dict, fasta_filename, target_directory,
41 dbkey, sequence_id, sequence_name, data_table_name,
42 sjdbOverhang, sjdbGTFfile, sjdbFileChrStartEnd,
43 sjdbGTFtagExonParentTranscript, sjdbGTFfeatureExon,
44 sjdbGTFchrPrefix, n_threads):
45 # TODO: allow multiple FASTA input files
46 fasta_base_name = os.path.basename( fasta_filename )
47 sym_linked_fasta_filename = os.path.join( target_directory, fasta_base_name )
48 os.symlink( fasta_filename, sym_linked_fasta_filename )
49 # print >> sys.stdout,'made',sym_linked_fasta_filename
50 cl = ['STAR', '--runMode', 'genomeGenerate', '--genomeFastaFiles', sym_linked_fasta_filename, '--genomeDir', target_directory, '--runThreadN', n_threads ]
51
52 if sjdbGTFfile:
53 cl += [ '--sjdbGTFfeatureExon', sjdbGTFfeatureExon, '--sjdbGTFtagExonParentTranscript', sjdbGTFtagExonParentTranscript]
54 if (sjdbGTFchrPrefix > ''):
55 cl += ['--sjdbGTFchrPrefix', sjdbGTFchrPrefix]
56 cl += ['--sjdbOverhang', sjdbOverhang, '--sjdbGTFfile', sjdbGTFfile]
57 elif sjdbFileChrStartEnd:
58 cl += ['--sjdbFileChrStartEnd', sjdbFileChrStartEnd, '--sjdbOverhang', sjdbOverhang]
59
60 tmp_stderr = tempfile.NamedTemporaryFile( prefix="tmp-data-manager-rnastar-index-builder-stderr" )
61 proc = subprocess.Popen( args=cl, shell=False, cwd=target_directory, stderr=tmp_stderr.fileno() )
62 return_code = proc.wait()
63 if return_code:
64 tmp_stderr.flush()
65 tmp_stderr.seek(0)
66 print >> sys.stderr, "Error building index:"
67 while True:
68 chunk = tmp_stderr.read( CHUNK_SIZE )
69 if not chunk:
70 break
71 sys.stderr.write( chunk )
72 sys.exit( return_code )
73 tmp_stderr.close()
74 data_table_entry = dict( value=sequence_id, dbkey=dbkey, name=sequence_name, path=fasta_base_name )
75 data_manager_dict = _add_data_table_entry( data_manager_dict, data_table_name, data_table_entry )
76 return data_manager_dict
77
78
79 def main():
80 # Parse Command Line
81 parser = optparse.OptionParser()
82 parser.add_option( '--fasta_filename', dest='fasta_filename', action='store', type="string", default=None, help='fasta_filename' )
83 parser.add_option( '--fasta_dbkey', dest='fasta_dbkey', action='store', type="string", default=None, help='fasta_dbkey' )
84 parser.add_option( '--fasta_description', dest='fasta_description', action='store', type="string", default=None, help='fasta_description' )
85 parser.add_option( '--data_table_name', dest='data_table_name', action='store', type="string", default=None, help='data_table_name' )
86 parser.add_option( '--sjdbGTFfile', type="string", default=None )
87 parser.add_option( '--sjdbGTFchrPrefix', type="string", default=None )
88 parser.add_option( '--sjdbGTFfeatureExon', type="string", default=None )
89 parser.add_option( '--sjdbGTFtagExonParentTranscript', type="string", default=None )
90 parser.add_option( '--sjdbFileChrStartEnd', type="string", default=None )
91 parser.add_option( '--sjdbOverhang', type="string", default='100' )
92 parser.add_option( '--runThreadN', type="string", default='4' )
93 (options, args) = parser.parse_args()
94 filename = args[0]
95 params = json.loads( open( filename ).read() )
96 target_directory = params[ 'output_data' ][0]['extra_files_path'].encode('ascii', 'replace')
97 dbkey = options.fasta_dbkey
98 if dbkey in [ None, '', '?' ]:
99 raise Exception( '"%s" is not a valid dbkey. You must specify a valid dbkey.' % ( dbkey ) )
100
101 sequence_id, sequence_name = get_id_name( params, dbkey=dbkey, fasta_description=options.fasta_description )
102
103 try:
104 os.mkdir( target_directory )
105 except OSError:
106 pass
107 # build the index
108 data_manager_dict = build_rnastar_index(
109 data_manager_dict={}, fasta_filename=options.fasta_filename,
110 target_directory=target_directory, dbkey=dbkey, sequence_id=sequence_id,
111 sequence_name=sequence_name, data_table_name=options.data_table_name,
112 sjdbOverhang=options.sjdbOverhang, sjdbGTFfile=options.sjdbGTFfile,
113 sjdbFileChrStartEnd=options.sjdbFileChrStartEnd,
114 sjdbGTFtagExonParentTranscript=options.sjdbGTFtagExonParentTranscript,
115 sjdbGTFfeatureExon=options.sjdbGTFfeatureExon,
116 sjdbGTFchrPrefix=options.sjdbGTFchrPrefix, n_threads=options.runThreadN)
117 open( filename, 'wb' ).write( json.dumps( data_manager_dict ) )
118
119
120 if __name__ == "__main__":
121 main()