Mercurial > repos > devteam > data_manager_sam_fasta_index_builder

comparison data_manager/data_manager_sam_fasta_index_builder.py @ 0:d6b6db69d9bf draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Uploaded data manager definition.
author devteam
date Thu, 27 Mar 2014 14:15:02 -0400
parents
children 2a1ac1abc3f7
comparison
equal deleted inserted replaced
-1:000000000000 0:d6b6db69d9bf
1 #!/usr/bin/env python
2 #Dan Blankenberg
3
4 import json
5 import optparse
6 import os
7 import subprocess
8 import sys
9 import tempfile
10
11 CHUNK_SIZE = 2**20
12
13 DEFAULT_DATA_TABLE_NAME = "fasta_indexes"
14
15 def get_id_name( params, dbkey, fasta_description=None):
16 #TODO: ensure sequence_id is unique and does not already appear in location file
17 sequence_id = params['param_dict']['sequence_id']
18 if not sequence_id:
19 sequence_id = dbkey
20
21 sequence_name = params['param_dict']['sequence_name']
22 if not sequence_name:
23 sequence_name = fasta_description
24 if not sequence_name:
25 sequence_name = dbkey
26 return sequence_id, sequence_name
27
28 def build_sam_index( data_manager_dict, fasta_filename, target_directory, dbkey, sequence_id, sequence_name, data_table_name=DEFAULT_DATA_TABLE_NAME ):
29 #TODO: allow multiple FASTA input files
30 fasta_base_name = os.path.split( fasta_filename )[-1]
31 sym_linked_fasta_filename = os.path.join( target_directory, fasta_base_name )
32 os.symlink( fasta_filename, sym_linked_fasta_filename )
33
34 args = [ 'samtools', 'faidx' ]
35 args.append( sym_linked_fasta_filename )
36 tmp_stderr = tempfile.NamedTemporaryFile( prefix = "tmp-data-manager-sam_fa_index_builder-stderr" )
37 proc = subprocess.Popen( args=args, shell=False, cwd=target_directory, stderr=tmp_stderr.fileno() )
38 return_code = proc.wait()
39 if return_code:
40 tmp_stderr.flush()
41 tmp_stderr.seek( 0 )
42 sys.stderr.write( "Error building index:\n" )
43 while True:
44 chunk = tmp_stderr.read( CHUNK_SIZE )
45 if not chunk:
46 break
47 sys.stderr.write( chunk )
48 sys.exit( return_code )
49 tmp_stderr.close()
50 data_table_entry = dict( value=sequence_id, dbkey=dbkey, name=sequence_name, path=fasta_base_name )
51 _add_data_table_entry( data_manager_dict, data_table_name, data_table_entry )
52
53 def _add_data_table_entry( data_manager_dict, data_table_name, data_table_entry ):
54 data_manager_dict['data_tables'] = data_manager_dict.get( 'data_tables', {} )
55 data_manager_dict['data_tables'][ data_table_name ] = data_manager_dict['data_tables'].get( data_table_name, [] )
56 data_manager_dict['data_tables'][ data_table_name ].append( data_table_entry )
57 return data_manager_dict
58
59 def main():
60 #Parse Command Line
61 parser = optparse.OptionParser()
62 parser.add_option( '-f', '--fasta_filename', dest='fasta_filename', action='store', type="string", default=None, help='fasta_filename' )
63 parser.add_option( '-d', '--fasta_dbkey', dest='fasta_dbkey', action='store', type="string", default=None, help='fasta_dbkey' )
64 parser.add_option( '-t', '--fasta_description', dest='fasta_description', action='store', type="string", default=None, help='fasta_description' )
65 parser.add_option( '-n', '--data_table_name', dest='data_table_name', action='store', type="string", default=None, help='data_table_name' )
66 (options, args) = parser.parse_args()
67
68 filename = args[0]
69
70 params = json.loads( open( filename ).read() )
71 target_directory = params[ 'output_data' ][0]['extra_files_path']
72 os.mkdir( target_directory )
73 data_manager_dict = {}
74
75 if options.fasta_dbkey in [ None, '', '?' ]:
76 raise Exception( '"%s" is not a valid dbkey. You must specify a valid dbkey.' % ( options.fasta_dbkey ) )
77
78 sequence_id, sequence_name = get_id_name( params, dbkey=options.fasta_dbkey, fasta_description=options.fasta_description )
79
80 #build the index
81 build_sam_index( data_manager_dict, options.fasta_filename, target_directory, options.fasta_dbkey, sequence_id, sequence_name, data_table_name=options.data_table_name or DEFAULT_DATA_TABLE_NAME )
82
83 #save info to json file
84 open( filename, 'wb' ).write( json.dumps( data_manager_dict ) )
85
86 if __name__ == "__main__": main()