comparison data_manager/salmon_index_builder.py @ 1:c5dea2080109 draft

Uploaded
author ieguinoa
date Tue, 14 Aug 2018 11:19:49 -0400
parents
children f7d9182bdcab
comparison
equal deleted inserted replaced
0:6cd60ba8a842 1:c5dea2080109
1 #!/usr/bin/env python
2 # Based heavily on the kallisto data manager wrapper script by iuc
3 from __future__ import print_function
4
5 import argparse
6 import os
7 import subprocess
8 import sys
9 from json import dumps, loads
10
11 DEFAULT_DATA_TABLE_NAME = "salmon_indexes"
12
13
14 def get_id_name( params, dbkey, fasta_description=None):
15 # TODO: ensure sequence_id is unique and does not already appear in location file
16 sequence_id = params['param_dict']['sequence_id']
17 if not sequence_id:
18 sequence_id = dbkey
19
20 sequence_name = params['param_dict']['sequence_name']
21 if not sequence_name:
22 sequence_name = fasta_description
23 if not sequence_name:
24 sequence_name = dbkey
25 return sequence_id, sequence_name
26
27
28 def build_salmon_index( data_manager_dict, options, params, sequence_id, sequence_name ):
29 data_table_name = options.data_table_name or DEFAULT_DATA_TABLE_NAME
30 target_directory = params[ 'output_data' ][0]['extra_files_path']
31 if not os.path.exists( target_directory ):
32 os.mkdir( target_directory )
33 if options.kmer_size != '':
34 args.append('-k')
35 args.append(options.kmer_size)
36 args.extend( [ '-t' , options.fasta_filename, '-i', sequence_id ] )
37 proc = subprocess.Popen( args=args, shell=False, cwd=target_directory )
38 return_code = proc.wait()
39 if return_code:
40 print("Error building index.", file=sys.stderr)
41 sys.exit( return_code )
42 data_table_entry = dict( value=sequence_id, dbkey=options.fasta_dbkey, name=sequence_name, path=sequence_id )
43 _add_data_table_entry( data_manager_dict, data_table_name, data_table_entry )
44
45
46 def _add_data_table_entry( data_manager_dict, data_table_name, data_table_entry ):
47 data_manager_dict['data_tables'] = data_manager_dict.get( 'data_tables', {} )
48 data_manager_dict['data_tables'][ data_table_name ] = data_manager_dict['data_tables'].get( data_table_name, [] )
49 data_manager_dict['data_tables'][ data_table_name ].append( data_table_entry )
50 return data_manager_dict
51
52
53 def main():
54 # Parse Command Line
55 parser = argparse.ArgumentParser()
56 parser.add_argument( '--output', dest='output', action='store', type=str, default=None )
57 parser.add_argument( '--fasta_filename', dest='fasta_filename', action='store', type=str, default=None )
58 parser.add_argument( '--fasta_dbkey', dest='fasta_dbkey', action='store', type=str, default=None )
59 parser.add_argument( '--fasta_description', dest='fasta_description', action='store', type=str, default=None )
60 parser.add_argument( '--data_table_name', dest='data_table_name', action='store', type=str, default='salmon_indexes' )
61 parser.add_argument( '-k', '--kmer_size', dest='kmer_size', action='store', type=str, help='kmer_size' )
62 options = parser.parse_args()
63
64 filename = options.output
65
66 params = loads( open( filename ).read() )
67 data_manager_dict = {}
68
69 if options.fasta_dbkey in [ None, '', '?' ]:
70 raise Exception( '"%s" is not a valid dbkey. You must specify a valid dbkey.' % ( options.fasta_dbkey ) )
71
72 sequence_id, sequence_name = get_id_name( params, dbkey=options.fasta_dbkey, fasta_description=options.fasta_description )
73
74 # build the index
75 build_salmon_index( data_manager_dict, options, params, sequence_id, sequence_name )
76
77 # save info to json file
78 open( filename, 'w' ).write( dumps( data_manager_dict ) )
79
80
81 if __name__ == "__main__":
82 main()