Mercurial > repos > devteam > data_manager_bowtie2_index_builder
annotate data_manager/bowtie2_index_builder.py @ 3:49331eedb1c5 draft
planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/data_managers/data_manager_bowtie2_index_builder commit c54b039220cc3d1767a424b6ae76852d40e67437
author | devteam |
---|---|
date | Mon, 16 Nov 2015 15:12:06 -0500 |
parents | 8de13c28d92e |
children | 8a892275932f |
rev | line source |
---|---|
0 | 1 #!/usr/bin/env python |
2 #Dan Blankenberg | |
3 | |
4 import sys | |
5 import os | |
6 import optparse | |
7 import subprocess | |
8 | |
9 from json import loads, dumps | |
10 | |
11 | |
2
8de13c28d92e
planemo upload commit 33927a87ba2eee9bf0ecdd376a66241b17b3d734
devteam
parents:
1
diff
changeset
|
12 DEFAULT_DATA_TABLE_NAMES = ["bowtie2_indexes"] |
0 | 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 | |
2
8de13c28d92e
planemo upload commit 33927a87ba2eee9bf0ecdd376a66241b17b3d734
devteam
parents:
1
diff
changeset
|
27 def build_bowtie2_index( data_manager_dict, fasta_filename, params, target_directory, dbkey, sequence_id, sequence_name, data_table_names=DEFAULT_DATA_TABLE_NAMES ): |
0 | 28 #TODO: allow multiple FASTA input files |
29 fasta_base_name = os.path.split( fasta_filename )[-1] | |
30 sym_linked_fasta_filename = os.path.join( target_directory, fasta_base_name ) | |
31 os.symlink( fasta_filename, sym_linked_fasta_filename ) | |
32 args = [ 'bowtie2-build', sym_linked_fasta_filename, sequence_id ] | |
33 proc = subprocess.Popen( args=args, shell=False, cwd=target_directory ) | |
34 return_code = proc.wait() | |
35 if return_code: | |
36 print >> sys.stderr, "Error building index." | |
37 sys.exit( return_code ) | |
1 | 38 data_table_entry = dict( value=sequence_id, dbkey=dbkey, name=sequence_name, path=sequence_id ) |
2
8de13c28d92e
planemo upload commit 33927a87ba2eee9bf0ecdd376a66241b17b3d734
devteam
parents:
1
diff
changeset
|
39 for data_table_name in data_table_names: |
8de13c28d92e
planemo upload commit 33927a87ba2eee9bf0ecdd376a66241b17b3d734
devteam
parents:
1
diff
changeset
|
40 _add_data_table_entry( data_manager_dict, data_table_name, data_table_entry ) |
0 | 41 |
42 def _add_data_table_entry( data_manager_dict, data_table_name, data_table_entry ): | |
43 data_manager_dict['data_tables'] = data_manager_dict.get( 'data_tables', {} ) | |
44 data_manager_dict['data_tables'][ data_table_name ] = data_manager_dict['data_tables'].get( data_table_name, [] ) | |
45 data_manager_dict['data_tables'][ data_table_name ].append( data_table_entry ) | |
46 return data_manager_dict | |
47 | |
48 def main(): | |
49 #Parse Command Line | |
50 parser = optparse.OptionParser() | |
51 parser.add_option( '-f', '--fasta_filename', dest='fasta_filename', action='store', type="string", default=None, help='fasta_filename' ) | |
52 parser.add_option( '-d', '--fasta_dbkey', dest='fasta_dbkey', action='store', type="string", default=None, help='fasta_dbkey' ) | |
53 parser.add_option( '-t', '--fasta_description', dest='fasta_description', action='store', type="string", default=None, help='fasta_description' ) | |
2
8de13c28d92e
planemo upload commit 33927a87ba2eee9bf0ecdd376a66241b17b3d734
devteam
parents:
1
diff
changeset
|
54 parser.add_option( '-n', '--data_table_name', dest='data_table_name', action='append', type="string", default=None, help='data_table_name' ) |
0 | 55 (options, args) = parser.parse_args() |
56 | |
57 filename = args[0] | |
58 | |
59 params = loads( open( filename ).read() ) | |
60 target_directory = params[ 'output_data' ][0]['extra_files_path'] | |
61 os.mkdir( target_directory ) | |
62 data_manager_dict = {} | |
63 | |
64 dbkey = options.fasta_dbkey | |
65 | |
66 if dbkey in [ None, '', '?' ]: | |
67 raise Exception( '"%s" is not a valid dbkey. You must specify a valid dbkey.' % ( dbkey ) ) | |
68 | |
69 sequence_id, sequence_name = get_id_name( params, dbkey=dbkey, fasta_description=options.fasta_description ) | |
70 | |
71 #build the index | |
2
8de13c28d92e
planemo upload commit 33927a87ba2eee9bf0ecdd376a66241b17b3d734
devteam
parents:
1
diff
changeset
|
72 build_bowtie2_index( data_manager_dict, options.fasta_filename, params, target_directory, dbkey, sequence_id, sequence_name, data_table_names=options.data_table_name or DEFAULT_DATA_TABLE_NAMES ) |
0 | 73 |
74 #save info to json file | |
75 open( filename, 'wb' ).write( dumps( data_manager_dict ) ) | |
76 | |
77 if __name__ == "__main__": main() |