Mercurial > repos > devteam > data_manager_sam_fasta_index_builder
annotate data_manager/data_manager_sam_fasta_index_builder.py @ 3:1f54e98616af draft
planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/data_managers/data_manager_sam_fasta_index_builder commit 8652f36a3a3838dca989426961561e81432acf4f
author | iuc |
---|---|
date | Tue, 04 Apr 2017 18:13:56 -0400 |
parents | 2a1ac1abc3f7 |
children | cc4ef4d38cf9 |
rev | line source |
---|---|
0 | 1 #!/usr/bin/env python |
3
1f54e98616af
planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/data_managers/data_manager_sam_fasta_index_builder commit 8652f36a3a3838dca989426961561e81432acf4f
iuc
parents:
2
diff
changeset
|
2 # Dan Blankenberg |
0 | 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 | |
3
1f54e98616af
planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/data_managers/data_manager_sam_fasta_index_builder commit 8652f36a3a3838dca989426961561e81432acf4f
iuc
parents:
2
diff
changeset
|
15 |
0 | 16 def get_id_name( params, dbkey, fasta_description=None): |
3
1f54e98616af
planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/data_managers/data_manager_sam_fasta_index_builder commit 8652f36a3a3838dca989426961561e81432acf4f
iuc
parents:
2
diff
changeset
|
17 # TODO: ensure sequence_id is unique and does not already appear in location file |
0 | 18 sequence_id = params['param_dict']['sequence_id'] |
19 if not sequence_id: | |
20 sequence_id = dbkey | |
3
1f54e98616af
planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/data_managers/data_manager_sam_fasta_index_builder commit 8652f36a3a3838dca989426961561e81432acf4f
iuc
parents:
2
diff
changeset
|
21 |
0 | 22 sequence_name = params['param_dict']['sequence_name'] |
23 if not sequence_name: | |
24 sequence_name = fasta_description | |
25 if not sequence_name: | |
26 sequence_name = dbkey | |
27 return sequence_id, sequence_name | |
28 | |
3
1f54e98616af
planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/data_managers/data_manager_sam_fasta_index_builder commit 8652f36a3a3838dca989426961561e81432acf4f
iuc
parents:
2
diff
changeset
|
29 |
0 | 30 def build_sam_index( data_manager_dict, fasta_filename, target_directory, dbkey, sequence_id, sequence_name, data_table_name=DEFAULT_DATA_TABLE_NAME ): |
3
1f54e98616af
planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/data_managers/data_manager_sam_fasta_index_builder commit 8652f36a3a3838dca989426961561e81432acf4f
iuc
parents:
2
diff
changeset
|
31 # TODO: allow multiple FASTA input files |
2
2a1ac1abc3f7
planemo upload commit 5ad726dc73203a704666033cd3bf70b82575978f
devteam
parents:
0
diff
changeset
|
32 assert os.path.exists( fasta_filename ), 'FASTA file "%s" is missing, cannot build samtools index.' % fasta_filename |
0 | 33 fasta_base_name = os.path.split( fasta_filename )[-1] |
34 sym_linked_fasta_filename = os.path.join( target_directory, fasta_base_name ) | |
35 os.symlink( fasta_filename, sym_linked_fasta_filename ) | |
3
1f54e98616af
planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/data_managers/data_manager_sam_fasta_index_builder commit 8652f36a3a3838dca989426961561e81432acf4f
iuc
parents:
2
diff
changeset
|
36 |
0 | 37 args = [ 'samtools', 'faidx' ] |
38 args.append( sym_linked_fasta_filename ) | |
3
1f54e98616af
planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/data_managers/data_manager_sam_fasta_index_builder commit 8652f36a3a3838dca989426961561e81432acf4f
iuc
parents:
2
diff
changeset
|
39 tmp_stderr = tempfile.NamedTemporaryFile(prefix="tmp-data-manager-sam_fa_index_builder-stderr") |
0 | 40 proc = subprocess.Popen( args=args, shell=False, cwd=target_directory, stderr=tmp_stderr.fileno() ) |
41 return_code = proc.wait() | |
42 if return_code: | |
43 tmp_stderr.flush() | |
44 tmp_stderr.seek( 0 ) | |
45 sys.stderr.write( "Error building index:\n" ) | |
46 while True: | |
47 chunk = tmp_stderr.read( CHUNK_SIZE ) | |
48 if not chunk: | |
49 break | |
50 sys.stderr.write( chunk ) | |
51 sys.exit( return_code ) | |
52 tmp_stderr.close() | |
53 data_table_entry = dict( value=sequence_id, dbkey=dbkey, name=sequence_name, path=fasta_base_name ) | |
54 _add_data_table_entry( data_manager_dict, data_table_name, data_table_entry ) | |
55 | |
3
1f54e98616af
planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/data_managers/data_manager_sam_fasta_index_builder commit 8652f36a3a3838dca989426961561e81432acf4f
iuc
parents:
2
diff
changeset
|
56 |
0 | 57 def _add_data_table_entry( data_manager_dict, data_table_name, data_table_entry ): |
58 data_manager_dict['data_tables'] = data_manager_dict.get( 'data_tables', {} ) | |
59 data_manager_dict['data_tables'][ data_table_name ] = data_manager_dict['data_tables'].get( data_table_name, [] ) | |
60 data_manager_dict['data_tables'][ data_table_name ].append( data_table_entry ) | |
61 return data_manager_dict | |
62 | |
3
1f54e98616af
planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/data_managers/data_manager_sam_fasta_index_builder commit 8652f36a3a3838dca989426961561e81432acf4f
iuc
parents:
2
diff
changeset
|
63 |
0 | 64 def main(): |
65 parser = optparse.OptionParser() | |
66 parser.add_option( '-f', '--fasta_filename', dest='fasta_filename', action='store', type="string", default=None, help='fasta_filename' ) | |
67 parser.add_option( '-d', '--fasta_dbkey', dest='fasta_dbkey', action='store', type="string", default=None, help='fasta_dbkey' ) | |
68 parser.add_option( '-t', '--fasta_description', dest='fasta_description', action='store', type="string", default=None, help='fasta_description' ) | |
69 parser.add_option( '-n', '--data_table_name', dest='data_table_name', action='store', type="string", default=None, help='data_table_name' ) | |
70 (options, args) = parser.parse_args() | |
3
1f54e98616af
planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/data_managers/data_manager_sam_fasta_index_builder commit 8652f36a3a3838dca989426961561e81432acf4f
iuc
parents:
2
diff
changeset
|
71 |
0 | 72 filename = args[0] |
3
1f54e98616af
planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/data_managers/data_manager_sam_fasta_index_builder commit 8652f36a3a3838dca989426961561e81432acf4f
iuc
parents:
2
diff
changeset
|
73 |
0 | 74 params = json.loads( open( filename ).read() ) |
75 target_directory = params[ 'output_data' ][0]['extra_files_path'] | |
76 os.mkdir( target_directory ) | |
77 data_manager_dict = {} | |
3
1f54e98616af
planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/data_managers/data_manager_sam_fasta_index_builder commit 8652f36a3a3838dca989426961561e81432acf4f
iuc
parents:
2
diff
changeset
|
78 |
0 | 79 if options.fasta_dbkey in [ None, '', '?' ]: |
80 raise Exception( '"%s" is not a valid dbkey. You must specify a valid dbkey.' % ( options.fasta_dbkey ) ) | |
3
1f54e98616af
planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/data_managers/data_manager_sam_fasta_index_builder commit 8652f36a3a3838dca989426961561e81432acf4f
iuc
parents:
2
diff
changeset
|
81 |
0 | 82 sequence_id, sequence_name = get_id_name( params, dbkey=options.fasta_dbkey, fasta_description=options.fasta_description ) |
3
1f54e98616af
planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/data_managers/data_manager_sam_fasta_index_builder commit 8652f36a3a3838dca989426961561e81432acf4f
iuc
parents:
2
diff
changeset
|
83 |
1f54e98616af
planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/data_managers/data_manager_sam_fasta_index_builder commit 8652f36a3a3838dca989426961561e81432acf4f
iuc
parents:
2
diff
changeset
|
84 # build the index |
0 | 85 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 ) |
3
1f54e98616af
planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/data_managers/data_manager_sam_fasta_index_builder commit 8652f36a3a3838dca989426961561e81432acf4f
iuc
parents:
2
diff
changeset
|
86 |
1f54e98616af
planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/data_managers/data_manager_sam_fasta_index_builder commit 8652f36a3a3838dca989426961561e81432acf4f
iuc
parents:
2
diff
changeset
|
87 # save info to json file |
0 | 88 open( filename, 'wb' ).write( json.dumps( data_manager_dict ) ) |
3
1f54e98616af
planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/data_managers/data_manager_sam_fasta_index_builder commit 8652f36a3a3838dca989426961561e81432acf4f
iuc
parents:
2
diff
changeset
|
89 |
1f54e98616af
planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/data_managers/data_manager_sam_fasta_index_builder commit 8652f36a3a3838dca989426961561e81432acf4f
iuc
parents:
2
diff
changeset
|
90 |
1f54e98616af
planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/data_managers/data_manager_sam_fasta_index_builder commit 8652f36a3a3838dca989426961561e81432acf4f
iuc
parents:
2
diff
changeset
|
91 if __name__ == "__main__": |
1f54e98616af
planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/data_managers/data_manager_sam_fasta_index_builder commit 8652f36a3a3838dca989426961561e81432acf4f
iuc
parents:
2
diff
changeset
|
92 main() |