Mercurial > repos > devteam > data_manager_picard_index_builder

comparison data_manager/picard_index_builder.py @ 5:acaa6719a478 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_picard_index_builder commit 31fc4022a8c2fef50a3aff9210091be21e8450ef"
author iuc
date Mon, 03 Feb 2020 22:27:13 +0000
parents 99847f7b3a32
children fa8c605bf385
comparison
equal deleted inserted replaced
4:99847f7b3a32 5:acaa6719a478
9 from json import dumps, loads 9 from json import dumps, loads
10 10
11 DEFAULT_DATA_TABLE_NAME = "picard_indexes" 11 DEFAULT_DATA_TABLE_NAME = "picard_indexes"
12 12
13 13
14 def get_id_name( params, dbkey, fasta_description=None): 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 15 # TODO: ensure sequence_id is unique and does not already appear in location file
16 sequence_id = params['param_dict']['sequence_id'] 16 sequence_id = params['param_dict']['sequence_id']
17 if not sequence_id: 17 if not sequence_id:
18 sequence_id = dbkey 18 sequence_id = dbkey
19 19
23 if not sequence_name: 23 if not sequence_name:
24 sequence_name = dbkey 24 sequence_name = dbkey
25 return sequence_id, sequence_name 25 return sequence_id, sequence_name
26 26
27 27
28 def build_picard_index( data_manager_dict, fasta_filename, params, target_directory, dbkey, sequence_id, sequence_name, data_table_name=DEFAULT_DATA_TABLE_NAME ): 28 def build_picard_index(data_manager_dict, fasta_filename, params, target_directory, dbkey, sequence_id, sequence_name, data_table_name=DEFAULT_DATA_TABLE_NAME):
29 # TODO: allow multiple FASTA input files 29 # TODO: allow multiple FASTA input files
30 fasta_base_name = os.path.split( fasta_filename )[-1] 30 fasta_base_name = os.path.split(fasta_filename)[-1]
31 prefix = os.path.splitext( fasta_base_name )[0] 31 prefix = os.path.splitext(fasta_base_name)[0]
32 sym_linked_fasta_filename = os.path.join( target_directory, fasta_base_name ) 32 sym_linked_fasta_filename = os.path.join(target_directory, fasta_base_name)
33 os.symlink( fasta_filename, sym_linked_fasta_filename ) 33 os.symlink(fasta_filename, sym_linked_fasta_filename)
34 args = [ 'picard', 'CreateSequenceDictionary', 'R=%s' % sym_linked_fasta_filename, 'O=%s.dict' % prefix ] 34 args = ['picard', 'CreateSequenceDictionary', 'R=%s' % sym_linked_fasta_filename, 'O=%s.dict' % prefix]
35 proc = subprocess.Popen( args=args, shell=False, cwd=target_directory ) 35 proc = subprocess.Popen(args=args, shell=False, cwd=target_directory)
36 return_code = proc.wait() 36 return_code = proc.wait()
37 if return_code: 37 if return_code:
38 print("Error building index.", file=sys.stderr) 38 print("Error building index.", file=sys.stderr)
39 sys.exit( return_code ) 39 sys.exit(return_code)
40 data_table_entry = dict( value=sequence_id, dbkey=dbkey, name=sequence_name, path=fasta_base_name ) 40 data_table_entry = dict(value=sequence_id, dbkey=dbkey, name=sequence_name, path=fasta_base_name)
41 _add_data_table_entry( data_manager_dict, data_table_name, data_table_entry ) 41 _add_data_table_entry(data_manager_dict, data_table_name, data_table_entry)
42 42
43 43
44 def _add_data_table_entry( data_manager_dict, data_table_name, data_table_entry ): 44 def _add_data_table_entry(data_manager_dict, data_table_name, data_table_entry):
45 data_manager_dict['data_tables'] = data_manager_dict.get( 'data_tables', {} ) 45 data_manager_dict['data_tables'] = data_manager_dict.get('data_tables', {})
46 data_manager_dict['data_tables'][ data_table_name ] = data_manager_dict['data_tables'].get( data_table_name, [] ) 46 data_manager_dict['data_tables'][data_table_name] = data_manager_dict['data_tables'].get(data_table_name, [])
47 data_manager_dict['data_tables'][ data_table_name ].append( data_table_entry ) 47 data_manager_dict['data_tables'][data_table_name].append(data_table_entry)
48 return data_manager_dict 48 return data_manager_dict
49 49
50 50
51 def main(): 51 def main():
52 parser = optparse.OptionParser() 52 parser = optparse.OptionParser()
53 parser.add_option( '-f', '--fasta_filename', dest='fasta_filename', action='store', type="string", default=None, help='fasta_filename' ) 53 parser.add_option('-f', '--fasta_filename', dest='fasta_filename', action='store', type="string", default=None, help='fasta_filename')
54 parser.add_option( '-d', '--fasta_dbkey', dest='fasta_dbkey', action='store', type="string", default=None, help='fasta_dbkey' ) 54 parser.add_option('-d', '--fasta_dbkey', dest='fasta_dbkey', action='store', type="string", default=None, help='fasta_dbkey')
55 parser.add_option( '-t', '--fasta_description', dest='fasta_description', action='store', type="string", default=None, help='fasta_description' ) 55 parser.add_option('-t', '--fasta_description', dest='fasta_description', action='store', type="string", default=None, help='fasta_description')
56 parser.add_option( '-n', '--data_table_name', dest='data_table_name', action='store', type="string", default=None, help='data_table_name' ) 56 parser.add_option('-n', '--data_table_name', dest='data_table_name', action='store', type="string", default=None, help='data_table_name')
57 (options, args) = parser.parse_args() 57 (options, args) = parser.parse_args()
58 58
59 filename = args[0] 59 filename = args[0]
60 60
61 params = loads( open( filename ).read() ) 61 params = loads(open(filename).read())
62 target_directory = params[ 'output_data' ][0]['extra_files_path'] 62 target_directory = params['output_data'][0]['extra_files_path']
63 os.mkdir( target_directory ) 63 os.mkdir(target_directory)
64 data_manager_dict = {} 64 data_manager_dict = {}
65 65
66 dbkey = options.fasta_dbkey 66 dbkey = options.fasta_dbkey
67 67
68 if dbkey in [ None, '', '?' ]: 68 if dbkey in [None, '', '?']:
69 raise Exception( '"%s" is not a valid dbkey. You must specify a valid dbkey.' % ( dbkey ) ) 69 raise Exception('"%s" is not a valid dbkey. You must specify a valid dbkey.' % (dbkey))
70 70
71 sequence_id, sequence_name = get_id_name( params, dbkey=dbkey, fasta_description=options.fasta_description ) 71 sequence_id, sequence_name = get_id_name(params, dbkey=dbkey, fasta_description=options.fasta_description)
72 72
73 # build the index 73 # build the index
74 build_picard_index( data_manager_dict, options.fasta_filename, params, target_directory, dbkey, sequence_id, sequence_name, data_table_name=options.data_table_name or DEFAULT_DATA_TABLE_NAME ) 74 build_picard_index(data_manager_dict, options.fasta_filename, params, target_directory, dbkey, sequence_id,
75 sequence_name, data_table_name=options.data_table_name or DEFAULT_DATA_TABLE_NAME)
75 76
76 # save info to json file 77 # save info to json file
77 open( filename, 'wb' ).write( dumps( data_manager_dict ) ) 78 open(filename, 'w').write(dumps(data_manager_dict, sort_keys=True))
78 79
79 80
80 if __name__ == "__main__": 81 if __name__ == "__main__":
81 main() 82 main()