Mercurial > repos > dave > data_manager_mummer4_index_builder
changeset 1:8e59926ddfe8 draft default tip
Uploaded
author | dave |
---|---|
date | Thu, 24 May 2018 21:51:15 -0400 |
parents | a2a04db1bbf3 |
children | |
files | .shed.yml data_manager/mummer4_index_builder.py |
diffstat | 2 files changed, 32 insertions(+), 39 deletions(-) [+] |
line wrap: on
line diff
--- a/.shed.yml Thu May 24 21:11:29 2018 -0400 +++ b/.shed.yml Thu May 24 21:51:15 2018 -0400 @@ -1,10 +1,10 @@ categories: - Data Managers -description: Data Manager for building BWA (0.6+) indexes +description: Data Manager for building Mummer4 indexes homepage_url: '' long_description: | - Data Manager for building BWA (0.6+) indexes. -name: data_manager_bwa_mem_index_builder -owner: devteam -remote_repository_url: https://github.com/galaxyproject/tools-iuc/tree/master/data_managers/data_manager_bwa_mem_index_builder + Data Manager for building Mummer 4 indexes. +name: data_manager_mummer4_index_builder +owner: iuc +remote_repository_url: https://github.com/galaxyproject/tools-iuc/tree/master/data_managers/data_manager_mummer4_index_builder type: unrestricted
--- a/data_manager/mummer4_index_builder.py Thu May 24 21:11:29 2018 -0400 +++ b/data_manager/mummer4_index_builder.py Thu May 24 21:51:15 2018 -0400 @@ -13,7 +13,7 @@ DEFAULT_DATA_TABLE_NAME = "mummer4_indexes" -def get_id_name( params, dbkey, fasta_description=None): +def get_id_name(params, dbkey, fasta_description=None): # TODO: ensure sequence_id is unique and does not already appear in location file sequence_id = params['param_dict']['sequence_id'] if not sequence_id: @@ -27,64 +27,57 @@ return sequence_id, sequence_name -def build_mummer4_index( data_manager_dict, fasta_filename, params, target_directory, dbkey, sequence_id, sequence_name, data_table_name=DEFAULT_DATA_TABLE_NAME ): +def build_mummer4_index(data_manager_dict, fasta_filename, params, target_directory, dbkey, sequence_id, sequence_name, data_table_name=DEFAULT_DATA_TABLE_NAME): # TODO: allow multiple FASTA input files - fasta_base_name = os.path.split( fasta_filename )[-1] - sym_linked_fasta_filename = os.path.join( target_directory, fasta_base_name ) - os.symlink( fasta_filename, sym_linked_fasta_filename ) - if params['param_dict']['index_algorithm'] == 'automatic': - if os.stat( fasta_filename ).st_size < TWO_GB: # use 2 GB as cut off for memory vs. max of 2gb database size; this is somewhat arbitrary - index_algorithm = 'is' - else: - index_algorithm = 'bwtsw' - else: - index_algorithm = params['param_dict']['index_algorithm'] + fasta_base_name = os.path.split(fasta_filename)[-1] + sym_linked_fasta_filename = os.path.join(target_directory, fasta_base_name) + os.symlink(fasta_filename, sym_linked_fasta_filename) - args = [ 'mummer4', '-save', dbkey, sym_linked_fasta_filename, sym_linked_fasta_filename ] - args.append( sym_linked_fasta_filename ) - proc = subprocess.Popen( args=args, shell=False, cwd=target_directory ) + args = ['mummer', '-save', dbkey, sym_linked_fasta_filename, sym_linked_fasta_filename] + args.append(sym_linked_fasta_filename) + proc = subprocess.Popen(args=args, shell=False, cwd=target_directory) return_code = proc.wait() if return_code: print("Error building index.", file=sys.stderr) - sys.exit( return_code ) - data_table_entry = dict( value=sequence_id, dbkey=dbkey, name=sequence_name, path=fasta_base_name ) - _add_data_table_entry( data_manager_dict, data_table_name, data_table_entry ) + sys.exit(return_code) + data_table_entry = dict(value=sequence_id, dbkey=dbkey, name=sequence_name, path=dbkey) + _add_data_table_entry(data_manager_dict, data_table_name, data_table_entry) -def _add_data_table_entry( data_manager_dict, data_table_name, data_table_entry ): - data_manager_dict['data_tables'] = data_manager_dict.get( 'data_tables', {} ) - data_manager_dict['data_tables'][ data_table_name ] = data_manager_dict['data_tables'].get( data_table_name, [] ) - data_manager_dict['data_tables'][ data_table_name ].append( data_table_entry ) +def _add_data_table_entry(data_manager_dict, data_table_name, data_table_entry): + data_manager_dict['data_tables'] = data_manager_dict.get('data_tables', {}) + data_manager_dict['data_tables'][data_table_name] = data_manager_dict['data_tables'].get(data_table_name, []) + data_manager_dict['data_tables'][data_table_name].append(data_table_entry) return data_manager_dict def main(): parser = optparse.OptionParser() - parser.add_option( '-f', '--fasta_filename', dest='fasta_filename', action='store', type="string", default=None, help='fasta_filename' ) - parser.add_option( '-d', '--fasta_dbkey', dest='fasta_dbkey', action='store', type="string", default=None, help='fasta_dbkey' ) - parser.add_option( '-t', '--fasta_description', dest='fasta_description', action='store', type="string", default=None, help='fasta_description' ) - parser.add_option( '-n', '--data_table_name', dest='data_table_name', action='store', type="string", default=None, help='data_table_name' ) + parser.add_option('-f', '--fasta_filename', dest='fasta_filename', action='store', type="string", default=None, help='fasta_filename') + parser.add_option('-d', '--fasta_dbkey', dest='fasta_dbkey', action='store', type="string", default=None, help='fasta_dbkey') + parser.add_option('-t', '--fasta_description', dest='fasta_description', action='store', type="string", default=None, help='fasta_description') + parser.add_option('-n', '--data_table_name', dest='data_table_name', action='store', type="string", default=None, help='data_table_name') (options, args) = parser.parse_args() filename = args[0] - params = loads( open( filename ).read() ) - target_directory = params[ 'output_data' ][0]['extra_files_path'] - os.mkdir( target_directory ) + params = loads(open(filename).read()) + target_directory = params['output_data'][0]['extra_files_path'] + os.mkdir(target_directory) data_manager_dict = {} dbkey = options.fasta_dbkey - if dbkey in [ None, '', '?' ]: - raise Exception( '"%s" is not a valid dbkey. You must specify a valid dbkey.' % ( dbkey ) ) + if dbkey in [None, '', '?']: + raise Exception('"%s" is not a valid dbkey. You must specify a valid dbkey.' % (dbkey)) - sequence_id, sequence_name = get_id_name( params, dbkey=dbkey, fasta_description=options.fasta_description ) + sequence_id, sequence_name = get_id_name(params, dbkey=dbkey, fasta_description=options.fasta_description) # build the index - build_mummer4_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 ) + build_mummer4_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) # save info to json file - open( filename, 'wb' ).write( dumps( data_manager_dict ) ) + open(filename, 'wb').write(dumps(data_manager_dict)) if __name__ == "__main__":