comparison data_manager/bwa_index_builder.py @ 0:ae10d0ddbbb1 draft default tip

planemo upload for repository https://github.com/galaxyproject/tools-devteam/tree/master/data_managers/data_manager_bwa_index_builder commit 86cf90107482cab1cb47fc0d42d6705f8077daa7

0:ae10d0ddbbb1

#!/usr/bin/env python
#Dan Blankenberg

import sys
import os
import tempfile
import optparse
import subprocess

from json import loads, dumps


CHUNK_SIZE = 2**20
ONE_GB = 2**30

DEFAULT_DATA_TABLE_NAME = "bwa_indexes"

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:
        sequence_id = dbkey
    
    sequence_name = params['param_dict']['sequence_name']
    if not sequence_name:
        sequence_name = fasta_description
        if not sequence_name:
            sequence_name = dbkey
    return sequence_id, sequence_name

def build_bwa_index( data_manager_dict, fasta_filename, params, target_directory, dbkey, sequence_id, sequence_name, data_table_name=DEFAULT_DATA_TABLE_NAME, color_space = False ):
    #TODO: allow multiple FASTA input files
    #tmp_dir = tempfile.mkdtemp( prefix='tmp-data-manager-bwa-index-builder-' )
    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 <= ONE_GB: #use 1 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']
    
    args = [ 'bwa', 'index', '-a', index_algorithm ]
    if color_space:
        args.append( '-c' )
    args.append( sym_linked_fasta_filename )
    proc = subprocess.Popen( args=args, shell=False, cwd=target_directory )
    return_code = proc.wait()
    if return_code:
        print >> sys.stderr, "Error building index."
        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 )

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():
    #Parse Command Line
    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( '-c', '--color_space', dest='color_space', action='store_true', default=False, help='color_space' )
    (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 )
    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 ) )
    
    sequence_id, sequence_name = get_id_name( params, dbkey=dbkey, fasta_description=options.fasta_description )
    
    #build the index
    build_bwa_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, color_space=options.color_space )
    
    #save info to json file
    open( filename, 'wb' ).write( dumps( data_manager_dict ) )
        
if __name__ == "__main__": main()