annotate sspace.py @ 0:b1be94418db9 draft

Uploaded

# -*- coding: utf-8 -*-
"""
SSPACE wrapper
"""

import logging
import optparse
import os
import shutil
import subprocess
import tempfile


def which(name, flags=os.X_OK):
    """ Search PATH for executable files with the given name. """
    result = []
    exts = filter(None, os.environ.get('PATHEXT', '').split(os.pathsep))
    path = os.environ.get('PATH', None)
    if path is None:
        return []
    for p in os.environ.get('PATH', '').split(os.pathsep):
        p = os.path.join(p, str(name))
        if os.access(p, flags):
            result.append(p)
            for e in exts:
                pext = p + e
                if os.access(pext, flags):
                    result.append(pext)
    return result


LOG_FORMAT = '%(asctime)s|%(levelname)-8s|%(message)s'
LOG_DATEFMT = '%Y-%m-%d %H:%M:%S'
LOG_LEVELS = ['DEBUG', 'INFO', 'WARNING', 'ERROR', 'CRITICAL']


def __main__():
    """ main function """
    parser = optparse.OptionParser()
    parser.add_option('-c', dest='contigs', help='contigs mandatory (-s)')
    parser.add_option('--r1', dest='reads1', help='')
    parser.add_option('--r2', dest='reads2', help='')
    parser.add_option('-i', dest='insert', type='int', help='')
    parser.add_option('-e', dest='error', type='float', help='')
    parser.add_option('-o', dest='orientation', choices=['FF', 'FR', 'RF', 'RR'], help='')
    parser.add_option('-x', action='store_true', dest='extension', help='whether to extend the contigs (-x)')
    parser.add_option('--minoverlap', dest='minoverlap', type='int', help='minoverlap (-m)')
    parser.add_option('--numofreads', dest='numofreads', type='int', help='num of reads to call an extension (-o)')
    parser.add_option('-t', dest='max_trim', type='int', help='maximum number of bases to trim on the contig end when all possibilities have been exhausted for an extension (-t)')
    parser.add_option('-u', dest='unpaired', help='unpaired reads (-u, optional)')
    parser.add_option('-r', dest='min_base_ratio', type='float', help='minimum base ratio used to accept a overhang consensus base (-r)')
    parser.add_option('--minlink', dest='minlink', type='int', help='min link (-k)')
    parser.add_option('--maxratio', dest='maxratio', type='float', help='max ratio (-a)')
    parser.add_option('--contigoverlap', dest='contigoverlap', type='int', help='contigoverlap (-n)')
    parser.add_option('--mincontig', dest='mincontig', type='int', help='mincontig (-z)')
    parser.add_option('-T', dest='n_threads', type='int', help='number of threads to use in Bowtie (-T)')
    parser.add_option('-p', dest='prefix', default='sspace_pre', help='prefix (-b)')
    parser.add_option('--lib', dest='libraryname', default='galx', help='libraryfile')
    parser.add_option('--fe', dest='finalevidence', help='')
    parser.add_option('--fs', dest='finalscaffolds', help='')
    parser.add_option('--lg', dest='logfile', help='')
    parser.add_option('--summ', dest='summaryfile', help='')
    parser.add_option('--loglevel', choices=LOG_LEVELS, help='logging level (default: INFO)', default='DEBUG')
    (options, args) = parser.parse_args()
    if len(args) > 0:
        parser.error('Wrong number of arguments')
    
    log_level = getattr(logging, options.loglevel)
    kwargs = {'format' : LOG_FORMAT,
              'datefmt' : LOG_DATEFMT,
              'level' : log_level}
    if options.logfile:
        kwargs['filename'] = options.logfile
    logging.basicConfig(**kwargs)
    logger = logging.getLogger('SSPACE scaffold assembly')
    
    logger.debug('Creating temp dir')
    contigs = options.contigs
    reads1 = options.reads1
    reads2 = options.reads2
    insert = options.insert
    error = options.error
    orientation = options.orientation
    extension = '-x 1' if options.extension else ''
    minoverlap = "-m %d" % options.minoverlap if options.minoverlap is not None else ''
    numofreads = "-o %d" % options.numofreads if options.numofreads is not None else ''
    max_trim = "-t %d" % options.max_trim if options.max_trim is not None else ''
    unpaired = "-u %s" % options.unpaired if options.unpaired else ''
    min_base_ratio = "-r %s" % options.min_base_ratio if options.min_base_ratio is not None else ''
    minlink = "-k %d" % options.minlink if options.minlink is not None else ''
    maxratio = "-a %s" % options.maxratio if options.maxratio is not None else ''
    contigoverlap = "-n %d" % options.contigoverlap if options.contigoverlap is not None else ''
    mincontig = "-z %d" % options.mincontig if options.mincontig is not None else ''
    n_threads = "-T %d" % options.n_threads if options.n_threads is not None else ''
    prefix = options.prefix
    libraryname = options.libraryname
    finalevidence = options.finalevidence
    finalscaffolds = options.finalscaffolds
    summaryfile = options.summaryfile
    exe_name = "SSPACE_Basic_v2.0.pl"
    
    # Set library
    outline = 'lib1 %s %s %d %s %s\n' % (reads1, reads2, insert, error, orientation)
    with open(libraryname, 'w') as out:
        out.write(outline)
    
    exe = which(exe_name, os.R_OK)
    command = "perl %s -l %s -s %s %s %s %s %s %s %s %s %s %s %s %s -b %s" % (exe.pop(), libraryname, contigs, extension, minoverlap, numofreads, max_trim, unpaired, min_base_ratio, minlink, maxratio, contigoverlap, mincontig, n_threads, prefix)
    wd = tempfile.mkdtemp()
    try:
        os.chdir(wd)
        logger.info("SSPACE running")
        logger.debug("executing %s" % command)
        subprocess.check_call(args=command, shell=True)
        
        # Clean environment
        with open("%s.logfile.txt" % os.path.join(wd, prefix), 'rb') as sspace_log_file:
            logger.info("\n".join(["Log from SSPACE", sspace_log_file.read()]))
        logger.info("Moving result files")
        shutil.move("%s.final.evidence" % os.path.join(wd, prefix), finalevidence)
        shutil.move("%s.final.scaffolds.fasta" % os.path.join(wd, prefix), finalscaffolds)
        shutil.move("%s.summaryfile.txt" % os.path.join(wd, prefix), summaryfile)
        logger.debug("Removing temp dir: %s" % wd)
    finally:
        shutil.rmtree(wd)


if __name__ == "__main__":
    __main__()