def __init__(self, input_fasta, reads_per_split, out_dir, out_prefix):
self.input_fasta = input_fasta
self.out_dir = out_dir
self.reads_per_split = reads_per_split # Number of reads per split
self.out_prefix = out_prefix
self.out_fns = None
mkdir(self.out_dir)
def __init__(self, input_fasta, reads_per_split, out_dir, out_prefix):
self.input_fasta = input_fasta
self.out_dir = out_dir
self.reads_per_split = reads_per_split # Number of reads per split
self.out_prefix = out_prefix
self.out_fns = None
mkdir(self.out_dir)
def _validate_inputs(self, root_dir, nfl_fa, N):
"""
Check inputs, return
(num_reads,
number_reads_per_chunk,
nfl_dir,
[i-th_chunk_nfl_fa for i in [0...N-1]])
"""
icef = IceFiles(prog_name="ice_partial_split", root_dir=root_dir, no_log_f=False)
nfl_dir = icef.nfl_dir
# root_dir/output/map_noFL/input.split_{0:02d}.fa
splitted_nfl_fas = [icef.nfl_fa_i(i) for i in range(0, N)]
mkdir(icef.nfl_dir)
# Check if inputs exist.
errMsg = ""
if not nfs_exists(nfl_fa):
errMsg = "The input non-full-length reads fasta file " + "{f} does not exists. ".format(f=nfl_fa)
if len(errMsg) != 0:
raise ValueError(errMsg)
num_reads = num_reads_in_fasta(nfl_fa)
reads_per_split = int(max(1, ceil(num_reads / N)))
return (num_reads, reads_per_split, nfl_dir, splitted_nfl_fas)
def _validate_inputs(self, root_dir, nfl_fa, N):
"""
Check inputs, return
(num_reads,
number_reads_per_chunk,
nfl_dir,
[i-th_chunk_nfl_fa for i in [0...N-1]])
"""
icef = IceFiles(prog_name="ice_partial_split",
root_dir=root_dir,
no_log_f=False)
nfl_dir = icef.nfl_dir
# root_dir/output/map_noFL/input.split_{0:02d}.fa
splitted_nfl_fas = [icef.nfl_fa_i(i) for i in range(0, N)]
mkdir(icef.nfl_dir)
# Check if inputs exist.
errMsg = ""
if not nfs_exists(nfl_fa):
errMsg = ("The input non-full-length reads fasta file " +
"{f} does not exists. ".format(f=nfl_fa))
if len(errMsg) != 0:
raise ValueError(errMsg)
num_reads = num_reads_in_fasta(nfl_fa)
reads_per_split = int(max(1, ceil(num_reads / N)))
return (num_reads, reads_per_split, nfl_dir, splitted_nfl_fas)
def run(self):
"""Run quiver for ICE."""
# Create directories: root_dir/quivered and root_dir/log_dir/quivered
mkdir(self.quivered_dir)
mkdir(self.quivered_log_dir)
files = get_files_from_fofn(self.fasta_fofn)
msg = "Indexing {0} fasta files, please wait.".format(len(files))
self.add_log(msg)
d = MetaSubreadFastaReader(files)
self.add_log("Fasta files indexing done.")
self.add_log("Loading uc from {f}.".format(f=self.final_pickle_fn))
a = load(open(self.final_pickle_fn))
uc = a['uc']
refs = a['refs']
self.add_log("Loading partial uc from {f}.".
format(f=self.nfl_all_pickle_fn))
partial_uc = load(open(self.nfl_all_pickle_fn))['partial_uc']
partial_uc2 = defaultdict(lambda: [])
partial_uc2.update(partial_uc)
# Write report to quivered/cluster_report.FL_nonFL.csv
self.add_log("Writing a csv report of cluster -> FL/NonFL reads to {f}".
format(f=self.report_fn), level=logging.INFO)
self.write_report(uc=uc, partial_uc=partial_uc2,
report_fn=self.report_fn)
good = [x for x in uc] #[x for x in uc if len(uc[x]) > 1 or len(partial_uc2[x]) >= 10]
keys = sorted(good) # sort good keys (cluster ids)
start = 0
end = len(keys)
submitted = [] # submitted jobs
todo = [] # to-do jobs
self.submit_quiver_jobs(d=d, uc=uc, partial_uc=partial_uc2,
refs=refs, keys=keys, start=start, end=end,
submitted=submitted, todo=todo,
use_sge=self.sge_opts.use_sge,
max_sge_jobs=self.sge_opts.max_sge_jobs,
quiver_nproc=self.sge_opts.quiver_nproc)
with open(self.submitted_quiver_jobs_log, 'w') as f:
f.write("\n".join(str(x[0]) + '\t' + str(x[1]) for x in submitted))
self.close_log()
return 0
def validate_inputs(self):
"""Validate input fofns, and root_dir, log_dir, tmp_dir,
create quivered_dir and quivered_log_dir"""
self.add_log("Validating inputs.")
# Create directories: root_dir/quivered and root_dir/log_dir/quivered
try:
mkdir(self.quivered_dir)
mkdir(self.quivered_log_dir)
except OSError:
# Multiple ice_quiver_i jobs may run at the same time and try to
# mkdir, race condition may happen, so ignore OSError here.
pass
errMsg = ""
if not nfs_exists(self.log_dir) or not op.isdir(self.log_dir):
errMsg = "Log dir {l} is not an existing directory.".\
format(l=self.log_dir)
elif self.bas_fofn is None:
errMsg = "Please specify bas_fofn (e.g. input.fofn)."
elif not nfs_exists(self.bas_fofn):
errMsg = "bas_fofn {f} ".format(f=self.bas_fofn) + \
"which contains bas/bax.h5 files does not exist."
elif self.fasta_fofn is None:
errMsg = "Please make sure ice_make_fasta_fofn has " + \
"been called, and specify fasta_fofn."
elif not nfs_exists(self.fasta_fofn):
errMsg = "Input fasta_fofn {f} does not exists.".\
format(f=self.fasta_fofn)
fasta_files = get_files_from_fofn(self.fasta_fofn)
for fasta_file in fasta_files:
if not nfs_exists(fasta_file):
errMsg = "A file {f} in fasta_fofn does not exist.".\
format(f=fasta_file)
elif not nfs_exists(self.nfl_all_pickle_fn):
#"output/map_noFL/noFL.ALL.partial_uc.pickle"):
errMsg = "Pickle file {f} ".format(f=self.nfl_all_pickle_fn) + \
"which assigns all non-full-length reads to isoforms " + \
"does not exist. Please check 'ice_partial.py *' are " + \
"all done."
elif not nfs_exists(self.final_pickle_fn):
errMsg = "Pickle file {f} ".format(f=self.final_pickle_fn) + \
"which assigns full-length non-chimeric reads to " + \
"isoforms does not exist."
if errMsg != "":
self.add_log(errMsg, level=logging.ERROR)
raise IOError(errMsg)
def validate_inputs(self):
"""Validate input fofns, and root_dir, log_dir, tmp_dir,
create quivered_dir and quivered_log_dir"""
self.add_log("Validating inputs.")
# Create directories: root_dir/quivered and root_dir/log_dir/quivered
try:
mkdir(self.quivered_dir)
mkdir(self.quivered_log_dir)
except OSError:
# Multiple ice_quiver_i jobs may run at the same time and try to
# mkdir, race condition may happen, so ignore OSError here.
pass
errMsg = ""
if not nfs_exists(self.log_dir) or not op.isdir(self.log_dir):
errMsg = "Log dir {l} is not an existing directory.".\
format(l=self.log_dir)
elif self.bas_fofn is None:
errMsg = "Please specify bas_fofn (e.g. input.fofn)."
elif not nfs_exists(self.bas_fofn):
errMsg = "bas_fofn {f} ".format(f=self.bas_fofn) + \
"which contains bas/bax.h5 files does not exist."
elif self.fasta_fofn is None:
errMsg = "Please make sure ice_make_fasta_fofn has " + \
"been called, and specify fasta_fofn."
elif not nfs_exists(self.fasta_fofn):
errMsg = "Input fasta_fofn {f} does not exists.".\
format(f=self.fasta_fofn)
fasta_files = get_files_from_fofn(self.fasta_fofn)
for fasta_file in fasta_files:
if not nfs_exists(fasta_file):
errMsg = "A file {f} in fasta_fofn does not exist.".\
format(f=fasta_file)
elif not nfs_exists(self.nfl_all_pickle_fn):
#"output/map_noFL/noFL.ALL.partial_uc.pickle"):
errMsg = "Pickle file {f} ".format(f=self.nfl_all_pickle_fn) + \
"which assigns all non-full-length reads to isoforms " + \
"does not exist. Please check 'ice_partial.py *' are " + \
"all done."
elif not nfs_exists(self.final_pickle_fn):
errMsg = "Pickle file {f} ".format(f=self.final_pickle_fn) + \
"which assigns full-length non-chimeric reads to " + \
"isoforms does not exist."
if errMsg != "":
self.add_log(errMsg, level=logging.ERROR)
raise IOError(errMsg)
def __init__(self,
root_dir,
fasta_filenames,
ref_fasta,
out_pickle,
sge_opts,
sa_file=None,
ccs_fofn=None):
"""
fasta_filenames --- a list of splitted nfl fasta files.
ref_fasta --- (unpolished) consensus isoforms
out_pickle --- a pickle file with all nfl fasta reads
ccs_fofn --- should be reads_of_insert.fofn or None
root_dir --- ICE root output directory
sge_opts --- params for SGE environment, including
use_sge : use SGE or not
max_sge_jobs: maximum number of gcon jobs submitted
unique_id : unique qsub job id, important that this
DOES NOT CONFLICT!
blasr_nproc: blasr -nproc param, number of threads per cpu.
gcon_nproc : number of gcon that can run at the same time
"""
self.prog_name = "IceAllPartials"
IceFiles.__init__(self, prog_name=self.prog_name, root_dir=root_dir)
self.fasta_filenames, self.ref_fasta, self.ccs_fofn, self.sa_file = \
self._validateInputs(fasta_filenames=fasta_filenames,
ref_fasta=ref_fasta,
ccs_fofn=ccs_fofn,
sa_file=sa_file)
self.out_pickle = out_pickle
self.sge_opts = sge_opts
self.add_log("Making dir for mapping noFL reads: " + self.nfl_dir)
mkdir(self.nfl_dir)
self.add_log("input fasta files are: " +
", ".join(self.fasta_filenames))
self.add_log("temp pickle files are: " +
", ".join(self.pickle_filenames))
self.add_log("out pickle file is: " + self.out_pickle)
def __init__(self,
prog_name,
root_dir,
bas_fofn=None,
ccs_fofn=None,
fasta_fofn=None,
no_log_f=False):
"""
prog_name --- name of a sub-class
root_dir --- root directory of the whole project. There will be
sub-directories under it, including:
tmp/ --- 0/ c0, c1, ..., c9999
--- 1/ c10000, c10001, ..., c19999
...
each c? folder contains data for a cluster id=c?
script/
--- 0/ gcon_job_?.sh, gcon jobs in the first iteration
--- 1/ gcon_job_?.sh, gcon jobs in the second iteration
...
log/
--- ICE.log Log of the ICE algorithm
--- 0/ log for jobs in the first iteration
...
output/ output files go here.
bas_fofn --- input.fofn which contains movie.bas|bax.h5 files.
ccs_fofn --- a fofn contains movie.ccs.h5 files.
fasta_fofn --- a fofn contains movie.bax.h5.fasta files.
script/
no_log_f --- DON'T write log to a log file.
"""
self.prog_name = str(prog_name)
self.root_dir = real_ppath(root_dir)
self.bas_fofn = real_ppath(bas_fofn)
self.ccs_fofn = real_ppath(ccs_fofn)
self.fasta_fofn = real_ppath(fasta_fofn)
mkdir(self.root_dir)
mkdir(self.tmp_dir)
mkdir(self.log_dir)
mkdir(self.script_dir)
mkdir(self.out_dir)
self.no_log_f = no_log_f
if not no_log_f:
self.log_f = open(self.log_fn, 'w', 0)
self.add_log(msg="Initializing {p}.".format(p=self.prog_name))
def __init__(self, prog_name, root_dir,
bas_fofn=None, ccs_fofn=None, fasta_fofn=None,
no_log_f=False):
"""
prog_name --- name of a sub-class
root_dir --- root directory of the whole project. There will be
sub-directories under it, including:
tmp/ --- 0/ c0, c1, ..., c9999
--- 1/ c10000, c10001, ..., c19999
...
each c? folder contains data for a cluster id=c?
script/
--- 0/ gcon_job_?.sh, gcon jobs in the first iteration
--- 1/ gcon_job_?.sh, gcon jobs in the second iteration
...
log/
--- ICE.log Log of the ICE algorithm
--- 0/ log for jobs in the first iteration
...
output/ output files go here.
bas_fofn --- input.fofn which contains movie.bas|bax.h5 files.
ccs_fofn --- a fofn contains movie.ccs.h5 files.
fasta_fofn --- a fofn contains movie.bax.h5.fasta files.
script/
no_log_f --- DON'T write log to a log file.
"""
self.prog_name = str(prog_name)
self.root_dir = real_ppath(root_dir)
self.bas_fofn = real_ppath(bas_fofn)
self.ccs_fofn = real_ppath(ccs_fofn)
self.fasta_fofn = real_ppath(fasta_fofn)
mkdir(self.root_dir)
mkdir(self.tmp_dir)
mkdir(self.log_dir)
mkdir(self.script_dir)
mkdir(self.out_dir)
self.no_log_f = no_log_f
if not no_log_f:
self.log_f = open(self.log_fn, 'w', 0)
self.add_log(msg="Initializing {p}.".format(p=self.prog_name))
def __init__(self, root_dir, fastq_filenames, ref_fasta,
out_pickle, sge_opts, sa_file=None, ccs_fofn=None):
"""
fastq_filenames --- a list of splitted nfl fastq files.
ref_fasta --- (unpolished) consensus isoforms
out_pickle --- a pickle file with all nfl fasta reads
ccs_fofn --- should be reads_of_insert.fofn or None
root_dir --- ICE root output directory
sge_opts --- params for SGE environment, including
use_sge : use SGE or not
max_sge_jobs: maximum number of sub-jobs submitted
unique_id : unique qsub job id, important that this
DOES NOT CONFLICT!
blasr_nproc: blasr -nproc param, number of threads per cpu.
"""
self.prog_name = "IceAllPartials"
IceFiles.__init__(self, prog_name=self.prog_name, root_dir=root_dir)
self.add_log("DEBUG: in IceAllPartials, ccs_fofn is {0}.".format(ccs_fofn), level=logging.INFO)
self.fastq_filenames, self.ref_fasta, self.ccs_fofn, self.sa_file = \
self._validate_inputs(fastq_filenames=fastq_filenames,
ref_fasta=ref_fasta,
ccs_fofn=ccs_fofn,
sa_file=sa_file)
self.out_pickle = out_pickle
self.sge_opts = sge_opts
self.add_log("Making dir for mapping noFL reads: " + self.nfl_dir)
mkdir(self.nfl_dir)
self.add_log("input fasta files are: " +
", ".join(self.fastq_filenames))
self.add_log("temp pickle files are: " +
", ".join(self.pickle_filenames))
self.add_log("out pickle file is: " + self.out_pickle)
def convert_fofn_to_fasta(fofn_filename,
out_filename,
fasta_out_dir,
force_overwrite=False):
"""
For each .bax.h5 file, create .bax.h5.fasta file and save paths to
out_filename, which should usually be 'input.fasta.fofn'
"""
logging.info("Converting fofn {fofn} to fasta.".format(fofn=fofn_filename))
in_fns = get_files_from_fofn(fofn_filename)
out_fns = []
mkdir(fasta_out_dir)
for in_fn in in_fns:
logging.debug("converting h5 file: {f}.".format(f=in_fn))
if not (in_fn.endswith('.bax.h5') or in_fn.endswith('.bas.h5')):
raise ValueError("fofn file {fofn} ".format(fofn=fofn_filename) +
"should only contain bax/bas.h5 files.")
# e.g. m111xxxx.1.bax.h5 ==>
# tmp_out_file = m11xxxx.1.bax.h5.fasta.tmp
# out_file = m11xxxx.1.bax.h5.fasta
in_basename = op.basename(in_fn)
tmp_out_file = op.join(fasta_out_dir, in_basename + '.fasta.tmp')
out_file = op.join(fasta_out_dir, in_basename + '.fasta')
if op.exists(out_file) and not force_overwrite:
logging.debug(
"File {0} already exists. skipping.".format(out_file))
else:
cmd = "pls2fasta {0} {1} ".format(in_fn, tmp_out_file) + \
"-minSubreadLength 300 -minReadScore 750 -trimByRegion"
logging.debug("CMD: {cmd}".format(cmd=cmd))
_out, _code, _msg = backticks(cmd)
if _code != 0:
raise RuntimeError("CMD failed: {cmd}\n".format(cmd=cmd) +
_msg)
trim_subread_flanks(tmp_out_file, out_file)
out_fns.append(out_file)
if op.exists(tmp_out_file):
os.remove(tmp_out_file)
write_files_to_fofn(out_fns, out_filename)
def convert_fofn_to_fasta(fofn_filename, out_filename, fasta_out_dir,
force_overwrite=False):
"""
For each .bax.h5 file, create .bax.h5.fasta file and save paths to
out_filename, which should usually be 'input.fasta.fofn'
"""
logging.info("Converting fofn {fofn} to fasta.".format(fofn=fofn_filename))
in_fns = get_files_from_fofn(fofn_filename)
out_fns = []
mkdir(fasta_out_dir)
for in_fn in in_fns:
logging.debug("converting h5 file: {f}.".format(f=in_fn))
if not (in_fn.endswith('.bax.h5') or in_fn.endswith('.bas.h5')):
raise ValueError("fofn file {fofn} ".format(fofn=fofn_filename) +
"should only contain bax/bas.h5 files.")
# e.g. m111xxxx.1.bax.h5 ==>
# tmp_out_file = m11xxxx.1.bax.h5.fasta.tmp
# out_file = m11xxxx.1.bax.h5.fasta
in_basename = op.basename(in_fn)
tmp_out_file = op.join(fasta_out_dir, in_basename + '.fasta.tmp')
out_file = op.join(fasta_out_dir, in_basename + '.fasta')
if op.exists(out_file) and not force_overwrite:
logging.debug("File {0} already exists. skipping.".format(out_file))
else:
cmd = "pls2fasta {0} {1} ".format(in_fn, tmp_out_file) + \
"-minSubreadLength 300 -minReadScore 750 -trimByRegion"
logging.debug("CMD: {cmd}".format(cmd=cmd))
_out, _code, _msg = backticks(cmd)
if _code != 0:
raise RuntimeError("CMD failed: {cmd}\n".format(cmd=cmd) + _msg)
trim_subread_flanks(tmp_out_file, out_file)
out_fns.append(out_file)
if op.exists(tmp_out_file):
os.remove(tmp_out_file)
write_files_to_fofn(out_fns, out_filename)
def convert_fofn_to_fasta(fofn_filename, out_filename, fasta_out_dir,
force_overwrite=False, cpus=1):
"""
For each .bax.h5 file, create .bax.h5.fasta file and save paths to
out_filename, which should usually be 'input.fasta.fofn'
"""
logging.info("Converting fofn {fofn} to fasta.".format(fofn=fofn_filename))
in_fns = get_files_from_fofn(fofn_filename)
#out_fns = []
mkdir(fasta_out_dir)
# multiprocessing worker stuff
manager = Manager()
in_queue = manager.Queue(len(in_fns))
in_queue_count = 0
outfile_track = {} # expected out file --> (cmd, tmp)
pool = []
out_fns = []
for in_fn in in_fns:
#print >> sys.stderr, "DEBUG: converting h5 file:", in_fn
logging.debug("converting h5 file: {f}.".format(f=in_fn))
if not (in_fn.endswith('.bax.h5') or in_fn.endswith('.bas.h5')):
raise ValueError("fofn file {fofn} ".format(fofn=fofn_filename) +
"should only contain bax/bas.h5 files.")
# e.g. m111xxxx.1.bax.h5 ==>
# tmp_out_file = m11xxxx.1.bax.h5.fasta.tmp
# out_file = m11xxxx.1.bax.h5.fasta
in_basename = op.basename(in_fn)
tmp_out_file = op.join(fasta_out_dir, in_basename + '.fasta.tmp')
out_file = op.join(fasta_out_dir, in_basename + '.fasta')
if op.exists(out_file) and not force_overwrite:
logging.debug("File {0} already exists. skipping.".format(out_file))
out_fns.append(out_file)
if op.exists(tmp_out_file):
os.remove(tmp_out_file)
else:
cmd = "pls2fasta {in_fn} ".format(in_fn=real_upath(in_fn)) + \
" {out} ".format(out=real_upath(tmp_out_file)) + \
"-minSubreadLength 300 -minReadScore 750 -trimByRegion"
print >> sys.stderr, "DEBUG: putting in queue:", cmd, tmp_out_file, out_file
in_queue.put((cmd, tmp_out_file, out_file))
in_queue_count += 1
outfile_track[out_file] = (cmd, tmp_out_file)
print >> sys.stderr, "DEBUG: put in queue:", cmd, tmp_out_file, out_file
cpus = min(cpus, in_queue_count) # cap max CPU if there's fewer files to convert
for i in xrange(cpus):
p = Process(target=convert_fofn_to_fasta_worker, args=(in_queue,))
pool.append(p)
#error_flag = False
# starting & joining pool worakers
for p in pool:
p.start()
#print >> sys.stderr, "Starting worker", p.name
for p in pool:
#print >> sys.stderr, "Waiting join", p.name
p.join(timeout=1200)
if p.is_alive(): p.terminate()
# check that all files exists
# if it does not, force to run locally
for out_file,(cmd, tmp_out_file) in outfile_track.iteritems():
in_queue.put((cmd, tmp_out_file, out_file))
convert_fofn_to_fasta_worker(in_queue)
out_fns.append(out_file)
#if error_flag:
# raise Exception, "Unable to successfuly run convert_fofn_to_fasta, ABORT!"
write_files_to_fofn(out_fns, out_filename)
def run(self):
"""Run quiver for ICE."""
# Create directories: root_dir/quivered and root_dir/log_dir/quivered
mkdir(self.quivered_dir)
mkdir(self.quivered_log_dir)
files = get_files_from_fofn(self.fasta_fofn)
msg = "Indexing {0} fasta files, please wait.".format(len(files))
self.add_log(msg)
d = MetaSubreadFastaReader(files)
self.add_log("Fasta files indexing done.")
self.add_log("Loading uc from {f}.".format(f=self.final_pickle_fn))
a = load(open(self.final_pickle_fn))
uc = a['uc']
refs = a['refs']
self.add_log(
"Loading partial uc from {f}.".format(f=self.nfl_all_pickle_fn))
partial_uc = load(open(self.nfl_all_pickle_fn))['partial_uc']
partial_uc2 = defaultdict(lambda: [])
partial_uc2.update(partial_uc)
# Write report to quivered/cluster_report.FL_nonFL.csv
self.add_log(
"Writing a csv report of cluster -> FL/NonFL reads to {f}".format(
f=self.report_fn),
level=logging.INFO)
self.write_report(uc=uc,
partial_uc=partial_uc2,
report_fn=self.report_fn)
good = [
x for x in uc
] #[x for x in uc if len(uc[x]) > 1 or len(partial_uc2[x]) >= 10]
keys = sorted(good) # sort good keys (cluster ids)
start = 0
end = len(keys)
submitted = [] # submitted jobs
todo = [] # to-do jobs
self.submit_quiver_jobs(d=d,
uc=uc,
partial_uc=partial_uc2,
refs=refs,
keys=keys,
start=start,
end=end,
submitted=submitted,
todo=todo,
use_sge=self.sge_opts.use_sge,
max_sge_jobs=self.sge_opts.max_sge_jobs,
quiver_nproc=self.sge_opts.quiver_nproc)
with open(self.submitted_quiver_jobs_log, 'w') as f:
f.write("\n".join(str(x[0]) + '\t' + str(x[1]) for x in submitted))
self.close_log()
return 0
def convert_fofn_to_fasta(fofn_filename, out_filename, fasta_out_dir,
force_overwrite=False, cpus=1):
"""
For each .bax.h5 file, create .bax.h5.fasta file and save paths to
out_filename, which should usually be 'input.fasta.fofn'
"""
logging.info("Converting fofn {fofn} to fasta.".format(fofn=fofn_filename))
in_fns = get_files_from_fofn(fofn_filename)
#out_fns = []
mkdir(fasta_out_dir)
# multiprocessing worker stuff
manager = Manager()
in_queue = manager.Queue(len(in_fns))
in_queue_count = 0
outfile_track = {} # expected out file --> (cmd, tmp)
pool = []
out_fns = []
for in_fn in in_fns:
#print >> sys.stderr, "DEBUG: converting h5 file:", in_fn
logging.debug("converting h5 file: {f}.".format(f=in_fn))
if not (in_fn.endswith('.bax.h5') or in_fn.endswith('.bas.h5')):
raise ValueError("fofn file {fofn} ".format(fofn=fofn_filename) +
"should only contain bax/bas.h5 files.")
# e.g. m111xxxx.1.bax.h5 ==>
# tmp_out_file = m11xxxx.1.bax.h5.fasta.tmp
# out_file = m11xxxx.1.bax.h5.fasta
in_basename = op.basename(in_fn)
tmp_out_file = op.join(fasta_out_dir, in_basename + '.fasta.tmp')
out_file = op.join(fasta_out_dir, in_basename + '.fasta')
if op.exists(out_file) and not force_overwrite:
logging.debug("File {0} already exists. skipping.".format(out_file))
out_fns.append(out_file)
if op.exists(tmp_out_file):
os.remove(tmp_out_file)
else:
cmd = "pls2fasta {in_fn} ".format(in_fn=real_upath(in_fn)) + \
" {out} ".format(out=real_upath(tmp_out_file)) + \
"-minSubreadLength 300 -minReadScore 750 -trimByRegion"
print >> sys.stderr, "DEBUG: putting in queue:", cmd, tmp_out_file, out_file
in_queue.put((cmd, tmp_out_file, out_file))
in_queue_count += 1
outfile_track[out_file] = (cmd, tmp_out_file)
print >> sys.stderr, "DEBUG: put in queue:", cmd, tmp_out_file, out_file
cpus = min(cpus, in_queue_count) # cap max CPU if there's fewer files to convert
for i in xrange(cpus):
p = Process(target=convert_fofn_to_fasta_worker, args=(in_queue,))
pool.append(p)
#error_flag = False
# starting & joining pool worakers
for p in pool:
p.start()
#print >> sys.stderr, "Starting worker", p.name
for p in pool:
#print >> sys.stderr, "Waiting join", p.name
p.join(timeout=1200)
if p.is_alive(): p.terminate()
# check that all files exists
# if it does not, force to run locally
for out_file,(cmd, tmp_out_file) in outfile_track.iteritems():
in_queue.put((cmd, tmp_out_file, out_file))
convert_fofn_to_fasta_worker(in_queue)
out_fns.append(out_file)
#if error_flag:
# raise Exception, "Unable to successfuly run convert_fofn_to_fasta, ABORT!"
write_files_to_fofn(out_fns, out_filename)
def convert_fofn_to_fasta(fofn_filename,
out_filename,
fasta_out_dir,
force_overwrite=False,
cpus=1):
"""
For each .bax.h5 file, create .bax.h5.fasta file and save paths to
out_filename, which should usually be 'input.fasta.fofn'
"""
logging.info("Converting fofn {fofn} to fasta.".format(fofn=fofn_filename))
in_fns = get_files_from_fofn(fofn_filename)
#out_fns = []
mkdir(fasta_out_dir)
# multiprocessing worker stuff
manager = Manager()
out_fns = manager.list()
in_queue = manager.Queue(99999)
pool = []
for i in xrange(cpus):
p = Process(target=convert_fofn_to_fasta_worker,
args=(in_queue, out_fns))
pool.append(p)
for in_fn in in_fns:
#print >> sys.stderr, "DEBUG: converting h5 file:", in_fn
logging.debug("converting h5 file: {f}.".format(f=in_fn))
if not (in_fn.endswith('.bax.h5') or in_fn.endswith('.bas.h5')):
raise ValueError("fofn file {fofn} ".format(fofn=fofn_filename) +
"should only contain bax/bas.h5 files.")
# e.g. m111xxxx.1.bax.h5 ==>
# tmp_out_file = m11xxxx.1.bax.h5.fasta.tmp
# out_file = m11xxxx.1.bax.h5.fasta
in_basename = op.basename(in_fn)
tmp_out_file = op.join(fasta_out_dir, in_basename + '.fasta.tmp')
out_file = op.join(fasta_out_dir, in_basename + '.fasta')
if op.exists(out_file) and not force_overwrite:
logging.debug(
"File {0} already exists. skipping.".format(out_file))
out_fns.append(out_file)
if op.exists(tmp_out_file):
os.remove(tmp_out_file)
else:
cmd = "pls2fasta {in_fn} ".format(in_fn=real_upath(in_fn)) + \
" {out} ".format(out=real_upath(tmp_out_file)) + \
"-minSubreadLength 300 -minReadScore 750 -trimByRegion"
print >> sys.stderr, "DEBUG: putting in queue:", cmd, tmp_out_file, out_file
in_queue.put((cmd, tmp_out_file, out_file))
print >> sys.stderr, "DEBUG: put in queue:", cmd, tmp_out_file, out_file
# logging.debug("CMD: {cmd}".format(cmd=cmd))
# _out, _code, _msg = backticks(cmd)
# if _code != 0:
# raise RuntimeError("CMD failed: {cmd}\n".format(cmd=cmd) + _msg)
# trim_subread_flanks(tmp_out_file, out_file)
# out_fns.append(out_file)
# if op.exists(tmp_out_file):
# os.remove(tmp_out_file)
# starting & joining pool worakers
for p in pool:
p.start()
#print >> sys.stderr, "Starting worker", p.name
for p in pool:
#print >> sys.stderr, "Waiting join", p.name
p.join()
write_files_to_fofn(out_fns, out_filename)
