def test_load_gz():
inpath = utils.get_test_data('random-20-a.fa')
savepath = utils.get_temp_filename('tempcountingsave1.ht')
loadpath = utils.get_temp_filename('tempcountingsave1.ht.gz')
sizes = list(PRIMES_1m)
sizes.append(1000005)
# save uncompressed hashtable.
hi = khmer._new_counting_hash(12, sizes)
hi.consume_fasta(inpath)
hi.save(savepath)
# compress.
in_file = open(savepath, 'rb')
out_file = gzip.open(loadpath, 'wb')
out_file.writelines(in_file)
out_file.close()
in_file.close()
# load compressed hashtable.
ht = khmer._new_counting_hash(12, sizes)
ht.load(loadpath)
tracking = khmer._new_hashbits(12, sizes)
x = hi.abundance_distribution(inpath, tracking)
tracking = khmer._new_hashbits(12, sizes)
y = ht.abundance_distribution(inpath, tracking)
assert sum(x) == 3966, sum(x)
assert x == y, (x, y)
def test_load_gz():
inpath = utils.get_test_data('random-20-a.fa')
savepath = utils.get_temp_filename('tempcountingsave1.ht')
loadpath = utils.get_temp_filename('tempcountingsave1.ht.gz')
sizes = list(PRIMES_1m)
sizes.append(1000005)
# save uncompressed hashtable.
hi = khmer._new_counting_hash(12, sizes)
hi.consume_fasta(inpath)
hi.save(savepath)
# compress.
in_file = open(savepath, 'rb')
out_file = gzip.open(loadpath, 'wb')
out_file.writelines(in_file)
out_file.close()
in_file.close()
# load compressed hashtable.
ht = khmer._new_counting_hash(12, sizes)
ht.load(loadpath)
tracking = khmer._new_hashbits(12, sizes)
x = hi.abundance_distribution(inpath, tracking)
tracking = khmer._new_hashbits(12, sizes)
y = ht.abundance_distribution(inpath, tracking)
assert sum(x) == 3966, sum(x)
assert x == y, (x, y)
def main():
info('abundance-dist.py', ['counting'])
args = get_parser().parse_args()
infiles = [
args.input_counting_table_filename, args.input_sequence_filename
]
for infile in infiles:
check_file_status(infile)
check_space(infiles)
print('hashtable from', args.input_counting_table_filename)
counting_hash = khmer.load_counting_hash(
args.input_counting_table_filename)
kmer_size = counting_hash.ksize()
hashsizes = counting_hash.hashsizes()
tracking = khmer._new_hashbits( # pylint: disable=protected-access
kmer_size, hashsizes)
print('K:', kmer_size)
print('HT sizes:', hashsizes)
print('outputting to', args.output_histogram_filename)
if os.path.exists(args.output_histogram_filename):
if not args.squash_output:
print('ERROR: %s exists; not squashing.' %
args.output_histogram_filename,
file=sys.stderr)
sys.exit(1)
print('** squashing existing file %s' % args.output_histogram_filename)
print('preparing hist...')
abundances = counting_hash.abundance_distribution(
args.input_sequence_filename, tracking)
total = sum(abundances)
if 0 == total:
print(
"ERROR: abundance distribution is uniformly zero; "
"nothing to report.",
file=sys.stderr)
print("\tPlease verify that the input files are valid.",
file=sys.stderr)
sys.exit(1)
hash_fp = open(args.output_histogram_filename, 'w')
sofar = 0
for _, i in enumerate(abundances):
if i == 0 and not args.output_zero:
continue
sofar += i
frac = sofar / float(total)
print(_, i, sofar, round(frac, 3), file=hash_fp)
if sofar == total:
break
def main():
info('abundance-dist.py', ['counting'])
args = get_parser().parse_args()
infiles = [args.input_counting_table_filename,
args.input_sequence_filename]
for infile in infiles:
check_file_status(infile)
check_space(infiles)
print('hashtable from', args.input_counting_table_filename)
counting_hash = khmer.load_counting_hash(
args.input_counting_table_filename)
kmer_size = counting_hash.ksize()
hashsizes = counting_hash.hashsizes()
tracking = khmer._new_hashbits( # pylint: disable=protected-access
kmer_size, hashsizes)
print('K:', kmer_size)
print('HT sizes:', hashsizes)
print('outputting to', args.output_histogram_filename)
if os.path.exists(args.output_histogram_filename):
if not args.squash_output:
print('ERROR: %s exists; not squashing.' %
args.output_histogram_filename,
file=sys.stderr)
sys.exit(1)
print('** squashing existing file %s' % args.output_histogram_filename)
print('preparing hist...')
abundances = counting_hash.abundance_distribution(
args.input_sequence_filename, tracking)
total = sum(abundances)
if 0 == total:
print("ERROR: abundance distribution is uniformly zero; "
"nothing to report.", file=sys.stderr)
print("\tPlease verify that the input files are valid.",
file=sys.stderr)
sys.exit(1)
hash_fp = open(args.output_histogram_filename, 'w')
sofar = 0
for _, i in enumerate(abundances):
if i == 0 and not args.output_zero:
continue
sofar += i
frac = sofar / float(total)
print(_, i, sofar, round(frac, 3), file=hash_fp)
if sofar == total:
break
def main():
parser = argparse.ArgumentParser(description="Output k-mer abundance distribution.")
parser.add_argument('hashname')
parser.add_argument('datafile')
parser.add_argument('histout')
parser.add_argument('-z', '--no-zero', dest='output_zero', default=True,
action='store_false',
help='Do not output 0-count bins')
parser.add_argument('-s', '--squash', dest='squash_output', default=False,
action='store_true',
help='Overwrite output file if it exists')
args = parser.parse_args()
hashfile = args.hashname
datafile = args.datafile
histout = args.histout
print 'hashtable from', hashfile
ht = khmer.load_counting_hash(hashfile)
K = ht.ksize()
sizes = ht.hashsizes()
tracking = khmer._new_hashbits(K, sizes)
print 'K:', K
print 'HT sizes:', sizes
print 'outputting to', histout
if os.path.exists(histout):
if not args.squash_output:
print >>sys.stderr, 'ERROR: %s exists; not squashing.' % histout
sys.exit(-1)
print '** squashing existing file %s' % histout
print 'preparing hist...'
z = ht.abundance_distribution(datafile, tracking)
total = sum(z)
fp = open(histout, 'w')
sofar = 0
for n, i in enumerate(z):
if i == 0 and not args.output_zero:
continue
sofar += i
frac = sofar / float(total)
print >>fp, n, i, sofar, round(frac, 3)
if sofar == total:
break
def test_save_load_gz():
inpath = utils.get_test_data('random-20-a.fa')
savepath = utils.get_temp_filename('tempcountingsave2.ht.gz')
sizes = list(PRIMES_1m)
sizes.append(1000005)
hi = khmer._new_counting_hash(12, sizes)
hi.consume_fasta(inpath)
hi.save(savepath)
ht = khmer._new_counting_hash(12, sizes)
ht.load(savepath)
tracking = khmer._new_hashbits(12, sizes)
x = hi.abundance_distribution(inpath, tracking)
tracking = khmer._new_hashbits(12, sizes)
y = ht.abundance_distribution(inpath, tracking)
assert sum(x) == 3966, sum(x)
assert x == y, (x, y)
def test_save_load_gz():
inpath = utils.get_test_data('random-20-a.fa')
savepath = utils.get_temp_filename('tempcountingsave2.ht.gz')
sizes = list(PRIMES_1m)
sizes.append(1000005)
hi = khmer._new_counting_hash(12, sizes)
hi.consume_fasta(inpath)
hi.save(savepath)
ht = khmer._new_counting_hash(12, sizes)
ht.load(savepath)
tracking = khmer._new_hashbits(12, sizes)
x = hi.abundance_distribution(inpath, tracking)
tracking = khmer._new_hashbits(12, sizes)
y = ht.abundance_distribution(inpath, tracking)
assert sum(x) == 3966, sum(x)
assert x == y, (x, y)
def test_save_load_gz():
thisdir = os.path.dirname(__file__)
inpath = os.path.join(thisdir, 'test-data/random-20-a.fa')
savepath = os.path.join(thisdir, 'tempcountingsave2.ht.gz')
sizes = list(PRIMES_1m)
sizes.append(1000005)
hi = khmer._new_counting_hash(12, sizes)
hi.consume_fasta(inpath)
hi.save(savepath)
ht = khmer._new_counting_hash(12, sizes)
ht.load(savepath)
tracking = khmer._new_hashbits(12, sizes)
x = hi.abundance_distribution(inpath, tracking)
tracking = khmer._new_hashbits(12, sizes)
y = ht.abundance_distribution(inpath, tracking)
assert sum(x) == 3966, sum(x)
assert x == y, (x,y)
def main():
parser = build_construct_args(
"Output k-mer abundance distribution (single file version).")
add_threading_args(parser)
parser.add_argument('datafile')
parser.add_argument('histout')
parser.add_argument('-z', '--no-zero', dest='output_zero', default=True,
action='store_false',
help='Do not output 0-count bins')
parser.add_argument('-b', '--no-bigcount', dest='bigcount', default=True,
action='store_false',
help='Do not count k-mers past 255')
parser.add_argument('-s', '--squash', dest='squash_output', default=False,
action='store_true',
help='Overwrite output file if it exists')
parser.add_argument('--savehash', dest='savehash', default='')
args = parser.parse_args()
report_on_config(args)
K = args.ksize
HT_SIZE = args.min_hashsize
N_HT = args.n_hashes
n_threads = int(args.n_threads)
datafile = args.datafile
histout = args.histout
print 'making hashtable'
ht = khmer.new_counting_hash(K, HT_SIZE, N_HT, n_threads)
ht.set_use_bigcount(args.bigcount)
print 'building tracking ht'
K = ht.ksize()
sizes = ht.hashsizes()
tracking = khmer._new_hashbits(K, sizes)
print 'K:', K
print 'HT sizes:', sizes
print 'outputting to', histout
config = khmer.get_config()
config.set_reads_input_buffer_size(n_threads * 64 * 1024)
# start loading
rparser = khmer.ReadParser(datafile, n_threads)
threads = []
print 'consuming input, round 1 --', datafile
for tnum in xrange(n_threads):
t = \
threading.Thread(
target=ht.consume_fasta_with_reads_parser,
args=(rparser, )
)
threads.append(t)
t.start()
for t in threads:
t.join()
z_list = []
def do_abundance_dist(r):
z = ht.abundance_distribution_with_reads_parser(r, tracking)
z_list.append(z)
print 'preparing hist from %s...' % datafile
rparser = khmer.ReadParser(datafile, n_threads)
threads = []
print 'consuming input, round 2 --', datafile
for tnum in xrange(n_threads):
t = \
threading.Thread(
target=do_abundance_dist,
args=(rparser,)
)
threads.append(t)
t.start()
for t in threads:
t.join()
assert len(z_list) == n_threads, len(z_list)
z = {}
for zz in z_list:
for i, count in enumerate(zz):
z[i] = z.get(i, 0) + count
total = sum(z.values())
if 0 == total:
print >>sys.stderr, \
"ERROR: abundance distribution is uniformly zero; " \
"nothing to report."
print >>sys.stderr, "\tPlease verify that the input files are valid."
sys.exit(-1)
fp = open(histout, 'w')
sofar = 0
for n, i in sorted(z.items()):
if i == 0 and not args.output_zero:
continue
sofar += i
frac = sofar / float(total)
print >>fp, n, i, sofar, round(frac, 3)
if sofar == total:
break
if args.savehash:
print 'Saving hashfile', args.savehash
print '...saving to', args.savehash
ht.save(args.savehash)
def test_bad_primes_list():
try:
coutingtable = khmer._new_hashbits(31, ["a", "b", "c"], 1)
assert 0, "Bad primes list should fail"
except TypeError, e:
print str(e)
def test_bad_primes_list():
try:
coutingtable = khmer._new_hashbits(31, ["a", "b", "c"], 1)
assert 0, "Bad primes list should fail"
except TypeError as e:
print str(e)
def main():
parser = argparse.ArgumentParser(
description="Output k-mer abundance distribution.")
parser.add_argument('hashname')
parser.add_argument('datafile')
parser.add_argument('histout')
parser.add_argument('-z',
'--no-zero',
dest='output_zero',
default=True,
action='store_false',
help='Do not output 0-count bins')
parser.add_argument('-s',
'--squash',
dest='squash_output',
default=False,
action='store_true',
help='Overwrite output file if it exists')
args = parser.parse_args()
hashfile = args.hashname
datafile = args.datafile
histout = args.histout
print 'hashtable from', hashfile
ht = khmer.load_counting_hash(hashfile)
K = ht.ksize()
sizes = ht.hashsizes()
tracking = khmer._new_hashbits(K, sizes)
print 'K:', K
print 'HT sizes:', sizes
print 'outputting to', histout
if os.path.exists(histout):
if not args.squash_output:
print >> sys.stderr, 'ERROR: %s exists; not squashing.' % histout
sys.exit(-1)
print '** squashing existing file %s' % histout
print 'preparing hist...'
z = ht.abundance_distribution(datafile, tracking)
total = sum(z)
if 0 == total:
print >> sys.stderr, "ERROR: abundance distribution is uniformly zero; nothing to report."
print >> sys.stderr, "\tPlease verify that the input files are valid."
sys.exit(-1)
fp = open(histout, 'w')
sofar = 0
for n, i in enumerate(z):
if i == 0 and not args.output_zero:
continue
sofar += i
frac = sofar / float(total)
print >> fp, n, i, sofar, round(frac, 3)
if sofar == total:
break
def main(filename):
global ht
n = 5
basename = os.path.basename(filename)
fd = open("log.txt", "w")
primes = []
below = khmer.get_n_primes_near_x(N_HT * n, HASHTABLE_SIZE)
above = khmer.get_n_primes_above_x(N_HT * n, HASHTABLE_SIZE)
primes = below + above
random.shuffle(primes)
for run in range(n):
print primes[run * N_HT:run * N_HT + N_HT]
ht = khmer._new_hashbits(K, primes[run * N_HT:run * N_HT + N_HT])
#ht = khmer.new_hashbits(K, HASHTABLE_SIZE, N_HT)
# populate the hash table and tag set
if not load_ht:
ht.consume_fasta_and_tag(filename)
# save to a file (optional)
if save_ht:
ht.save(basename + '.ht')
ht.save_tagset(basename + '.tagset')
# calculate the hashtable occupancy
print '---'
print 'hashtable occupancy:', ht.n_occupied() / float(
HASHTABLE_SIZE)
print '---'
else:
ht.load(basename + '.ht')
ht.load_tagset(basename + '.tagset')
# did we just want to load the ht/tagset?
if stop_after_n_subsets == 0:
sys.exit(0)
#stop_tags = pickle.load(open(sys.argv[2]))
#for stop_tag in stop_tags:
# ht.add_stop_tag(stop_tag)
# divide the tags up into subsets
divvy = ht.divide_tags_into_subsets(SUBSET_SIZE)
n_subsets = len(divvy)
divvy.append(0)
# build a queue of tasks:
worker_q = Queue.Queue()
for i in range(0, n_subsets):
if stop_after_n_subsets is not None and i >= stop_after_n_subsets:
break
start = divvy[i]
end = divvy[i + 1]
worker_q.put((ht, i, start, end))
open('%s.info' % basename,
'w').write('%d subsets total\n' % (n_subsets))
threads = []
for th in range(N_THREADS):
t = threading.Thread(target=worker, args=(worker_q, basename))
threads.append(t)
t.start()
# wait for threads
for t in threads:
t.join()
###
del ht
gc.collect()
# create a new, empty ht object for merging; K matters, but not
# hashtable size.
ht = khmer.new_hashbits(K, 1, 1)
# load & merge all pmap files
for i in range(0, n_subsets):
pmap_file = basename + '.subset.%d.pmap' % (i, )
ht.merge_subset_from_disk(pmap_file)
# save merged partitionmap
if save_merged_pmap:
ht.save_partitionmap(basename + '.pmap.merged')
if remove_orig_pmap:
for i in range(0, n_subsets):
pmap_file = basename + '.subset.%d.pmap' % (i, )
os.unlink(pmap_file)
# output partitions!
n_partitions = ht.output_partitions(filename, basename + '.part')
(n_partitions, n_singletons) = ht.count_partitions()
print n_partitions
fd.write(str(n_partitions) + "\n")
#print os.listdir(os.getcwd())
for file in glob.glob(os.getcwd() + "/*pmap*"):
os.remove(file)
for file in glob.glob(os.getcwd() + "/*.info"):
os.remove(file)
for file in glob.glob(os.getcwd() + "/*.part"):
os.remove(file)
fd.close()
def main():
parser = build_construct_args(
"Output k-mer abundance distribution (single file version).")
add_threading_args(parser)
parser.add_argument('datafile')
parser.add_argument('histout')
parser.add_argument('-z',
'--no-zero',
dest='output_zero',
default=True,
action='store_false',
help='Do not output 0-count bins')
parser.add_argument('-b',
'--no-bigcount',
dest='bigcount',
default=True,
action='store_false',
help='Do not count k-mers past 255')
parser.add_argument('-s',
'--squash',
dest='squash_output',
default=False,
action='store_true',
help='Overwrite output file if it exists')
parser.add_argument('--savehash', dest='savehash', default='')
args = parser.parse_args()
report_on_config(args)
K = args.ksize
HT_SIZE = args.min_hashsize
N_HT = args.n_hashes
n_threads = int(args.n_threads)
datafile = args.datafile
histout = args.histout
print 'making hashtable'
ht = khmer.new_counting_hash(K, HT_SIZE, N_HT, n_threads)
ht.set_use_bigcount(args.bigcount)
print 'building tracking ht'
K = ht.ksize()
sizes = ht.hashsizes()
tracking = khmer._new_hashbits(K, sizes)
print 'K:', K
print 'HT sizes:', sizes
print 'outputting to', histout
config = khmer.get_config()
config.set_reads_input_buffer_size(n_threads * 64 * 1024)
# start loading
rparser = khmer.ReadParser(datafile, n_threads)
threads = []
print 'consuming input, round 1 --', datafile
for tnum in xrange(n_threads):
t = \
threading.Thread(
target=ht.consume_fasta_with_reads_parser,
args=(rparser, )
)
threads.append(t)
t.start()
for t in threads:
t.join()
z_list = []
def do_abundance_dist(r):
z = ht.abundance_distribution_with_reads_parser(r, tracking)
z_list.append(z)
print 'preparing hist from %s...' % datafile
rparser = khmer.ReadParser(datafile, n_threads)
threads = []
print 'consuming input, round 2 --', datafile
for tnum in xrange(n_threads):
t = \
threading.Thread(
target=do_abundance_dist,
args=(rparser,)
)
threads.append(t)
t.start()
for t in threads:
t.join()
assert len(z_list) == n_threads, len(z_list)
z = {}
for zz in z_list:
for i, count in enumerate(zz):
z[i] = z.get(i, 0) + count
total = sum(z.values())
if 0 == total:
print >>sys.stderr, \
"ERROR: abundance distribution is uniformly zero; " \
"nothing to report."
print >> sys.stderr, "\tPlease verify that the input files are valid."
sys.exit(-1)
fp = open(histout, 'w')
sofar = 0
for n, i in sorted(z.items()):
if i == 0 and not args.output_zero:
continue
sofar += i
frac = sofar / float(total)
print >> fp, n, i, sofar, round(frac, 3)
if sofar == total:
break
if args.savehash:
print 'Saving hashfile', args.savehash
print '...saving to', args.savehash
ht.save(args.savehash)
def main(filename):
global ht
n = 5
basename = os.path.basename(filename)
fd = open("log.txt", "w")
primes = []
below = khmer.get_n_primes_near_x(N_HT * n, HASHTABLE_SIZE)
above = khmer.get_n_primes_above_x(N_HT * n, HASHTABLE_SIZE)
primes = below + above
random.shuffle(primes)
for run in range(n):
print primes[run*N_HT:run*N_HT+N_HT]
ht = khmer._new_hashbits(K, primes[run*N_HT:run*N_HT+N_HT])
#ht = khmer.new_hashbits(K, HASHTABLE_SIZE, N_HT)
# populate the hash table and tag set
if not load_ht:
ht.consume_fasta_and_tag(filename)
# save to a file (optional)
if save_ht:
ht.save(basename + '.ht')
ht.save_tagset(basename + '.tagset')
# calculate the hashtable occupancy
print '---'
print 'hashtable occupancy:', ht.n_occupied() / float(HASHTABLE_SIZE)
print '---'
else:
ht.load(basename + '.ht')
ht.load_tagset(basename + '.tagset')
# did we just want to load the ht/tagset?
if stop_after_n_subsets == 0:
sys.exit(0)
#stop_tags = pickle.load(open(sys.argv[2]))
#for stop_tag in stop_tags:
# ht.add_stop_tag(stop_tag)
# divide the tags up into subsets
divvy = ht.divide_tags_into_subsets(SUBSET_SIZE)
n_subsets = len(divvy)
divvy.append(0)
# build a queue of tasks:
worker_q = Queue.Queue()
for i in range(0, n_subsets):
if stop_after_n_subsets is not None and i >= stop_after_n_subsets:
break
start = divvy[i]
end = divvy[i+1]
worker_q.put((ht, i, start, end))
open('%s.info' % basename, 'w').write('%d subsets total\n' % (n_subsets))
threads = []
for th in range(N_THREADS):
t = threading.Thread(target=worker, args=(worker_q, basename))
threads.append(t)
t.start()
# wait for threads
for t in threads:
t.join()
###
del ht
gc.collect()
# create a new, empty ht object for merging; K matters, but not
# hashtable size.
ht = khmer.new_hashbits(K, 1, 1)
# load & merge all pmap files
for i in range(0, n_subsets):
pmap_file = basename + '.subset.%d.pmap' % (i,)
ht.merge_subset_from_disk(pmap_file)
# save merged partitionmap
if save_merged_pmap:
ht.save_partitionmap(basename + '.pmap.merged')
if remove_orig_pmap:
for i in range(0, n_subsets):
pmap_file = basename + '.subset.%d.pmap' % (i,)
os.unlink(pmap_file)
# output partitions!
n_partitions = ht.output_partitions(filename, basename + '.part')
(n_partitions, n_singletons) = ht.count_partitions()
print n_partitions
fd.write(str(n_partitions) + "\n")
#print os.listdir(os.getcwd())
for file in glob.glob(os.getcwd() + "/*pmap*"):
os.remove(file)
for file in glob.glob(os.getcwd() + "/*.info"):
os.remove(file)
for file in glob.glob(os.getcwd() + "/*.part"):
os.remove(file)
fd.close()
def main():
parser = argparse.ArgumentParser(description="Output k-mer abundance distribution.")
parser.add_argument("hashname")
parser.add_argument("datafile")
parser.add_argument("histout")
parser.add_argument(
"-z", "--no-zero", dest="output_zero", default=True, action="store_false", help="Do not output 0-count bins"
)
parser.add_argument(
"-s",
"--squash",
dest="squash_output",
default=False,
action="store_true",
help="Overwrite output file if it exists",
)
args = parser.parse_args()
hashfile = args.hashname
datafile = args.datafile
histout = args.histout
print "hashtable from", hashfile
ht = khmer.load_counting_hash(hashfile)
K = ht.ksize()
sizes = ht.hashsizes()
tracking = khmer._new_hashbits(K, sizes)
print "K:", K
print "HT sizes:", sizes
print "outputting to", histout
if os.path.exists(histout):
if not args.squash_output:
print >> sys.stderr, "ERROR: %s exists; not squashing." % histout
sys.exit(-1)
print "** squashing existing file %s" % histout
print "preparing hist..."
z = ht.abundance_distribution(datafile, tracking)
total = sum(z)
if 0 == total:
print >> sys.stderr, "ERROR: abundance distribution is uniformly zero; nothing to report."
print >> sys.stderr, "\tPlease verify that the input files are valid."
sys.exit(-1)
fp = open(histout, "w")
sofar = 0
for n, i in enumerate(z):
if i == 0 and not args.output_zero:
continue
sofar += i
frac = sofar / float(total)
print >> fp, n, i, sofar, round(frac, 3)
if sofar == total:
break
