tmpdir += '/'
hashobject = Fastq_Reader(inputdir,outputdir)
cp = np.load(hashobject.output_path+'cluster_probs.npy')
cluster_probs = dict(enumerate(cp))
Hashq_Files = glob.glob(os.path.join(hashobject.input_path,'*.hashq.*'))
Hashq_Files = [fp for fp in Hashq_Files if '.tmp' not in fp]
Hashq_Files.sort()
infile = Hashq_Files[fr]
outpart = infile[-6:-3]
sample_id = infile[infile.rfind('/')+1:infile.index('.hashq')]
tmpdir += str(fr) + '/'
os.system('mkdir '+tmpdir)
G = [open('%s%s.%s.cols.%d' % (tmpdir,sample_id,outpart,i),'w') for i in range(0,2**hashobject.hash_size,2**hashobject.hash_size/50)]
f = gzip.open(infile)
r_id = 0
for a in hashobject.hash_read_generator(f):
for x in a[2]:
G[int(x*50/2**hashobject.hash_size)].write('%d\t%d\n' % (x,r_id))
r_id += 1
R = r_id
f.close()
for g in G:
g.close()
if R < 50:
print 'Fewer than 50 reads...doing nothing'
else:
ClusterFile = open(hashobject.output_path+'cluster_cols.npy')
ValueFile = open(hashobject.output_path+'cluster_vals.npy')
G = [open('%s%s.%s.ids.%d' % (tmpdir,sample_id,outpart,i),'w') for i in range(0,R,R/50)]
# If sharing ClusterFile among many jobs is not practical, we may aggregate jobs below by 1/50 ClusterFile fractions across samples (so each job reads 1 fraction)
for i in range(0,2**hashobject.hash_size,2**hashobject.hash_size/50):
Hashq_Files = glob.glob(os.path.join(hashobject.input_path, '*.hashq.*'))
Hashq_Files = [fp for fp in Hashq_Files if '.tmp' not in fp]
Hashq_Files.sort()
infile = Hashq_Files[fr]
outpart = infile[-6:-3]
sample_id = infile[infile.rfind('/') + 1:infile.index('.hashq')]
tmpdir += str(fr) + '/'
os.system('mkdir ' + tmpdir)
G = [
open('%s%s.%s.cols.%d' % (tmpdir, sample_id, outpart, i), 'w')
for i in range(0, 2**hashobject.hash_size, 2**hashobject.hash_size /
50)
]
f = gzip.open(infile)
r_id = 0
for a in hashobject.hash_read_generator(f):
for x in a[2]:
G[int(x * 50 / 2**hashobject.hash_size)].write('%d\t%d\n' %
(x, r_id))
r_id += 1
R = r_id
f.close()
for g in G:
g.close()
if R < 50:
print 'Fewer than 50 reads...doing nothing'
else:
ClusterFile = open(hashobject.output_path + 'cluster_cols.npy')
ValueFile = open(hashobject.output_path + 'cluster_vals.npy')
G = [
open('%s%s.%s.ids.%d' % (tmpdir, sample_id, outpart, i), 'w')
