def downloadFiles(infiles, outfile):
infile = infiles
basefile = os.path.basename(infile)
filename = "temp_bams/%s" % basefile
baseoutfile = os.path.basename(outfile)
outdir = os.path.dirname(outfile)
if infile.endswith(".remote"):
for line in IOTools.open_file(infile):
repo, acc = line.strip().split("\t")[:2]
if repo == "SRA":
if not os.path.isfile(outfile + ".1.gz"):
statement = "; ".join(
[Sra.prefetch(acc),
Sra.extract(acc, outdir)])
P.run(statement)
else:
pass
elif repo == "GDC":
base = os.path.splitext(basefile)
outfile = "bam.dir/" + base[0] + ".bam"
token = glob.glob("gdc-user-token*")
if len(token) > 0:
token = token[0]
else:
token = None
s, infile = Sra.process_remote_BAM(
infile,
token,
filename,
filter_bed=os.path.join(
PARAMS["annotations_dir"],
PARAMS["annotations_interface_contigs_bed"]))
infile = " ".join(infile)
if not os.path.isfile(outfile):
statement = "; ".join([
"mkdir -p %(filename)s", s,
'''cp %(infile)s %(outfile)s;
rm -r %(filename)s'''
])
P.run(statement)
else:
pass
else:
raise ValueError("Unknown repository: %s" % repo)
else:
pass
def downloadSequinsNeatData(outfile):
''' Download the neat Sequins data from NCBI'''
address_base = 'ftp://ftp-trace.ncbi.nlm.nih.gov/sra/sra-instant/reads/ByExp/sra/SRX/SRX189'
outfile2srr = {
'neat-A.fastq.1.gz': 'SRR3743147',
'neat-B.fastq.1.gz': 'SRR3743148'
}
srr2srx = {'SRR3743147': 'SRX1897294', 'SRR3743148': 'SRX1897295'}
outfile_base = os.path.basename(outfile)
srr = outfile2srr[outfile_base]
srx = srr2srx[srr]
outfile_name = P.snip(outfile_base, '.fastq.1.gz')
statement = '''
wget %(address_base)s/%(srx)s/%(srr)s/%(srr)s.sra
-O %(outfile_name)s.sra
'''
P.run()
outdir = os.path.dirname(outfile)
statement = Sra.extract(outfile_name + '.sra', outdir)
P.run()
statement = '''
mv %(outdir)s/%(outfile_name)s_1.fastq.gz
%(outdir)s/%(outfile_name)s.fastq.1.gz; checkpoint;
mv %(outdir)s/%(outfile_name)s_2.fastq.gz
%(outdir)s/%(outfile_name)s.fastq.2.gz'''
P.run()
os.unlink(outfile_name + '.sra')
def extractGSE65525(infile, outfile):
''' extract fastqs '''
statement = SRA.extract(infile, "GSE65525/fastqs.dir")
P.run()
def extractGGSE53638(infile, outfile):
''' extract the fastqs from the SRA '''
statement = SRA.extract(infile, "GSE53638/fastqs.dir")
P.run()
def extractGSE65525(infile, outfile):
''' extract fastqs '''
statement = SRA.extract(infile, "GSE65525/fastqs.dir")
P.run()
def extractGGSE53638(infile, outfile):
''' extract the fastqs from the SRA '''
statement = SRA.extract(infile, "GSE53638/fastqs.dir")
P.run()