def run(parser, args):
args = parser.parse_args()
# args = check_args(args)
if args.quiet:
logger.setLevel(logging.ERROR)
else:
logger.setLevel(logging.INFO)
DBNAME = 'atlas-%s' % (args.db_name)
db = client[DBNAME]
connect(DBNAME)
logger.debug("Using DB %s" % DBNAME)
reference_set_name = ".".join(
os.path.basename(args.reference_set).split(".")[:-1])
try:
reference_set = ReferenceSet.objects.get(name=reference_set_name)
except DoesNotExist:
reference_set = ReferenceSet.create_and_save(name=reference_set_name)
# Hack
try:
reference = Reference.create_and_save(name=reference_set_name,
reference_sets=[reference_set],
md5checksum="NA")
except:
pass
vcf = VCF(args.vcf, reference_set.id, method=args.method, force=args.force)
vcf.add_to_database()
def test_add_add_variants_and_calls(self):
vcf = VCF(f="tests/vcf_tests/test.vcf",
reference_set_id=self.reference_set.id,
method="CORTEX")
vcf.add_to_database()
assert VariantCall.objects().count() == 21
assert Variant.objects().count() == 21
def run(parser, args):
args = parser.parse_args()
# args = check_args(args)
if args.quiet:
logger.setLevel(logging.ERROR)
else:
logger.setLevel(logging.INFO)
DBNAME = "%s-%s" % (DB_PREFIX, args.db_name)
connect(DBNAME, host=args.db_uri)
logger.debug("Using DB %s" % DBNAME)
with open(args.reference_set) as fp:
reference_set_name = get_first_chrom_name(fp)
try:
reference_set = ReferenceSet.objects.get(name=reference_set_name)
except DoesNotExist:
reference_set = ReferenceSet.create_and_save(name=reference_set_name)
try:
reference = Reference.create_and_save(name=reference_set_name,
reference_sets=[reference_set],
md5checksum="NA")
except:
pass
vcf = VCF(args.vcf, reference_set.id, method=args.method, force=args.force)
vcf.add_to_database()
def test_add_new_vcf_variant_set(self):
vcf = VCF(f="tests/vcf_tests/test.vcf",
reference_set_id=self.reference_set.id,
method="CORTEX")
vcf.add_to_database()
assert VariantSetMetadata.objects().count() >= 2
assert VariantSetMetadata.objects(key="KMER").count() == 2
def test_add_new_vcf_variant_set(self):
vcf = VCF(f="tests/vcf_tests/test.vcf",
reference_set_id=self.reference_set.id,
method="CORTEX")
vcf.add_to_database()
# We create a global variant set as well as one for the individual VCF
assert VariantSet.objects().count() == 2
vs = VariantSet.objects()[0]
assert len(Variant.objects()[0].variant_sets) == 2
assert vs.name == "test.vcf"
def setup(self):
DB.drop_database('mykrobe-test')
self.reference_set = ReferenceSet().create_and_save(name="ref_set")
self.reference = Reference().create_and_save(
name="NC_000962.3",
md5checksum="sre",
reference_sets=[self.reference_set])
vcf = VCF(f="tests/vcf_tests/test.vcf",
reference_set_id=self.reference_set.id,
method="CORTEX")
vcf.add_to_database()
def test_add_new_call_set(self):
vcf = VCF(f="tests/vcf_tests/test.vcf",
reference_set_id=self.reference_set.id,
method="CORTEX")
vcf.add_to_database()
# Only one callset but the callset should belong to multiple variant
# sets
assert VariantCallSet.objects().count() == 1
assert VariantCallSet.objects()[0].created_at <= datetime.datetime.now(
)
assert len(VariantCallSet.objects()[0].variant_sets) == 2
def test_add_second_vcf_variant_set(self):
# This VCF only has one Variant which is not in the first VCF
vcf = VCF(f="tests/vcf_tests/test2.vcf",
reference_set_id=self.reference_set.id,
method="CORTEX")
vcf.add_to_database()
assert VariantSet.objects().count() == 3
assert VariantCallSet.objects().count() == 2
assert VariantCall.objects().count() == 42
assert Variant.objects().count() == 22
assert len(Variant.objects()[0].variant_sets) == 3
assert len(
Variant.objects.get(
names="UNION_BC_k31_var_147").variant_sets) == 3
def test_add_second_vcf_variant_set(self):
# This VCF only has one Variant which is not in the first VCF
vcf = VCF(f="tests/vcf_tests/test3.vcf",
reference_set_id=self.reference_set.id,
method="CORTEX")
vcf.add_to_database()
assert VariantSet.objects().count() == 2
assert VariantCallSet.objects().count() == 1
assert VariantCall.objects().count() == 106
assert Variant.objects().count() == 106
assert Variant.snps().count() == 89
assert Variant.indels().count() == 17
assert Variant.insertions().count() == 8
assert Variant.deletions().count() == 8
assert Variant.ph_snps.count() == 1
def run_make_probes_from_vcf_file(args):
# Make VariantSet from vcf
reference = os.path.basename(args.reference_filepath).split(".fa")[0]
try:
reference_set = ReferenceSet.objects.get(name=reference)
except DoesNotExist:
reference_set = ReferenceSet.create_and_save(name=reference)
# Hack
try:
reference = Reference.create_and_save(
name=reference,
reference_sets=[reference_set],
md5checksum=reference)
except NotUniqueError:
pass
vcf = VCF(
args.vcf,
reference_set.id,
method="tmp",
force=True,
append_to_global_variant_set=False)
vcf.add_to_database()