def test_call_end2end(self, compressed_inputs):
FLAGS.infile = make_golden_dataset(compressed_inputs)
FLAGS.ref = test_utils.CHR20_FASTA
FLAGS.outfile = test_utils.test_tmpfile('calls.vcf')
postprocess_variants.main(['postprocess_variants.py'])
self.assertEqual(
tf.gfile.FastGFile(FLAGS.outfile).readlines(),
tf.gfile.FastGFile(
test_utils.GOLDEN_POSTPROCESS_OUTPUT).readlines())
def test_catches_bad_argv(self):
# Define valid flags to ensure raise occurs due to argv issues.
FLAGS.infile = make_golden_dataset(False)
FLAGS.ref = testdata.CHR20_FASTA
FLAGS.outfile = test_utils.test_tmpfile('nonempty_outfile.vcf')
with mock.patch.object(logging, 'error') as mock_logging,\
mock.patch.object(sys, 'exit') as mock_exit:
postprocess_variants.main(['postprocess_variants.py', 'extra_arg'])
mock_logging.assert_called_once_with(
'Command line parsing failure: postprocess_variants does not accept '
'positional arguments but some are present on the command line: '
'"[\'postprocess_variants.py\', \'extra_arg\']".')
mock_exit.assert_called_once_with(errno.ENOENT)
def test_reading_empty_input_should_raise_error(self):
empty_shard_one = test_utils.test_tmpfile(
'no_records.tfrecord-00000-of-00002')
empty_shard_two = test_utils.test_tmpfile(
'no_records.tfrecord-00001-of-00002')
io_utils.write_tfrecords([], empty_shard_one)
io_utils.write_tfrecords([], empty_shard_two)
FLAGS.infile = test_utils.test_tmpfile('no_records.tfrecord@2')
FLAGS.ref = testdata.CHR20_FASTA
FLAGS.outfile = test_utils.test_tmpfile('no_records.vcf')
with self.assertRaisesRegexp(ValueError, 'Cannot find any records in'):
postprocess_variants.main(['postprocess_variants.py'])
def test_catches_bad_argv(self):
# Define valid flags to ensure raise occurs due to argv issues.
FLAGS.infile = make_golden_dataset(False)
FLAGS.ref = testdata.CHR20_FASTA
FLAGS.outfile = test_utils.test_tmpfile('nonempty_outfile.vcf')
with mock.patch.object(logging, 'error') as mock_logging,\
mock.patch.object(sys, 'exit') as mock_exit:
postprocess_variants.main(['postprocess_variants.py', 'extra_arg'])
mock_logging.assert_called_once_with(
'Command line parsing failure: postprocess_variants does not accept '
'positional arguments but some are present on the command line: '
'"[\'postprocess_variants.py\', \'extra_arg\']".')
mock_exit.assert_called_once_with(errno.ENOENT)
def test_catches_bad_flags(self):
FLAGS.infile = make_golden_dataset(False)
FLAGS.ref = testdata.CHR20_FASTA
FLAGS.outfile = 'nonempty_outfile.vcf'
FLAGS.nonvariant_site_tfrecord_path = (
testdata.GOLDEN_POSTPROCESS_GVCF_INPUT)
# This is the bad flag.
FLAGS.gvcf_outfile = ''
with mock.patch.object(logging, 'error') as mock_logging, \
mock.patch.object(sys, 'exit') as mock_exit:
postprocess_variants.main(['postprocess_variants.py'])
mock_logging.assert_called_once_with(
'gVCF creation requires both nonvariant_site_tfrecord_path and '
'gvcf_outfile flags to be set.')
mock_exit.assert_called_once_with(errno.ENOENT)
def test_reading_sharded_input_with_empty_shards_does_not_crash(self):
valid_variants = io_utils.read_tfrecords(
testdata.GOLDEN_POSTPROCESS_INPUT,
proto=deepvariant_pb2.CallVariantsOutput)
empty_shard_one = test_utils.test_tmpfile(
'reading_empty_shard.tfrecord-00000-of-00002')
none_empty_shard_two = test_utils.test_tmpfile(
'reading_empty_shard.tfrecord-00001-of-00002')
io_utils.write_tfrecords([], empty_shard_one)
io_utils.write_tfrecords(valid_variants, none_empty_shard_two)
FLAGS.infile = test_utils.test_tmpfile('reading_empty_shard.tfrecord@2')
FLAGS.ref = testdata.CHR20_FASTA
FLAGS.outfile = test_utils.test_tmpfile('calls_reading_empty_shard.vcf')
postprocess_variants.main(['postprocess_variants.py'])
def test_catches_bad_flags(self):
FLAGS.infile = make_golden_dataset(False)
FLAGS.ref = testdata.CHR20_FASTA
FLAGS.outfile = 'nonempty_outfile.vcf'
FLAGS.nonvariant_site_tfrecord_path = (
testdata.GOLDEN_POSTPROCESS_GVCF_INPUT)
# This is the bad flag.
FLAGS.gvcf_outfile = ''
with mock.patch.object(logging, 'error') as mock_logging, \
mock.patch.object(sys, 'exit') as mock_exit:
postprocess_variants.main(['postprocess_variants.py'])
mock_logging.assert_called_once_with(
'gVCF creation requires both nonvariant_site_tfrecord_path and '
'gvcf_outfile flags to be set.')
mock_exit.assert_called_once_with(errno.ENOENT)
def test_call_end2end(self, compressed_inputs):
FLAGS.infile = make_golden_dataset(compressed_inputs)
FLAGS.ref = testdata.CHR20_FASTA
FLAGS.outfile = test_utils.test_tmpfile('calls.vcf')
FLAGS.nonvariant_site_tfrecord_path = (
testdata.GOLDEN_POSTPROCESS_GVCF_INPUT)
FLAGS.gvcf_outfile = test_utils.test_tmpfile('gvcf_calls.vcf')
postprocess_variants.main(['postprocess_variants.py'])
self.assertEqual(
tf.gfile.FastGFile(FLAGS.outfile).readlines(),
tf.gfile.FastGFile(testdata.GOLDEN_POSTPROCESS_OUTPUT).readlines())
self.assertEqual(
tf.gfile.FastGFile(FLAGS.gvcf_outfile).readlines(),
tf.gfile.FastGFile(testdata.GOLDEN_POSTPROCESS_GVCF_OUTPUT).readlines())
def test_call_end2end(self, compressed_inputs_and_outputs):
FLAGS.infile = make_golden_dataset(compressed_inputs_and_outputs)
FLAGS.ref = testdata.CHR20_FASTA
FLAGS.outfile = create_outfile('calls.vcf', compressed_inputs_and_outputs)
FLAGS.nonvariant_site_tfrecord_path = (
testdata.GOLDEN_POSTPROCESS_GVCF_INPUT)
FLAGS.gvcf_outfile = create_outfile('gvcf_calls.vcf',
compressed_inputs_and_outputs)
postprocess_variants.main(['postprocess_variants.py'])
self.assertEqual(
tf.gfile.GFile(FLAGS.outfile).readlines(),
tf.gfile.GFile(testdata.GOLDEN_POSTPROCESS_OUTPUT_COMPRESSED
if compressed_inputs_and_outputs else testdata
.GOLDEN_POSTPROCESS_OUTPUT).readlines())
self.assertEqual(
tf.gfile.GFile(FLAGS.gvcf_outfile).readlines(),
tf.gfile.GFile(testdata.GOLDEN_POSTPROCESS_GVCF_OUTPUT_COMPRESSED
if compressed_inputs_and_outputs else testdata
.GOLDEN_POSTPROCESS_GVCF_OUTPUT).readlines())
if compressed_inputs_and_outputs:
self.assertTrue(tf.gfile.Exists(FLAGS.outfile + '.tbi'))
self.assertTrue(tf.gfile.Exists(FLAGS.gvcf_outfile + '.tbi'))