def test_make_read(self):
bases = 'ACG'
quals = [30, 40, 50]
cigar = '3M'
mapq = 42
chrom = 'chr10'
start = 123
name = 'myname'
read = test_utils.make_read(bases,
quals=quals,
cigar=cigar,
mapq=mapq,
chrom=chrom,
start=start,
name=name)
self.assertEqual(read.aligned_sequence, bases)
self.assertEqual(read.aligned_quality, quals)
self.assertEqual(list(read.alignment.cigar), [
cigar_pb2.CigarUnit(operation_length=3,
operation=cigar_pb2.CigarUnit.ALIGNMENT_MATCH)
])
self.assertEqual(read.alignment.mapping_quality, mapq)
self.assertEqual(read.alignment.position.reference_name, chrom)
self.assertEqual(read.alignment.position.position, start)
self.assertEqual(read.fragment_name, name)
def to_cigar_unit(source):
"""Creates a cigar_pb2 CigarUnit from source.
This function attempts to convert source into a CigarUnit protobuf. If
source is a string, it must be a single CIGAR string specification like
'12M'. If source is a tuple or a list, must have exactly two elements
(operation_length, opstr). operation_length can be a string or int, and must
be >= 1. opstr should be a single character CIGAR specification (e.g., 'M').
If source is already a CigarUnit, it is just passed through unmodified.
Args:
source: many types allowed. The object we want to convert to a CigarUnit
proto.
Returns:
CigarUnit proto with operation_length and operation set to values from
source.
Raises:
ValueError: if source cannot be converted or is malformed.
"""
try:
if isinstance(source, cigar_pb2.CigarUnit):
return source
elif isinstance(source, six.string_types):
l, op = source[:-1], source[-1]
elif isinstance(source, (tuple, list)):
l, op = source
else:
raise ValueError('Unexpected source', source)
if isinstance(op, six.string_types):
op = CHAR_TO_CIGAR_OPS[op]
l = int(l)
if l < 1:
raise ValueError('Length must be >= 1', l)
return cigar_pb2.CigarUnit(operation=op, operation_length=int(l))
except (KeyError, IndexError):
raise ValueError(
'Failed to convert {} into a CigarUnit'.format(source))
from __future__ import absolute_import
from __future__ import division
from __future__ import print_function
import itertools
from absl.testing import absltest
from absl.testing import parameterized
from nucleus.protos import cigar_pb2
from nucleus.util import cigar
_CIGAR_TUPLES_AND_CIGAR_UNITS = [
((1, 'M'),
cigar_pb2.CigarUnit(operation=cigar_pb2.CigarUnit.ALIGNMENT_MATCH,
operation_length=1)),
((2, 'I'),
cigar_pb2.CigarUnit(operation=cigar_pb2.CigarUnit.INSERT,
operation_length=2)),
((3, 'D'),
cigar_pb2.CigarUnit(operation=cigar_pb2.CigarUnit.DELETE,
operation_length=3)),
((4, 'N'),
cigar_pb2.CigarUnit(operation=cigar_pb2.CigarUnit.SKIP,
operation_length=4)),
((5, 'S'),
cigar_pb2.CigarUnit(operation=cigar_pb2.CigarUnit.CLIP_SOFT,
operation_length=5)),
((6, 'H'),
cigar_pb2.CigarUnit(operation=cigar_pb2.CigarUnit.CLIP_HARD,
operation_length=6)),