def test_iterate_bed_reader(self, bed_filename):
bed_path = test_utils.genomics_core_testdata(bed_filename)
expected = [('chr1', 10, 20), ('chr1', 100, 200)]
with bed.BedReader(bed_path) as reader:
records = list(reader.iterate())
self.assertLen(records, 2)
self.assertEqual([(r.reference_name, r.start, r.end) for r in records],
expected)
def test_roundtrip_writer(self, filename):
output_path = test_utils.test_tmpfile(filename)
with bed.BedWriter(output_path,
header=bed_pb2.BedHeader(num_fields=5)) as writer:
for record in self.records:
writer.write(record)
with bed.BedReader(output_path) as reader:
v2_records = list(reader.iterate())
self.assertEqual(self.records, v2_records)
def test_roundtrip_num_fields(self, num_fields):
all_num_fields_in_file = [
n for n in _VALID_NUM_BED_FIELDS if n >= num_fields
]
for num_fields_in_file in all_num_fields_in_file:
lines = ['\t'.join(line[:num_fields_in_file]) for line in self.tokens]
contents = '{}\n'.format('\n'.join(lines))
input_path = test_utils.test_tmpfile('test_field.bed', contents=contents)
with bed.BedReader(input_path, num_fields=num_fields) as reader:
records = list(reader.iterate())
output_path = test_utils.test_tmpfile('test_field2.bed')
with bed.BedWriter(output_path, header=reader.header) as writer:
for record in records:
writer.write(record)
with bed.BedReader(output_path) as reader2:
v2_records = list(reader2.iterate())
self.assertLen(records, 3)
self.assertEqual(records, v2_records)
def bed_parser(filename):
"""Parses Range objects from a BED-formatted file object.
See http://bedtools.readthedocs.org/en/latest/content/general-usage.html
for more information on the BED format.
Args:
filename: file name of a BED-formatted file.
Yields:
nucleus.genomics.v1.Range protobuf objects.
"""
with bed.BedReader(filename) as fin:
for r in fin.iterate():
yield make_range(r.reference_name, r.start, r.end)
def test_invalid_num_fields(self, invalid_num_fields):
bed_path = test_utils.genomics_core_testdata('test_regions.bed')
with self.assertRaisesRegexp(ValueError,
'Invalid requested number of fie'):
reader = bed.BedReader(bed_path, num_fields=invalid_num_fields)
def test_native_bed_header(self, bed_filename):
bed_path = test_utils.genomics_core_testdata(bed_filename)
with bed.BedReader(bed_path) as reader:
self.assertEqual(reader.header.num_fields, 12)
with bed.NativeBedReader(bed_path) as native_reader:
self.assertEqual(native_reader.header.num_fields, 12)