def test_vcf_to_csv():
import vcf, csv
vcfreader = vcf.Reader(open('tests/data/snv.vcf'))
columns = ['sample:DP', 'info:VAF', 'info:DP', 'something']
filename = vcf_to_csv(vcfreader, columns, None,
default_values={'something': True})
with open(filename) as fd:
rows = list(csv.reader(fd))
asserts.eq_(rows[0], ['44', '', '81', 'True'])
asserts.eq_(rows[-1], ['40', '', '74', 'True'])
asserts.eq_(len(rows), 20)
def test_vcf_to_csv():
vcfreader = vcf.Reader(open('tests/data/snv.vcf'))
columns = ['contig', 'position', 'sample:DP', 'info:VAF', 'info:DP', 'something']
filename = vcf_to_csv(vcfreader, columns, None,
default_values={'something': 'Something!'})
with open(filename) as f:
rows = list(csv.reader(f))
# first two are from the same sample
assert rows[0] == ['20', '61795', '44', '', '81', 'Something!']
assert rows[1] == ['20', '61795', '37', '', '81', 'Something!']
# this is the last sample
assert rows[-1] == ['20', '75254', '40', '', '74', 'Something!']
assert len(rows) == 20 # 10 records, 2 samples each = 20 rows
def test_csv_to_vcf():
# Here we just write the VCF to a CSV like we did above, and then read it
# and compare it to the original VCF to make sure we haven't lost any data
# in the conversation.
vcfreader = vcf.Reader(open('tests/data/one-sample.vcf'))
columns = ['contig', 'position', 'id', 'reference', 'alternates',
'quality', 'filters', 'info:DP', 'info:SS',
'info:SSC', 'info:GPV', 'info:SPV', 'info:SOMATIC',
'sample:GT','sample:GQ','sample:DP','sample:RD',
'sample:AD','sample:FREQ','sample:DP4',
'sample_name']
filename = vcf_to_csv(vcfreader, columns, None)
# Now we convert the rows in this CSV into dict (converting the position as
# well), which is what it would look like select from Postgres.
with open(filename) as f:
rows = list(csv.reader(f))
relations = []
for row in rows:
relation = {
col: val if val != '' else None
for col, val in zip(columns, row)
}
relation['position'] = int(relation['position'])
relations.append(relation)
text = open('tests/data/one-sample.vcf').readlines()
header = (line for line in text if line.startswith('#'))
template = vcf.Reader(header)
# Here we convert those dicts to vcf.Records, for later writing.
records = genotypes_to_records(relations, template, columns)
# Now we write the output of those to a string buffer...
vcf_sb = StringIO.StringIO()
writer = vcf.Writer(vcf_sb, template)
for record in records:
writer.write_record(record)
# And see if we get the same VCF.
#
# We can't just e.g. test that the text of
# both files is equal, as pyVCF will e.g. emit header lines in different
# orders, or e.g. include fields in the INFO field which may be blank and
# thus omitted in the original VCF. Scientific notation, for example, also
# can change, e.g. 1.05 can become 0.105e-1.
#
# So, instead, we test that the same number of records, exists, that they
# all have the correct CHROM, POS, REF, ALT and select sample and info
# fields.
original_vcf = list(vcf.Reader(open('tests/data/one-sample.vcf')))
vcf_sb.seek(0)
saved_vcf = list(vcf.Reader(vcf_sb))
for o, s in zip(original_vcf, saved_vcf):
assert o.CHROM == s.CHROM
assert o.POS == s.POS
assert o.REF == s.REF
assert o.ALT == s.ALT
assert o.samples[0].data.DP == s.samples[0].data.DP
assert o.samples[0].data.GT == s.samples[0].data.GT
assert o.samples[0].data.DP4 == s.samples[0].data.DP4
assert o.INFO['DP'] == s.INFO['DP']
assert o.INFO['SS'] == s.INFO['SS']