def testSubtractNothing(self):
"""
When two input files have no overlap, subtraction must result in the
same reads as are in the first input.
"""
fasta1 = '\n'.join([
'>one',
'agtcagtcagtc',
'>two',
'acctg',
'>three',
'atgggtc',
'>four',
'atggctattgaactgtatct',
])
fasta2 = '\n'.join([
'>five',
'agtcagtcagtc',
'>six',
'acctg',
])
result = list(fastaSubtract([StringIO(fasta1), StringIO(fasta2)]))
self.assertEqual(['four', 'one', 'three', 'two'],
sorted([seq.id for seq in result]))
def testThreeFiles(self):
"""
Subtraction of three files must work correctly.
"""
fasta1 = '\n'.join([
'>one',
'agtcagtcagtc',
'>two',
'acctg',
'>three',
'atgggtc',
'>four',
'atggctattgaactgtatct',
])
fasta2 = '\n'.join([
'>one',
'agtcagtcagtc',
])
fasta3 = '\n'.join([
'>two',
'acctg',
'>three',
'atgggtc',
])
result = list(fastaSubtract([StringIO(fasta1),
StringIO(fasta2),
StringIO(fasta3)]))
self.assertEqual(len(result), 1)
self.assertEqual(str(result[0].seq), 'atggctattgaactgtatct')
self.assertEqual(str(result[0].id), 'four')
def testSubtractNothing(self):
"""
When two input files have no overlap, subtraction must result in the
same reads as are in the first input.
"""
fasta1 = '\n'.join([
'>one',
'agtcagtcagtc',
'>two',
'acctg',
'>three',
'atgggtc',
'>four',
'atggctattgaactgtatct',
])
fasta2 = '\n'.join([
'>five',
'agtcagtcagtc',
'>six',
'acctg',
])
result = list(fastaSubtract([StringIO(fasta1), StringIO(fasta2)]))
self.assertEqual(['four', 'one', 'three', 'two'],
sorted([seq.id for seq in result]))
def testThreeFiles(self):
"""
Subtraction of three files must work correctly.
"""
fasta1 = '\n'.join([
'>one',
'agtcagtcagtc',
'>two',
'acctg',
'>three',
'atgggtc',
'>four',
'atggctattgaactgtatct',
])
fasta2 = '\n'.join([
'>one',
'agtcagtcagtc',
])
fasta3 = '\n'.join([
'>two',
'acctg',
'>three',
'atgggtc',
])
result = list(
fastaSubtract(
[StringIO(fasta1),
StringIO(fasta2),
StringIO(fasta3)]))
self.assertEqual(len(result), 1)
self.assertEqual(str(result[0].seq), 'atggctattgaactgtatct')
self.assertEqual(str(result[0].id), 'four')
def testSubtractFromNothing(self):
"""
When the first file is empty, the result shoud be too.
"""
fasta1 = ''
fasta2 = '\n'.join([
'>five',
'agtcagtcagtc',
'>six',
'acctg',
])
result = list(fastaSubtract([StringIO(fasta1), StringIO(fasta2)]))
self.assertEqual([], result)
def testSubtractFromNothing(self):
"""
When the first file is empty, the result shoud be too.
"""
fasta1 = ''
fasta2 = '\n'.join([
'>five',
'agtcagtcagtc',
'>six',
'acctg',
])
result = list(fastaSubtract([StringIO(fasta1), StringIO(fasta2)]))
self.assertEqual([], result)
def testOneFile(self):
"""
When just one file is passed we should get a result that has as many
reads as was in the single input file.
"""
fasta1 = '\n'.join([
'>one',
'agtcagtcagtc',
'>two',
'acctg',
'>three',
'atgggtc',
'>four',
'atggctattgaactgtatct',
])
result = list(fastaSubtract([StringIO(fasta1)]))
self.assertEqual(len(result), 4)
def testSubtractEverything(self):
"""
When two input files have the same reads, subtraction must result in an
empty (no reads) output.
"""
fasta1 = '\n'.join([
'>one',
'agtcagtcagtc',
'>two',
'acctg',
'>three',
'atgggtc',
'>four',
'atggctattgaactgtatct',
])
result = list(fastaSubtract([StringIO(fasta1), StringIO(fasta1)]))
self.assertEqual([], result)
def testSubtractEverything(self):
"""
When two input files have the same reads, subtraction must result in an
empty (no reads) output.
"""
fasta1 = '\n'.join([
'>one',
'agtcagtcagtc',
'>two',
'acctg',
'>three',
'atgggtc',
'>four',
'atggctattgaactgtatct',
])
result = list(fastaSubtract([StringIO(fasta1), StringIO(fasta1)]))
self.assertEqual([], result)
def testOneFile(self):
"""
When just one file is passed we should get a result that has as many
reads as was in the single input file.
"""
fasta1 = '\n'.join([
'>one',
'agtcagtcagtc',
'>two',
'acctg',
'>three',
'atgggtc',
'>four',
'atggctattgaactgtatct',
])
result = list(fastaSubtract([StringIO(fasta1)]))
self.assertEqual(len(result), 4)
#!/usr/bin/env python
import sys
from Bio import SeqIO
from dark import fasta
if __name__ == '__main__':
if len(sys.argv) == 1:
print >>sys.stderr, 'Usage: %s 1.fasta, 2.fasta, ... > seq.fasta' % (
sys.argv[0])
sys.exit(1)
else:
reads = fasta.fastaSubtract(map(open, sys.argv[1:]))
SeqIO.write(reads, sys.stdout, 'fasta')
#!/usr/bin/env python
from __future__ import print_function
import sys
from Bio import SeqIO
from dark import fasta
if __name__ == '__main__':
if len(sys.argv) == 1:
print('Usage: %s 1.fasta, 2.fasta, ... > seq.fasta' % (sys.argv[0]),
file=sys.stderr)
sys.exit(1)
else:
reads = fasta.fastaSubtract(map(open, sys.argv[1:]))
SeqIO.write(reads, sys.stdout, 'fasta')
