def testGetRestOfPaths(self):
pred_str = """##fileformat=VCFv4.0\n
##FORMAT=<ID=GT,Number=1,Type=String,Description="Genotype">\n
##source=TVsim\n
#CHROM POS ID REF ALT QUAL FILTER INFO FORMAT NA00001\n
chr19 11 . ACT A 20 PASS . GT 1/1\n
chr19 15 . ACGATT AA 20 PASS . GT 1/1\n
chr19 16 . ACG A 20 PASS . GT 1/1\n
chr19 22 . ATT A 20 PASS . GT 0/1\n
"""
pred_io = StringIO.StringIO(pred_str)
pred_vcf = vcf.Reader(pred_io)
pred_vars = Variants(pred_vcf,MAX_INDEL_LEN)
viw = extract_range_and_filter(pred_vars.on_chrom('chr19'),10,25,11)
paths = _getRestOfPath([], _getOverlaps([],viw))
#all paths take variants at pos 11 and 22; one takes pos 15, one pos 16
self.assertEqual(len(paths),2)
self.assertEqual(len(paths[0]),3)
self.assertEqual(len(paths[1]),3)
self.assertTrue(all(map(lambda e: any(map(lambda x: x.pos == 11, e)), paths)))
self.assertTrue(all(map(lambda e: any(map(lambda x: x.pos == 22, e)), paths)))
self.assertTrue(any(map(lambda x: x.pos == 15, paths[0])))
self.assertFalse(any(map(lambda x: x.pos == 16, paths[0])))
self.assertFalse(any(map(lambda x: x.pos == 15, paths[1])))
self.assertTrue(any(map(lambda x: x.pos == 16, paths[1])))
def testGetOverlaps(self):
pred_str = """##fileformat=VCFv4.0\n
##FORMAT=<ID=GT,Number=1,Type=String,Description="Genotype">\n
##source=TVsim\n
#CHROM POS ID REF ALT QUAL FILTER INFO FORMAT NA00001\n
chr19 10 . ACT A 20 PASS . GT 1/1\n
chr19 13 . AC A 20 PASS . GT 1/1\n
chr19 14 . TAGG TA 20 PASS . GT 1/1\n
chr19 15 . AGG A 20 PASS . GT 0/1\n
chr19 19 . T TAAAC 20 PASS . GT 0/1
"""
pred_io = StringIO.StringIO(pred_str)
pred_vcf = vcf.Reader(pred_io)
pred_vars = Variants(pred_vcf,MAX_INDEL_LEN)
variants_in_window = extract_range_and_filter(pred_vars.on_chrom('chr19'),10,20,10)
#the three overlapping variants should be in same group
overlaps = _getOverlaps([], variants_in_window)
self.assertEqual(len(overlaps),3)
self.assertEqual(map(lambda o: len(o),overlaps),[1,3,1])