def test_split_byte_murmur():
# check the byte is correctly split when using murmur hash
sct = SmallCounttable(4, 4, 1)
# these kmers were carefully chosen to have hash values that produce
# consecutive indices in the count table.
a = "AAAC"
b = "AAAG"
assert sct.get_kmer_hashes(a) == [11898086063751343884]
assert sct.get_kmer_hashes(b) == [10548630838975263317]
sct.add(a)
assert sct.get(a) == 1
assert sct.get(b) == 0
def test_save_load(Tabletype):
kh = Tabletype(5)
ttype = type(kh)
savefile = utils.get_temp_filename('tablesave.out')
# test add(dna)
x = kh.add("ATGGC")
z = kh.get("ATGGC")
assert z == 1
kh.save(savefile)
# should we provide a single load function here? yes, probably. @CTB
if ttype == _Countgraph:
loaded = khmer.load_countgraph(savefile)
elif ttype == Counttable:
loaded = Counttable.load(savefile)
elif ttype == _SmallCountgraph:
loaded = khmer.load_countgraph(savefile, small=True)
elif ttype == SmallCounttable:
loaded = SmallCounttable.load(savefile)
elif ttype == _Nodegraph:
loaded = khmer.load_nodegraph(savefile)
elif ttype == Nodetable:
loaded = Nodetable.load(savefile)
else:
raise Exception("unknown tabletype")
z = loaded.get('ATGGC')
assert z == 1
def test_read_write():
rng = random.Random(1)
sct = SmallCounttable(20, 1e2, 4)
kmers = ["".join(rng.choice("ACGT") for _ in range(20))
for n in range(400)]
for kmer in kmers:
sct.add(kmer)
fname = utils.get_temp_filename('zzz')
sct.save(fname)
# on purpose choose parameters that are different from sct
sct2 = SmallCounttable.load(fname)
assert sct.ksize() == sct2.ksize()
for kmer in kmers:
assert sct.get(kmer) == sct2.get(kmer)
def test_overflow():
# check that we do not overflow into other part of the byte
sct = SmallCounttable(4, 1e6, 4)
a = "AAAA"
b = "AAAT"
# try to overflow our 4bit counter
for n in range(17):
sct.add(a)
assert sct.get(a) == 15
assert sct.get(b) == 0
# repeat with the other kmer that hashes to the other half of the byte
sct = SmallCounttable(4, 1e6, 4)
a = "AAAA"
b = "AAAT"
# try to overflow our 4bit counter
for n in range(17):
sct.add(b)
assert sct.get(b) == 15
assert sct.get(a) == 0
def test_random_kmers():
# check for out-of-bounds errors and similar with random kmers
rng = random.Random(1)
sct = SmallCounttable(20, 1e2, 4)
kmers = ["".join(rng.choice("ACGT") for _ in range(20))
for n in range(400)]
for kmer in kmers:
sct.add(kmer)
for kmer in kmers:
sct.get(kmer)
def test_split_byte_murmur():
# check the byte is correctly split when using murmur hash
sct = SmallCounttable(4, 4, 1)
# these kmers were carefully chosen to have hash values that produce
# consecutive indices in the count table.
a = "AAAC"
b = "AAAG"
assert sct.get_kmer_hashes(a) == [11898086063751343884]
assert sct.get_kmer_hashes(b) == [10548630838975263317]
sct.add(a)
assert sct.get(a) == 1
assert sct.get(b) == 0
def test_overflow():
# check that we do not overflow into other part of the byte
sct = SmallCounttable(4, 1e6, 4)
a = "AAAA"
b = "AAAT"
# try to overflow our 4bit counter
for n in range(17):
sct.add(a)
assert sct.get(a) == 15
assert sct.get(b) == 0
# repeat with the other kmer that hashes to the other half of the byte
sct = SmallCounttable(4, 1e6, 4)
a = "AAAA"
b = "AAAT"
# try to overflow our 4bit counter
for n in range(17):
sct.add(b)
assert sct.get(b) == 15
assert sct.get(a) == 0
def test_single_add():
sct = SmallCounttable(4, 1e6, 4)
sct.add("AAAA")
assert sct.get("AAAA") == 1
def test_single_add():
sct = SmallCounttable(4, 1e6, 4)
sct.add("AAAA")
assert sct.get("AAAA") == 1