def test_get_adj_str_ndarray(self, tmpdir):
cw = CountsWeighter(k=1)
array = np.array([[1, 2, 3, 4], [5, 6, 7, 8]])
array_path = str(Path(tmpdir, "out.npy"))
np.save(array_path, array)
counts = cw.get_counts(array_path)
expected = pd.DataFrame(array, columns=["A", "G", "T", "C"])
assert expected.equals(counts)
def test_build_weights_dict_minimal(self):
pwm = self.minimal_pwm()
cw = CountsWeighter(k=2)
kmer2weight = cw.build_weights_dict(pwm)
assert kmer2weight["AA"] == (0.5 * 0.5) + (0.5 * 0.95)
assert kmer2weight["GG"] == (0.1 * 0.2) + (0.2 * 0.05)
assert kmer2weight["CC"] == (0.1 * 0.3) + (0.3 * 0)
assert kmer2weight["AG"] == (0.5 * 0.2) + (0.5 * 0.05)
assert len(kmer2weight) == 16
def test_run(self):
pwm_dir = pkg_resources.resource_filename("seekr", "tests/data/pwms/")
counts = pkg_resources.resource_filename(
"seekr", "tests/data/example_2mers.npy")
cw = CountsWeighter(pwm_dir, counts, k=2)
cw.run()
expected = np.array(
[[4.56194883, 5.65991574, 6.5880013, 6.47169944, 8.14747599]])
assert np.allclose(cw.df.values, expected)
def test_weight_counts(self):
kmers = "AGTC"
kmer2weight = dict(zip(kmers, range(4)))
cw = CountsWeighter(k=1)
counts = [[1, 2, 3, 4], [1, 1, 1, 1], [1, 2, 1, 2]]
counts = pd.DataFrame(np.array(counts), columns=list(kmers))
cw.counts = counts
sum_scores = cw.weight_counts(kmer2weight)
assert np.all(sum_scores == np.array([20, 6, 10]))
def test_build_weights_dict_minimal(self):
pwm = self.minimal_pwm()
cw = CountsWeighter(k=2)
kmer2weight = cw.build_weights_dict(pwm)
assert kmer2weight['AA'] == (.5 * .5) + (.5 * .95)
assert kmer2weight['GG'] == (.1 * .2) + (.2 * .05)
assert kmer2weight['CC'] == (.1 * .3) + (.3 * 0)
assert kmer2weight['AG'] == (.5 * .2) + (.5 * 0.05)
assert len(kmer2weight) == 16
def test_build_weights_dict_full(self):
pwm = pkg_resources.resource_filename("seekr",
"tests/data/pwms/M001_0.6.txt")
pwm = pd.read_csv(pwm, sep="\t").rename(columns={"U": "T"}).to_dict()
cw = CountsWeighter(k=2)
kmer2weight = cw.build_weights_dict(pwm)
assert kmer2weight["AA"] == 0.9749391864711447
assert kmer2weight["CG"] == 0.00473602191097646
assert kmer2weight["CT"] == 0.1186835711075973
assert kmer2weight["AG"] == 0.02190317370024123
assert len(kmer2weight) == 16
def test_set_kmer2weight(self):
kmer2weight = defaultdict(int)
pwm = self.minimal_pwm()
cw = CountsWeighter(k=2)
for kmer in ("".join(p) for p in product("AGTC", repeat=2)):
cw.set_kmer2weight(kmer2weight, pwm, kmer, kmer, 2)
assert kmer2weight["AA"] == (0.5 * 0.5) + (0.5 * 0.95)
assert kmer2weight["GG"] == (0.1 * 0.2) + (0.2 * 0.05)
assert kmer2weight["CC"] == (0.1 * 0.3) + (0.3 * 0)
assert kmer2weight["AG"] == (0.5 * 0.2) + (0.5 * 0.05)
assert len(kmer2weight) == 16
def test_get_pwm_dicts(self):
pwm_dir = pkg_resources.resource_filename("seekr", "tests/data/pwms/")
cw = CountsWeighter(pwm_dir, k=1)
pwm_path, pwm = next(cw.gen_pwm_dicts())
expected = Path(
pkg_resources.resource_filename("seekr",
"tests/data/pwms/M001_0.6.txt"))
assert pwm_path == expected
assert len(pwm) == 4
assert len(pwm["A"]) == 7
assert pwm["A"][0] == 0.39532879396435
def test_build_weights_dict_full(self):
pwm = pkg_resources.resource_filename('seekr',
'tests/data/pwms/M001_0.6.txt')
pwm = pd.read_csv(pwm, sep='\t').rename(columns={'U': 'T'}).to_dict()
cw = CountsWeighter(k=2)
kmer2weight = cw.build_weights_dict(pwm)
assert kmer2weight['AA'] == 0.9749391864711447
assert kmer2weight['CG'] == 0.00473602191097646
assert kmer2weight['CT'] == 0.1186835711075973
assert kmer2weight['AG'] == 0.02190317370024123
assert len(kmer2weight) == 16
def test_set_kmer2weight(self):
kmer2weight = defaultdict(int)
pwm = self.minimal_pwm()
cw = CountsWeighter(k=2)
for kmer in (''.join(p) for p in product('AGTC', repeat=2)):
cw.set_kmer2weight(kmer2weight, pwm, kmer, kmer, 2)
assert kmer2weight['AA'] == (.5 * .5) + (.5 * .95)
assert kmer2weight['GG'] == (.1 * .2) + (.2 * .05)
assert kmer2weight['CC'] == (.1 * .3) + (.3 * 0)
assert kmer2weight['AG'] == (.5 * .2) + (.5 * 0.05)
assert len(kmer2weight) == 16
def _run_pwms(pwm_dir, counts, kmer, out_path):
# TODO (Dan) name this function
# Note: This function is separated for testing purposes.
counts_weighter = CountsWeighter(pwm_dir, counts, kmer, out_path)
counts_weighter.run()
def test_get_adj_df(self):
cw = CountsWeighter(k=1)
array = np.array([[1, 2, 3, 4], [5, 6, 7, 8]])
expected = pd.DataFrame(array, columns=["A", "G", "T", "C"])
counts = cw.get_counts(expected)
assert expected.equals(counts)
def test_get_adj_ndarray(self):
cw = CountsWeighter(k=1)
array = np.array([[1, 2, 3, 4], [5, 6, 7, 8]])
counts = cw.get_counts(array)
expected = pd.DataFrame(array, columns=['A', 'G', 'T', 'C'])
assert expected.equals(counts)