def it_computes_zero_pr_for_subset():
p, r, s, a = cb_all_wrong.pr_curve(pep_iz_subset=[1], n_steps=2)
assert utils.np_array_same([p, r, s, a], zero_pr_result)
p, r, s, a = cb_all_wrong.pr_curve(pep_iz_subset=[2], n_steps=2)
assert utils.np_array_same([p, r, s, a], zero_pr_result)
# peptide 2 does not show up in true/pred at all so should get zero pr curve
p, r, s, a = cb_all_right.pr_curve(pep_iz_subset=[2], n_steps=2)
assert utils.np_array_same([p, r, s, a], zero_pr_result)
def it_can_mock():
# A mock cycle is the same as a edman_failure no matter what (p_edman_failure=0.0)
# and yet I should still see failure behavior
result = sim_v1_worker._step_3a_cycle_edman(samples,
is_mock=True,
p_edman_failure=0.0)
assert np_array_same(samples, result)
def it_bleaches_channels_independently():
result = sim_v1_worker._step_3b_photobleach(
samples, p_bleach_by_channel=[1.0, 0.0])
expected = npf([
[[0, 0, 0], [0, 1, 0]],
[[n, 0, 0], [n, 0, 1]],
])
assert np_array_same(expected, result)
def it_returns_train_dyepeps():
sim_v2_result, sim_v2_params = _sim()
# The order of dyts is not guaranteed, so remove them
assert utils.np_array_same(
sim_v2_result.train_dyepeps[:, (0, 2)],
np.array([[0, 0], [1, 5000], [2, 5000], [3, 5000],], dtype=np.uint64),
)
def it_detach():
result = sim_v1_worker._step_3c_detach(samples, 1.0)
expected = np.full_like(samples, n)
assert np_array_same(expected, result)
def it_make_a_copy():
# The original should not have changed
assert samples[0, 0, 0] == 1.0
zest()
def it_proceeds_with_no_error():
evolution = sim_v1_worker._step_3_evolve_cycles(samples, sim_params)
# fmt: off
n = np.nan
expected = npf([
[
[[1, 1, 0], [0, 1, 0]],
[[0, 1, 1], [0, 0, 1]],
],
[
[[n, 1, 0], [n, 1, 0]],
[[n, 1, 1], [n, 0, 1]],
],
])
# fmt: on
if not np_array_same(evolution, expected):
debug(evolution)
debug(sim_params)
assert np_array_same(evolution, expected)
def it_degrades_at_first_non_nan():
samples = npf([
[[n, 1, 0], [n, 1, 0]],
[[0, 1, 1], [0, 0, 1]],
])
result = sim_v1_worker._step_3a_cycle_edman(samples,
is_mock=False,
p_edman_failure=0.0)
expected = npf([
[[n, n, 0], [n, n, 0]],
[[n, 1, 1], [n, 0, 1]],
])
assert np_array_same(expected, result)
def it_proceeds_with_errors():
# Example error: detach 100%
sim_params.error_model.p_detach = 1.0
sim_params._build_join_dfs()
evolution = sim_v1_worker._step_3_evolve_cycles(samples, sim_params)
expected = np.full_like(evolution, np.nan)
assert np_array_same(evolution, expected)
def it_makes_a_copy():
assert samples[0, 0, 0] == 1.0
zest()
def it_bleaches_all():
result = sim_v1_worker._step_3b_photobleach(
samples, p_bleach_by_channel=[1.0, 1.0])
expected = npf([
[[0, 0, 0], [0, 0, 0]],
[[n, 0, 0], [n, 0, 0]],
])
assert np_array_same(expected, result)
def it_makes_a_copy():
# The original should not have changed
assert samples[0, 0, 0] == 1.0
zest()
def it_gets_flat_test_radmat():
flat_test_radmat = result.flat_test_radmat()
assert flat_test_radmat.shape == (
n_peps * n_test_samples,
n_channels * n_cycles,
)
expected = npf([
[4.0, 2.0, 1.0, 1.0, 1.0, 0.0],
[5.0, 2.0, 1.0, 1.0, 1.0, 1.0],
[6.0, 3.0, 0.9, 1.0, 0.0, 1.0],
[0.4, 0.2, 0.1, 0.3, 0.2, 0.1],
[0.5, 0.2, 0.1, 0.3, 0.2, 0.1],
[0.6, 0.3, 2.9, 0.3, 0.2, 0.1],
])
assert np_array_same(flat_test_radmat, expected)
def it_returns_train_dyemat():
# Because it has no errors, there's only a perfect dyemats
sim_v2_result, sim_v2_params = _sim()
assert sim_v2_result.train_dyemat.shape == (4, 5 * 2) # 5 cycles, 2 channels
assert utils.np_array_same(
sim_v2_result.train_dyemat,
np.array(
[
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 2, 1, 0, 0, 0],
[0, 0, 0, 0, 0, 2, 1, 1, 0, 0],
[1, 0, 0, 0, 0, 1, 1, 0, 0, 0],
],
dtype=np.uint8,
),
)
def it_can_degrade_one_sample_and_not_another():
samples = npf([
[[n, 1, 0], [n, 1, 0]],
[[0, 1, 1], [0, 0, 1]],
])
n_samples = 2
with zest.mock(sim_v1_worker._rand1,
returns=np.full((n_samples, ), [0.0, 1.0])):
result = sim_v1_worker._step_3a_cycle_edman(samples,
is_mock=False,
p_edman_failure=0.5)
expected = npf([
[[n, 1, 0], [n, 1, 0]],
[[n, 1, 1], [n, 0, 1]],
])
assert np_array_same(expected, result)
def it_can_edman_degrade():
# Degrade = Edman success; thus p_edman_failure=0.0
result = sim_v1_worker._step_3a_cycle_edman(samples,
is_mock=False,
p_edman_failure=0.0)
expected = npf([
[[n, 1, 0], [n, 1, 0]],
[[n, 1, 1], [n, 0, 1]],
])
assert np_array_same(expected, result)
def it_makes_a_copy():
# The original should not have changed
assert samples[0, 0, 0] == 1.0
zest()
def it_zeros_nans():
# fmt: off
evolution = npf([
[
[[n, n, n], [n, n, n]],
[[n, n, n], [n, n, n]],
],
])
dyemat = sim_v1_worker._step_4_make_dyemat(evolution)
expected = npf([
[
[0, 0],
[0, 0],
],
])
# fmt: on
assert np_array_same(dyemat, expected)
def it_makes_pcbs():
# fmt: off
params = SimV2Params(
is_survey=False,
dyes=[
Munch(dye_name="dye1", channel_name="640"),
Munch(dye_name="dye2", channel_name="549"),
],
labels=[
Munch(aa="A", dye_name="dye1", label_name="A", ptm_only=False),
Munch(aa="B", dye_name="dye2", label_name="B", ptm_only=False),
],
priors_desc={
"p_non_fluorescent.dye1":
dict(class_name="MLEPrior", params=dict(value=0.5)),
"p_non_fluorescent.dye2":
dict(class_name="MLEPrior", params=dict(value=0.3)),
},
channels={
"640": 0,
"549": 1,
},
)
# fmt: on
pep_seq = pd.DataFrame(
dict(
pep_i=[1, 1, 1, 2, 2],
aa=["X", "A", "X", "B", "X"],
pep_offset_in_pro=[0, 1, 2, 0, 1],
))
pcbs = params.pcbs(pep_seq)
assert utils.np_array_same(
pcbs,
[
[1.0, np.nan, np.nan],
[1.0, 0.0, 0.5],
[1.0, np.nan, np.nan],
[2.0, 1.0, 1.0 - 0.3],
[2.0, np.nan, np.nan],
],
)
def it_returns_train_dyemat_for_uncleaved_cterm_labels():
prep_cterm = prep_fixtures.result_cterm_label_fixture()
# dyemat when allow_edman_cterm is False (default)
sim_v2_result, sim_v2_params = _sim(
_prep_result=prep_cterm, sim_kwargs=Munch(allow_edman_cterm=False)
)
assert sim_v2_result.train_dyemat.shape == (3, 5 * 2) # 5 cycles, 2 channels
assert utils.np_array_same(
sim_v2_result.train_dyemat,
np.array(
[
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 1, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 1, 1, 1, 1, 1],
],
dtype=np.uint8,
),
)
def it_makes_a_flu():
assert np_array_same(compact_flu, [1, n, 0, 1])
def it_pads_an_array():
assert utils.np_array_same(utils.np_1d_end_pad(a, 4), np.array([1, 2, 3, 0]))
def it_repl_2d_vecs_1_col():
r = utils.repl_vec_over_cols(np.array([[1, 2, 3]]).T, 2)
assert utils.np_array_same(r, expected)
def it_repl_1d_vecs():
r = utils.repl_vec_over_cols(np.array([1, 2, 3]), 2)
assert utils.np_array_same(r, expected)
def it_matches_beyond_100_epsilon():
assert not utils.np_array_same(
[1 + 100 * np.finfo(np.float32).eps, np.nan], [1, np.nan]
)
def it_matches_epsilon():
assert utils.np_array_same([1 + np.finfo(np.float32).eps, np.nan], [1, np.nan])
def it_matches_nan():
assert utils.np_array_same([1, np.nan], [1, np.nan])
def it_can_edman_fail():
# Failure means Edman did not happen; p_edman_failure=1.0 and we no change
result = sim_v1_worker._step_3a_cycle_edman(samples,
is_mock=False,
p_edman_failure=1.0)
assert np_array_same(samples, result)
def it_fail_to_detach():
result = sim_v1_worker._step_3c_detach(samples, 0.0)
assert np_array_same(samples, result)
def it_sums():
# fmt: off
# Using a huge array here to make sure all my axes
# are getting set properly
evolution = npf([
[
[[1, 1, 0], [0, 1, 0]],
[[0, 1, 1], [0, 0, 1]],
[[1, 1, 1], [1, 0, 1]],
[[1, 1, 1], [1, 0, 1]],
],
[
[[n, 1, 0], [n, 1, 0]],
[[n, 1, 1], [n, 0, 1]],
[[n, n, 1], [n, n, 1]],
[[n, 1, 1], [n, 0, 1]],
],
[
[[n, 1, 0], [n, 1, 0]],
[[n, 1, 1], [n, 0, 1]],
[[n, n, 1], [n, n, 1]],
[[n, 1, 1], [n, 0, 1]],
],
[
[[n, 1, 0], [n, 1, 0]],
[[n, 1, 1], [n, 0, 1]],
[[n, n, 1], [n, n, 1]],
[[n, 1, 1], [n, 0, 1]],
],
[
[[n, 3, 0], [n, 4, 0]],
[[n, 1, 5], [n, 0, 1]],
[[n, n, 9], [n, n, 1]],
[[n, 1, 1], [n, 0, 2]],
],
])
dyemat = sim_v1_worker._step_4_make_dyemat(evolution)
# The axes are now rearranged so that there's
# 4 samples, 2 channels, 5 cycles
# I've thrown enough noise in to make sure I've got
# the dimensions as I want
expected = npf([
[
[2, 1, 1, 1, 3],
[1, 1, 1, 1, 4],
],
[
[2, 2, 2, 2, 6],
[1, 1, 1, 1, 1],
],
[
[3, 1, 1, 1, 9],
[2, 1, 1, 1, 1],
],
[
[3, 2, 2, 2, 2],
[2, 1, 1, 1, 2],
],
])
# fmt: on
assert np_array_same(dyemat, expected)
def it_makes_remainder():
assert np_array_same(remainder_flu, [1, 2])
def it_np_signal_radmat():
assert np_array_same(res.signal_radmat(), props.signal_radmat)
def it_np_noise_radmat():
assert np_array_same(res.noise_radmat(), props.noise_radmat)
def it_computes_zero_pr():
p, r, s, a = cb_all_wrong.pr_curve(n_steps=2)
assert utils.np_array_same([p, r, s, a], zero_pr_result)
