def test_invalid_ranks(data, init_signal_ranks, individual_ranks):
# invalid number of samples
dat = data["diff_views"]
with pytest.raises(ValueError):
_ = AJIVE(init_signal_ranks=init_signal_ranks).fit_transform(dat)
with pytest.raises(ValueError):
_ = AJIVE(individual_ranks=individual_ranks).fit_transform(dat)
def test_indiv_rank_0(data):
dat = data["diff_views"]
ajive = AJIVE(init_signal_ranks=[2, 2], individual_ranks=[0, 0])
_ = ajive.fit_transform(dat)
Is = ajive.individual_mats_
assert_allclose(Is[0], 0)
assert_allclose(Is[1], 0)
def test_signal_rank_warning():
# Throws warning signal rank is larger than possible rank, sets to max-1
ajive = AJIVE(init_signal_ranks=[2, 2], joint_rank=2)
a = np.vstack([[1, 1], [1, 1]])
with pytest.warns(RuntimeWarning):
ajive.fit([a, a])
assert ajive.init_signal_ranks_ == [1, 1]
def test_wrong_signal_ranks(data):
# rank < 0
dat = data["diff_views"]
with pytest.raises(ValueError):
_ = AJIVE(init_signal_ranks=[-1, -4]).fit_transform(dat)
with pytest.raises(ValueError):
_ = AJIVE(init_signal_ranks=[max(dat[0].shape) +
1, -4]).fit_transform(dat)
def test_same_indiv(data, init_signal_ranks, individual_ranks, joint_rank):
# Test same indiv input result across varying inputs
dat = data["same_views"]
ajive = AJIVE(
init_signal_ranks=init_signal_ranks,
joint_rank=joint_rank,
individual_ranks=individual_ranks,
)
ajive = ajive.fit(Xs=dat)
Is = ajive.individual_mats_
assert_allclose(Is[0], Is[1])
def test_same_joint(data, init_signal_ranks, individual_ranks, joint_rank):
# Test same joint input result across varying inputs
dat = data["same_views"]
ajive = AJIVE(
init_signal_ranks=init_signal_ranks,
joint_rank=joint_rank,
individual_ranks=individual_ranks,
)
Js = ajive.fit_transform(Xs=dat)
for i in np.arange(20):
j = np.sum(Js[0][i] == Js[1][i])
assert_equal(j, 100)
def test_fit_elbows():
n = 10
elbows = 3
np.random.seed(1)
x = np.random.binomial(1, 0.6, (n**2)).reshape(n, n)
xorth = orth(x)
d = np.zeros(xorth.shape[0])
for i in range(0, len(d), int(len(d) / (elbows + 1))):
d[:i] += 10
X = xorth.T.dot(np.diag(d)).dot(xorth)
Xs = [X, X]
ajive = AJIVE(n_elbows=2)
ajive = ajive.fit(Xs)
assert_equal(ajive.init_signal_ranks_[0], 4)
def test_random_state(data):
# Tests reproducible simulations
dat = data["same_views"]
ajive1 = AJIVE(init_signal_ranks=[2, 2], random_state=0)
ajive1 = ajive1.fit(Xs=dat)
ajive2 = AJIVE(init_signal_ranks=[2, 2], random_state=0)
ajive2 = ajive2.fit(Xs=dat)
assert_allclose(ajive1.wedin_samples_, ajive2.wedin_samples_)
assert_allclose(ajive1.random_sv_samples_, ajive2.random_sv_samples_)
def test_joint_rank_0(data):
dat = data["diff_views"]
ajive = AJIVE(init_signal_ranks=[2, 2], joint_rank=0)
Js = ajive.fit_transform(dat)
assert_allclose(Js[0], 0)
assert_allclose(Js[1], 0)
def ajive(data):
aj = AJIVE(init_signal_ranks=[3, 2], random_state=0)
joint_mats = aj.fit_transform(Xs=data['diff_views'])
aj.joint_mats = joint_mats
return aj
def test_invalid_samples(n_wedin_samples, n_randdir_samples):
# invalid number of samples
with pytest.raises(ValueError):
_ = AJIVE(n_wedin_samples=n_wedin_samples)
with pytest.raises(ValueError):
_ = AJIVE(n_randdir_samples=n_randdir_samples)
def test_indiv_rank(data):
dat = data["same_views"]
ajive = AJIVE(init_signal_ranks=[2, 2], individual_ranks=[2, 1])
ajive = ajive.fit(Xs=dat)
assert_equal(ajive.individual_ranks_[0], 2)
def test_signal_ranks_None():
# Both init rank inputs are None
with pytest.raises(ValueError):
_ = AJIVE(init_signal_ranks=None, n_elbows=None)
def test_zero_rank_warn(data):
# warn returning rank 0 joint
dat = data["diff_views"]
ajive = AJIVE(init_signal_ranks=[2, 2], joint_rank=0)
with pytest.warns(RuntimeWarning):
_ = ajive.fit_transform(dat)
def test_check_joint_rank_large():
# Joint rank < sum(init_signal_ranks)
with pytest.raises(ValueError):
_ = AJIVE(init_signal_ranks=[2, 2], joint_rank=5)
def test_joint_rank(data):
dat = data["same_views"]
ajive = AJIVE(init_signal_ranks=[2, 2], joint_rank=2)
ajive = ajive.fit(Xs=dat)
assert_equal(ajive.joint_rank, 2)
axes[0].set_ylabel('MVMDS component 2')
axes[0].set_title('Multiview Kmeans Clusters')
plt.tight_layout()
plt.show()
###############################################################################
# Decomposition using AJIVE
# -------------------------
#
# We can also apply joint decomposition tools to find features across views
# that are jointly related. Using :class:`mvlearn.decomposition.AJIVE`, we can
# find genes and lipids that are jointly related.
from mvlearn.decomposition import AJIVE # noqa: E402
ajive = AJIVE()
Xs_joint = ajive.fit_transform(Xs)
f, axes = plt.subplots(2, 1, figsize=(5, 5))
sort_idx = np.hstack((np.argsort(y[:20, 1]), np.argsort(y[20:, 1]) + 20))
y_ticks = [
diet_names[j] + f' ({genotype_names[i]})' if idx % 4 == 0 else ''
for idx, (i, j) in enumerate(y[sort_idx])
]
gene_ticks = [
n if i in [31, 36, 76, 94] else ''
for i, n in enumerate(dataset['gene_feature_names'])
]
g = sns.heatmap(Xs_joint[0][sort_idx],
yticklabels=y_ticks,
vars1 = S**2 / np.sum(S**2)
U, S, V = np.linalg.svd(X2)
vars2 = S**2 / np.sum(S**2)
axes[0].plot(np.arange(10) + 1, vars1[:10], 'ro-', linewidth=2)
axes[1].plot(np.arange(10) + 1, vars2[:10], 'ro-', linewidth=2)
axes[0].set_title('Scree Plot View 1')
axes[1].set_title('Scree Plot View 2')
axes[0].set_xlabel('Number of top singular values')
axes[1].set_xlabel('Number of top singular values')
axes[0].set_ylabel('Percent variance explained')
plt.show()
# Based on the scree plots, we fit AJIVE with both initial signal ranks set to
# 2.
ajive1 = AJIVE(init_signal_ranks=[2, 2])
Js_1 = ajive1.fit_transform(Xs_same)
ajive2 = AJIVE(init_signal_ranks=[2, 2])
Js_2 = ajive2.fit_transform(Xs_diff)
###############################################################################
# Heatmap Visualizations
# ----------------------
#
# Here we are using heatmaps to visualize the decomposition of our views. As we
# can see when we use two of the same views there is no Individualized
# Variation displayed. When we create two different views, the algorithm finds
# different decompositions where common and individual structural artifacts
# can be seen in their corresponding heatmaps.
def test_same_joint_indiv_length(data):
dat = data["same_views"]
ajive = AJIVE(init_signal_ranks=[2, 2])
Js = ajive.fit_transform(Xs=dat)
Is = ajive.individual_mats_
assert_equal(Js[0].shape, Is[0].shape)