def test_collapse_taxa(self, stool_biom):
otu_inst = Otu(stool_biom)
otu_collapse, children_map = otu_inst.collapse_taxa("Family")
family_members = set(otu_inst.obs_metadata.Family)
assert len(family_members) == otu_collapse.otu_data.shape[0]
assert otu_inst.otu_data.shape[1] == otu_collapse.otu_data.shape[1]
assert family_members == set(otu_collapse.obs_metadata.Family)
group_dict = otu_inst.obs_metadata.groupby("Family").groups
assert sorted(list(i) for i in group_dict.values()) == sorted(
list(i) for i in children_map.values())
def test_rm_sparse_samples(self, stool_biom):
otu_inst = Otu(stool_biom)
rm_samples_otu = otu_inst.rm_sparse_samples()
assert otu_inst.otu_data.shape[0] == rm_samples_otu.otu_data.shape[0]
assert otu_inst.otu_data.shape[1] > rm_samples_otu.otu_data.shape[1]
assert otu_inst.otu_data.shape[1] - rm_samples_otu.otu_data.shape[
1] == 1
norm_otu = otu_inst.normalize()
with pytest.raises(ValueError):
norm_otu.rm_sparse_samples()
def test_filter(self, stool_biom):
otu_inst = Otu(stool_biom)
query = "Firmicutes"
func = lambda values, id_, md: Lineage(**md).Phylum == query
md = otu_inst.obs_metadata
ind = md.index[md.Phylum == query]
otu_filtered = otu_inst.filter(func=func)
assert otu_filtered.otu_data.shape[1] == otu_inst.otu_data.shape[1]
assert otu_filtered.otu_data.shape[0] < otu_inst.otu_data.shape[0]
assert set(ind) == set(otu_filtered.otu_data.ids("observation"))
def test_partition(self, stool_biom):
otu_inst = Otu(stool_biom)
func = lambda id_, md: Lineage(**md).get_superset("Phylum")
md = otu_inst.obs_metadata
gen = otu_inst.partition(axis="observation", func=func)
partition_dict = {k.name[1]: v for k, v in gen}
assert set(partition_dict) == set(md.Phylum)
assert len(set(v.otu_data.shape[1]
for v in partition_dict.values())) == 1
assert (sum(
v.otu_data.shape[0]
for v in partition_dict.values()) == otu_inst.otu_data.shape[0])
def test_normalize(self, stool_biom):
otu_inst = Otu(stool_biom)
sample_norm = otu_inst.normalize()
assert (otu_inst.otu_data.to_dataframe().sum(axis=0) > 1).all()
assert np.isclose(sample_norm.otu_data.to_dataframe().sum(axis=0),
1.0).all()
assert sample_norm.is_norm(axis="sample")
obs_norm = otu_inst.normalize(axis="observation")
assert (otu_inst.otu_data.to_dataframe().sum(axis=1) > 1).all()
assert obs_norm.is_norm(axis="observation")
assert np.isclose(obs_norm.otu_data.to_dataframe().sum(axis=1),
1.0).all()
with pytest.raises(ValueError):
otu_inst.normalize(method="random_method")
with pytest.raises(NotImplementedError):
otu_inst.normalize(method="css")
with pytest.raises(NotImplementedError):
otu_inst.normalize(method="rarefy")
def test_write(self, stool_biom, tmpdir):
otu_inst = Otu(stool_biom)
fol = tmpdir.mkdir("results")
otu_inst.write("biom_test", str(fol))
otu_load1 = Otu.load_data(fol.join("biom_test.biom"))
assert otu_inst.otu_data.shape == otu_load1.otu_data.shape
otu_inst.write("tsv_test", str(fol), "tsv")
otu_load2 = Otu.load_data(
fol.join("tsv_test_otu.tsv"),
fol.join("tsv_test_sample_metadata.tsv"),
fol.join("tsv_test_obs_metadata.csv"),
dtype="tsv",
)
assert otu_inst.otu_data.shape == otu_load2.otu_data.shape
def test_rm_sparse_obs(self, stool_biom):
otu_inst = Otu(stool_biom)
rm_obs_otu = otu_inst.rm_sparse_obs(prevalence_thres=0.4)
assert otu_inst.otu_data.shape[1] == rm_obs_otu.otu_data.shape[1]
assert otu_inst.otu_data.shape[0] >= rm_obs_otu.otu_data.shape[0]
assert otu_inst.otu_data.shape[0] - rm_obs_otu.otu_data.shape[0] == 10
def test_load_data_tsv(self, tsv_files):
for otu, sample, tax in tsv_files["good"]:
otu_inst = Otu.load_data(otu, sample, tax, dtype="tsv")
assert hasattr(otu_inst, "otu_data")
assert hasattr(otu_inst, "sample_metadata")
assert hasattr(otu_inst, "obs_metadata")
def test_load_data_biom(self, biom_files):
for biom in biom_files["good"]:
otu_inst = Otu.load_data(biom)
assert hasattr(otu_inst, "otu_data")
assert hasattr(otu_inst, "sample_metadata")
assert hasattr(otu_inst, "obs_metadata")
def test_init(self, biom_data):
for biom in biom_data:
otu_inst = Otu(biom)
assert (otu_inst.otu_data.to_dataframe() == biom.to_dataframe()
).any().any()