def test_vectorize_counts_and_tree(self):
t = TreeNode.read(io.StringIO("((a:1, b:2)c:3)root;"))
counts = np.array([[0, 1], [1, 5], [10, 1]])
count_array, indexed, branch_lengths = \
_vectorize_counts_and_tree(counts, np.array(['a', 'b']), t)
exp_counts = np.array([[0, 1, 10], [1, 5, 1], [1, 6, 11], [1, 6, 11]])
npt.assert_equal(count_array, exp_counts.T)
def test_vectorize_counts_and_tree(self):
t = TreeNode.read(io.StringIO("((a:1, b:2)c:3)root;"))
counts = np.array([[0, 1], [1, 5], [10, 1]])
count_array, indexed, branch_lengths = \
_vectorize_counts_and_tree(counts, np.array(['a', 'b']), t)
exp_counts = np.array([[0, 1, 10], [1, 5, 1], [1, 6, 11], [1, 6, 11]])
npt.assert_equal(count_array, exp_counts.T)
def _setup_multiple_unifrac(counts, otu_ids, tree, validate):
if validate:
_validate_otu_ids_and_tree(counts[0], otu_ids, tree)
counts_by_node, tree_index, branch_lengths = _vectorize_counts_and_tree(counts, otu_ids, tree)
return counts_by_node, tree_index, branch_lengths
def _setup_multiple_unifrac(counts, otu_ids, tree, validate):
if validate:
_validate_otu_ids_and_tree(counts[0], otu_ids, tree)
counts_by_node, tree_index, branch_lengths = \
_vectorize_counts_and_tree(counts, otu_ids, tree)
return counts_by_node, tree_index, branch_lengths
def _setup_faith_pd(counts, otu_ids, tree, validate, single_sample):
if validate:
if single_sample:
# only validate count if operating in single sample mode, they
# will have already been validated otherwise
counts = _validate_counts_vector(counts)
_validate_otu_ids_and_tree(counts, otu_ids, tree)
else:
_validate_otu_ids_and_tree(counts[0], otu_ids, tree)
counts_by_node, tree_index, branch_lengths = _vectorize_counts_and_tree(counts, otu_ids, tree)
return counts_by_node, branch_lengths
def _setup_faith_pd(counts, otu_ids, tree, validate, single_sample):
if validate:
if single_sample:
# only validate count if operating in single sample mode, they
# will have already been validated otherwise
counts = _validate_counts_vector(counts)
_validate_otu_ids_and_tree(counts, otu_ids, tree)
else:
_validate_otu_ids_and_tree(counts[0], otu_ids, tree)
counts_by_node, tree_index, branch_lengths = \
_vectorize_counts_and_tree(counts, otu_ids, tree)
return counts_by_node, branch_lengths
def _run_unifrac(counts, otu_ids, pruned_phylo, method):
_, tree_index, branch_lengths = \
_vectorize_counts_and_tree(counts[0, :], otu_ids, pruned_phylo)
if method == 'weighted_unifrac':
features = _weighted_unifrac_features(counts, otu_ids, pruned_phylo,
tree_index, branch_lengths)
elif method == 'unweighted_unifrac':
features = _unweighted_unifrac_features(counts, branch_lengths,
otu_ids, pruned_phylo)
else:
raise ValueError(
'Method not implemented. Options are '
'weighted_unifrac or unweighted_unifrac, but', method, 'is given')
return features, tree_index
def _setup_pairwise_unifrac(u_counts, v_counts, otu_ids, tree, validate, normalized, unweighted):
if validate:
_validate(u_counts, v_counts, otu_ids, tree)
# temporarily store u_counts and v_counts in a 2-D array as that's what
# _vectorize_counts_and_tree takes
u_counts = np.asarray(u_counts)
v_counts = np.asarray(v_counts)
counts = np.vstack([u_counts, v_counts])
counts_by_node, tree_index, branch_lengths = _vectorize_counts_and_tree(counts, otu_ids, tree)
# unpack counts vectors for single pairwise UniFrac calculation
u_node_counts = counts_by_node[0]
v_node_counts = counts_by_node[1]
u_total_count = u_counts.sum()
v_total_count = v_counts.sum()
return (u_node_counts, v_node_counts, u_total_count, v_total_count, tree_index)
def _setup_pairwise_unifrac(u_counts, v_counts, otu_ids, tree, validate,
normalized, unweighted):
if validate:
_validate(u_counts, v_counts, otu_ids, tree)
# temporarily store u_counts and v_counts in a 2-D array as that's what
# _vectorize_counts_and_tree takes
u_counts = np.asarray(u_counts)
v_counts = np.asarray(v_counts)
counts = np.vstack([u_counts, v_counts])
counts_by_node, tree_index, branch_lengths = \
_vectorize_counts_and_tree(counts, otu_ids, tree)
# unpack counts vectors for single pairwise UniFrac calculation
u_node_counts = counts_by_node[0]
v_node_counts = counts_by_node[1]
u_total_count = u_counts.sum()
v_total_count = v_counts.sum()
return (u_node_counts, v_node_counts, u_total_count, v_total_count,
tree_index)
