def get_response_content(fs):
# read the matrix
D = np.array(fs.matrix)
n = len(D)
# read the ordered labels
ordered_labels = Util.get_stripped_lines(StringIO(fs.labels))
selected_labels = Util.get_stripped_lines(StringIO(fs.selection))
# validate the input
if n != len(ordered_labels):
raise HandlingError(
'the number of taxon labels should match the number of rows in the distance matrix'
)
# get the two sets of indices
index_set_A = set(i for i, label in enumerate(ordered_labels)
if label in selected_labels)
index_set_B = set(range(n)) - index_set_A
# get internal values related to the split
R, alpha, beta, gamma = get_R_alpha_beta_gamma(D, index_set_B)
# get the two new distance matrices
D_A = BuildTreeTopology.update_generalized_nj(D, index_set_B)
D_B = BuildTreeTopology.update_generalized_nj(D, index_set_A)
# get the names associated with the indices of the new distance matrices
all_names = [set([name]) for name in ordered_labels]
D_A_names = [
set_to_string(x) for x in SchurAlgebra.vmerge(all_names, index_set_B)
]
D_B_names = [
set_to_string(x) for x in SchurAlgebra.vmerge(all_names, index_set_A)
]
# show the results
out = StringIO()
print >> out, 'alpha:', alpha
print >> out, 'beta:', beta
print >> out, 'gamma:', gamma
print >> out
print >> out, 'new distance matrix corresponding to the selected names:'
print >> out, MatrixUtil.m_to_string(D_A)
print >> out
print >> out, 'ordered labels corresponding to this matrix:'
for name in D_A_names:
print >> out, name
print >> out
print >> out, 'new distance matrix corresponding to the non-selected names:'
print >> out, MatrixUtil.m_to_string(D_B)
print >> out
print >> out, 'ordered labels corresponding to this matrix:'
for name in D_B_names:
print >> out, name
# return the response
return out.getvalue()
def get_response_content(fs):
# read the matrix
D = np.array(fs.matrix)
n = len(D)
# read the ordered labels
ordered_labels = Util.get_stripped_lines(StringIO(fs.labels))
selected_labels = Util.get_stripped_lines(StringIO(fs.selection))
# validate the input
if n != len(ordered_labels):
raise HandlingError("the number of taxon labels should match the number of rows in the distance matrix")
# get the two sets of indices
index_set_A = set(i for i, label in enumerate(ordered_labels) if label in selected_labels)
index_set_B = set(range(n)) - index_set_A
# get internal values related to the split
R, alpha, beta, gamma = get_R_alpha_beta_gamma(D, index_set_B)
# get the two new distance matrices
D_A = BuildTreeTopology.update_generalized_nj(D, index_set_B)
D_B = BuildTreeTopology.update_generalized_nj(D, index_set_A)
# get the names associated with the indices of the new distance matrices
all_names = [set([name]) for name in ordered_labels]
D_A_names = [set_to_string(x) for x in SchurAlgebra.vmerge(all_names, index_set_B)]
D_B_names = [set_to_string(x) for x in SchurAlgebra.vmerge(all_names, index_set_A)]
# show the results
out = StringIO()
print >> out, "alpha:", alpha
print >> out, "beta:", beta
print >> out, "gamma:", gamma
print >> out
print >> out, "new distance matrix corresponding to the selected names:"
print >> out, MatrixUtil.m_to_string(D_A)
print >> out
print >> out, "ordered labels corresponding to this matrix:"
for name in D_A_names:
print >> out, name
print >> out
print >> out, "new distance matrix corresponding to the non-selected names:"
print >> out, MatrixUtil.m_to_string(D_B)
print >> out
print >> out, "ordered labels corresponding to this matrix:"
for name in D_B_names:
print >> out, name
# return the response
return out.getvalue()
def _do_analysis(self, use_generalized_nj):
"""
Do some splits of the tree.
@param use_generalized_nj: True if we use an old method of outgrouping
"""
# define the distance matrix
D = np.array(self.pruned_tree.get_distance_matrix(self.pruned_names))
# get the primary split of the criterion matrix
L = Euclid.edm_to_laplacian(D)
v = BuildTreeTopology.laplacian_to_fiedler(L)
eigensplit = BuildTreeTopology.eigenvector_to_split(v)
# assert that the first split cleanly separates the bacteria from the rest
left_indices, right_indices = eigensplit
left_domains = self._get_domains([self.pruned_names[x] for x in left_indices])
right_domains = self._get_domains([self.pruned_names[x] for x in right_indices])
if ('bacteria' in left_domains) and ('bacteria' in right_domains):
raise HandlingError('bacteria were not defined by the first split')
# now we have enough info to define the first supplementary csv file
self.first_split_object = SupplementarySpreadsheetObject(self.pruned_names, L, v)
# define the bacteria indices vs the non-bacteria indices for the second split
if 'bacteria' in left_domains:
bacteria_indices = left_indices
non_bacteria_indices = right_indices
elif 'bacteria' in right_domains:
bacteria_indices = right_indices
non_bacteria_indices = left_indices
# get the secondary split of interest
if use_generalized_nj:
D_secondary = BuildTreeTopology.update_generalized_nj(D, bacteria_indices)
L_secondary = Euclid.edm_to_laplacian(D_secondary)
else:
L_secondary = SchurAlgebra.mmerge(L, bacteria_indices)
full_label_sets = [set([i]) for i in range(len(self.pruned_names))]
next_label_sets = SchurAlgebra.vmerge(full_label_sets, bacteria_indices)
v_secondary = BuildTreeTopology.laplacian_to_fiedler(L_secondary)
eigensplit_secondary = BuildTreeTopology.eigenvector_to_split(v_secondary)
left_subindices, right_subindices = eigensplit_secondary
pruned_names_secondary = []
for label_set in next_label_sets:
if len(label_set) == 1:
label = list(label_set)[0]
pruned_names_secondary.append(self.pruned_names[label])
else:
pruned_names_secondary.append('all-bacteria')
# assert that the second split cleanly separates the eukaryota from the rest
left_subdomains = self._get_domains([pruned_names_secondary[x] for x in left_subindices])
right_subdomains = self._get_domains([pruned_names_secondary[x] for x in right_subindices])
if ('eukaryota' in left_subdomains) and ('eukaryota' in right_subdomains):
raise HandlingError('eukaryota were not defined by the second split')
# now we have enough info to define the second supplementary csv file
self.second_split_object = SupplementarySpreadsheetObject(pruned_names_secondary, L_secondary, v_secondary)
def _do_analysis(self, use_generalized_nj):
"""
Do some splits of the tree.
@param use_generalized_nj: True if we use an old method of outgrouping
"""
# define the distance matrix
D = np.array(self.pruned_tree.get_distance_matrix(self.pruned_names))
# get the primary split of the criterion matrix
L = Euclid.edm_to_laplacian(D)
v = BuildTreeTopology.laplacian_to_fiedler(L)
eigensplit = BuildTreeTopology.eigenvector_to_split(v)
# assert that the first split cleanly separates the bacteria from the rest
left_indices, right_indices = eigensplit
left_domains = self._get_domains(
[self.pruned_names[x] for x in left_indices])
right_domains = self._get_domains(
[self.pruned_names[x] for x in right_indices])
if ('bacteria' in left_domains) and ('bacteria' in right_domains):
raise HandlingError('bacteria were not defined by the first split')
# now we have enough info to define the first supplementary csv file
self.first_split_object = SupplementarySpreadsheetObject(
self.pruned_names, L, v)
# define the bacteria indices vs the non-bacteria indices for the second split
if 'bacteria' in left_domains:
bacteria_indices = left_indices
non_bacteria_indices = right_indices
elif 'bacteria' in right_domains:
bacteria_indices = right_indices
non_bacteria_indices = left_indices
# get the secondary split of interest
if use_generalized_nj:
D_secondary = BuildTreeTopology.update_generalized_nj(
D, bacteria_indices)
L_secondary = Euclid.edm_to_laplacian(D_secondary)
else:
L_secondary = SchurAlgebra.mmerge(L, bacteria_indices)
full_label_sets = [set([i]) for i in range(len(self.pruned_names))]
next_label_sets = SchurAlgebra.vmerge(full_label_sets,
bacteria_indices)
v_secondary = BuildTreeTopology.laplacian_to_fiedler(L_secondary)
eigensplit_secondary = BuildTreeTopology.eigenvector_to_split(
v_secondary)
left_subindices, right_subindices = eigensplit_secondary
pruned_names_secondary = []
for label_set in next_label_sets:
if len(label_set) == 1:
label = list(label_set)[0]
pruned_names_secondary.append(self.pruned_names[label])
else:
pruned_names_secondary.append('all-bacteria')
# assert that the second split cleanly separates the eukaryota from the rest
left_subdomains = self._get_domains(
[pruned_names_secondary[x] for x in left_subindices])
right_subdomains = self._get_domains(
[pruned_names_secondary[x] for x in right_subindices])
if ('eukaryota' in left_subdomains) and ('eukaryota'
in right_subdomains):
raise HandlingError(
'eukaryota were not defined by the second split')
# now we have enough info to define the second supplementary csv file
self.second_split_object = SupplementarySpreadsheetObject(
pruned_names_secondary, L_secondary, v_secondary)
