def runTest(self):
# rooted tree: so clade bitmasks
tree_list = dendropy.TreeList.get_from_stream(
StringIO("""[&R]((t5:0.161175,t6:0.161175):0.392293,((t4:0.104381,(t2:0.075411,t1:0.075411):1):0.065840,t3:0.170221):0.383247);"""),
"newick")
for i in tree_list:
_LOG.debug(i._get_indented_form())
i.encode_bipartitions()
_LOG.debug(i._get_indented_form(splits=True))
i._debug_check_tree(splits=True, logger_obj=_LOG)
root1 = tree_list[0].seed_node
root1e = root1.edge
self.assertEqual(bitprocessing.indexes_of_set_bits(root1e.split_bitmask), list(range(6)))
self.assertEqual(bitprocessing.indexes_of_set_bits(root1e.split_bitmask, one_based=True), list(range(1,7)))
self.assertEqual(bitprocessing.indexes_of_set_bits(root1e.split_bitmask, fill_bitmask=21, one_based=True), [1, 3, 5])
self.assertEqual(bitprocessing.indexes_of_set_bits(root1e.split_bitmask, fill_bitmask=21), [0, 2, 4])
self.assertEqual(bitprocessing.num_set_bits(root1e.split_bitmask), 6)
fc1 = root1.child_nodes()[0]
fc1e = fc1.edge
self.assertEqual(bitprocessing.indexes_of_set_bits(fc1e.split_bitmask), [0, 1])
self.assertEqual(bitprocessing.indexes_of_set_bits(fc1e.split_bitmask, one_based=True), [1, 2])
self.assertEqual(bitprocessing.indexes_of_set_bits(fc1e.split_bitmask, fill_bitmask=0x15, one_based=True), [1])
self.assertEqual(bitprocessing.indexes_of_set_bits(fc1e.split_bitmask, fill_bitmask=0x15), [0])
self.assertEqual(bitprocessing.num_set_bits(fc1e.split_bitmask), 2)
def runTest(self):
# rooted tree: so clade bitmasks
tree_list = dendropy.TreeList.get_from_stream(
StringIO("""[&R]((t5:0.161175,t6:0.161175):0.392293,((t4:0.104381,(t2:0.075411,t1:0.075411):1):0.065840,t3:0.170221):0.383247);"""),
"newick")
for i in tree_list:
_LOG.debug(i._get_indented_form())
i.encode_bipartitions()
_LOG.debug(i._get_indented_form(splits=True))
i._debug_check_tree(splits=True, logger_obj=_LOG)
root1 = tree_list[0].seed_node
root1e = root1.edge
self.assertEqual(bitprocessing.indexes_of_set_bits(root1e.split_bitmask), list(range(6)))
self.assertEqual(bitprocessing.indexes_of_set_bits(root1e.split_bitmask, one_based=True), list(range(1,7)))
self.assertEqual(bitprocessing.indexes_of_set_bits(root1e.split_bitmask, fill_bitmask=21, one_based=True), [1, 3, 5])
self.assertEqual(bitprocessing.indexes_of_set_bits(root1e.split_bitmask, fill_bitmask=21), [0, 2, 4])
self.assertEqual(bitprocessing.num_set_bits(root1e.split_bitmask), 6)
fc1 = root1.child_nodes()[0]
fc1e = fc1.edge
self.assertEqual(bitprocessing.indexes_of_set_bits(fc1e.split_bitmask), [0, 1])
self.assertEqual(bitprocessing.indexes_of_set_bits(fc1e.split_bitmask, one_based=True), [1, 2])
self.assertEqual(bitprocessing.indexes_of_set_bits(fc1e.split_bitmask, fill_bitmask=0x15, one_based=True), [1])
self.assertEqual(bitprocessing.indexes_of_set_bits(fc1e.split_bitmask, fill_bitmask=0x15), [0])
self.assertEqual(bitprocessing.num_set_bits(fc1e.split_bitmask), 2)
def getCentroidEdgeRandom(tree, minBound=5):
fullMask = tree.seed_node.tree_leafset_bitmask
numLeaves = bitprocessing.num_set_bits(fullMask)
candidates = []
for edge in tree.postorder_internal_edge_iter():
if edge.tail_node is None:
continue
mask = edge.bipartition.leafset_bitmask
numMask1 = bitprocessing.num_set_bits(mask)
numMask2 = numLeaves - numMask1
if numMask1 >= minBound and numMask2 >= minBound:
candidates.append(edge)
return np.random.choice(candidates)
def count_bits(a):
deprecate.dendropy_deprecation_warning(
preamble=
"Deprecated since DendroPy 4: 'dendropy.treesplit.count_bits()'.",
old_construct=
"from dendropy import treesplit\nd = treesplit.count_bits(...)",
new_construct=
"from dendropy.utility import bitprocessing\nd = bitprocessing.num_set_bits(...)"
)
return bitprocessing.num_set_bits(a)
def getCentroidEdge(tree):
numLeaves = bitprocessing.num_set_bits(tree.seed_node.tree_leafset_bitmask)
# numLeaves = len(tree.seed_node.leaf_nodes())
bestBalance = float('inf')
# sys.stderr.write("searching for best edge in num leaves:")
# sys.stderr.write(str(numLeaves) + str("\n"))
for edge in tree.postorder_edge_iter():
if edge.tail_node is None:
continue
balance = abs(
numLeaves / 2 -
bitprocessing.num_set_bits(edge.bipartition.leafset_bitmask))
# sys.stderr.write("current_balance:")
# sys.stderr.write(str(balance) + "\n")
if balance < bestBalance:
bestBalance = balance
bestEdge = edge
# sys.stderr.write(str(bestEdge.head_node))
# sys.stderr.write(str(bestEdge.length))
# sys.stderr.write(str(bestEdge.head_node.label))
return bestEdge
def getBestHeuristicEdge(tree, max_subset_size, num_taxa):
num_leaves = bitprocessing.num_set_bits(
tree.seed_node.tree_leafset_bitmask)
current_best_score = -1
current_best_edge = None
subset_L_size = 0
subset_r_size = 0
best_L_score = 0
best_R_score = 0
best_subset_L_size = 0
best_subset_R_size = 0
for edge in tree.postorder_edge_iter():
if edge.tail_node is None:
continue
if edge.head_node.label is not None:
subset_L_size = bitprocessing.num_set_bits(
edge.bipartition.leafset_bitmask)
subset_R_size = num_leaves - bitprocessing.num_set_bits(
edge.bipartition.leafset_bitmask)
# sys.stderr.write(str(subset_L_size) + ":" + str(subset_R_size) + "\n")
current_L_score = heuristic(float(edge.head_node.label),
subset_L_size, max_subset_size,
num_taxa)
current_R_score = heuristic(float(edge.head_node.label),
subset_R_size, max_subset_size,
num_taxa)
if current_best_score < sum([current_L_score, current_R_score]):
best_L_score = current_L_score
best_R_score = current_R_score
best_subset_L_size = subset_L_size
best_subset_R_size = subset_R_size
current_best_score = sum([current_L_score, current_R_score])
current_best_edge = edge
# sys.stderr.write(str(current_best_edge.head_node) + "\n")
# sys.stderr.write(str(current_best_edge.length) + "\n")
# sys.stderr.write(str(current_best_edge.head_node.label) + "\n")
# sys.stderr.write(str(best_subset_L_size) + ":" + str(best_subset_R_size) + "\n")
# sys.stderr.write(str(best_L_score) + ":" + str(best_R_score) + "\n")
return current_best_edge
def getLongestEdge(tree):
numLeaves = bitprocessing.num_set_bits(tree.seed_node.tree_leafset_bitmask)
# numLeaves = len(tree.seed_node.leaf_nodes())
longeth_edge = None
longest_edge_length = 0
for edge in tree.postorder_edge_iter():
if edge.tail_node is None:
continue
current_length = edge.length
if longest_edge_length < current_length:
longest_edge_length = current_length
longest_edge = edge
return longest_edge
def frequency_of_bipartition_inclusive(tree_list, taxon_labels,
return_locus_list):
is_bipartitions_updated = False
split = tree_list.taxon_namespace.taxa_bitmask(labels=taxon_labels)
k = len(taxon_labels)
if bitprocessing.num_set_bits(split) != k:
raise IndexError('Not all taxa could be mapped to bipartition (%s): %s' \
% (tree_list.taxon_namespace.bitmask_as_bitstring(split), k))
found = 0
total = 0
if return_locus_list:
locus_list = []
for tree in tree_list:
tree_labels = [leaf.taxon.label for leaf in tree.leaf_nodes()]
labels_present = [name for name in taxon_labels if name in tree_labels]
modified_split = tree_list.taxon_namespace.taxa_bitmask(
labels=labels_present)
unnormalized_split = modified_split
normalized_split = treemodel.Bipartition.normalize_bitmask(
bitmask=modified_split,
fill_bitmask=tree_list.taxon_namespace.all_taxa_bitmask(),
lowest_relevant_bit=1)
if not is_bipartitions_updated or not tree.bipartition_encoding:
tree.encode_bipartitions()
bipartition_encoding = set(b.split_bitmask
for b in tree.bipartition_encoding)
total += 1
if tree.is_unrooted and (normalized_split in bipartition_encoding):
found += 1
if return_locus_list:
locus_list.append(tree._label)
elif (not tree.is_unrooted) and (unnormalized_split
in bipartition_encoding):
found += 1
if return_locus_list:
locus_list.append(tree._label)
if return_locus_list:
try:
return float(found) / total, locus_list
except ZeroDivisionError:
return 0, 0
else:
try:
return float(found) / total
except ZeroDivisionError:
return 0
def masked_frequency_of_bipartition(self, **kwargs):
"""Adaptation of dendropy.TreeList.frequency_of_bipartition that takes a taxon mask.
This allows identifying splits on a subset of taxa within a larger tree without
pruning any tree structures, which is much slower.
Given a split or bipartition specified as:
- a split bitmask given the keyword 'split_bitmask'
- a list of `Taxon` objects given with the keyword `taxa`
- a list of taxon labels given with the keyword `labels`
- a list of oids given with the keyword `oids`
this function returns the proportion of trees in self in which the
split is found.
"""
partialMask = kwargs["mask"] if "mask" in kwargs else self.taxon_namespace.all_taxa_bitmask()
if "split_bitmask" in kwargs:
targetSplit = kwargs["split_bitmask"]
else:
targetSplit = self.taxon_namespace.get_taxa_bitmask(**kwargs)
k = kwargs.values()[0]
if bitprocessing.num_set_bits(targetSplit) != len(k):
raise IndexError('Not all taxa could be mapped to split (%s): %s'
% (self.taxon_namespace.split_bitmask_string(targetSplit), k))
found = 0
total = 0
for tree in self:
tree.compat_encode_bipartitions()
total += 1
compSplit = (~targetSplit & partialMask)
#for test_split in tree.split_edges:
for test_split in tree.reference_tree.bipartition_encoding:
if not treesplit.is_compatible(test_split, targetSplit, partialMask):
break
masked_test = (test_split & partialMask)
if targetSplit == masked_test or compSplit == masked_test:
found += 1
break
return float(found) / total
def count_bits(a):
deprecate.dendropy_deprecation_warning(
preamble="Deprecated since DendroPy 4: 'dendropy.treesplit.count_bits()'.",
old_construct="from dendropy import treesplit\nd = treesplit.count_bits(...)",
new_construct="from dendropy.utility import bitprocessing\nd = bitprocessing.num_set_bits(...)")
return bitprocessing.num_set_bits(a)
def runTest(self):
self.assertEqual(bitprocessing.num_set_bits(21), 3)
def runTest(self):
self.assertEqual(bitprocessing.num_set_bits(21), 3)
