def test_common_intervals_len_5_c(alg):
sig1 = (0, 1, 2, 3, 4)
sig2 = (0, 2, 4, 1, 3)
commons = [
# len 4
CommonInterval((1, 4), (1, 4)),
# len 5
CommonInterval((0, 4), (0, 4)),
]
test_results(commons, alg, sig1, sig2)
def test_common_intervals_3_strings_len_5(alg):
pi_1 = [0, 1, 2, 3, 4]
pi_2 = [1, 0, 3, 4, 2]
pi_3 = [1, 4, 3, 0, 2]
commons = [
# len 2
CommonInterval((3, 4), (2, 3), (1, 2)),
# len 5
CommonInterval((0, 4), (0, 4), (0, 4)),
]
test_results(commons, alg, pi_1, pi_2, pi_3)
def test_common_intervals_len_4(alg):
sig1 = (0, 1, 2, 3)
sig2 = (3, 0, 2, 1)
commons = [
# len 2
CommonInterval((1, 2), (2, 3)),
# len 3
CommonInterval((0, 2), (1, 3)),
# len 4
CommonInterval((0, 3), (0, 3)),
]
test_results(commons, alg, sig1, sig2)
def common_k(perm_id: Sequence, perms: Sequence[Sequence], sub_len: int):
k_index = {} # index will be lazily updated on request
for w1 in window(perm_id, sub_len):
charset = frozenset(w1)
if in_others := find_in_others(k_index, charset, perms[1:]):
yield CommonInterval((w1[0], w1[-1]), *in_others)
def trivial_common(perm_a: Sequence, perm_b: Sequence) -> List[CommonInterval]:
commons = []
for l in range(2, len(perm_a) + 1):
for w1 in window(perm_a, l):
chars1 = set(w1)
if in_other := find_in_other(chars1, perm_b):
commons.append(CommonInterval((w1[0], w1[-1]), in_other))
def _intersects_with(cls, intervals: CommonIntervals,
ci: CommonInterval) -> IntervalList:
"""Find intervals that intersect with ci and are 'After' ci"""
return [
other for other in intervals
if CommonInterval.intersect(ci, other)
and ci.first_end <= other.first_end
]
def test_common_intervals_len_5_b(alg):
sig1 = (0, 1, 2, 3, 4)
sig2 = (3, 1, 4, 0, 2)
commons = [
# len 5
CommonInterval((0, 4), (0, 4)),
]
test_results(commons, alg, sig1, sig2)
def test_common_intervals_len_5(alg):
sig1 = (0, 1, 2, 3, 4)
sig2 = (4, 3, 0, 2, 1)
commons = [
# len 2
CommonInterval((1, 2), (3, 4)),
CommonInterval((3, 4), (0, 1)),
# len 3
CommonInterval((0, 2), (2, 4)),
# len 4
CommonInterval((0, 3), (1, 4)),
# len 5
CommonInterval((0, 4), (0, 4)),
]
test_results(commons, alg, sig1, sig2)
def chain_starting_with(self, ci: CommonInterval) -> List[CommonInterval]:
lvl = self.reverse_index[ci]
chain = []
ci_index = self.nesting_levels[lvl].index(ci)
items_after_ci = self.nesting_levels[lvl][ci_index + 1:]
cur = ci
for other in items_after_ci:
if CommonInterval.intersect(cur, other):
chain.append(other)
cur = other
else:
break
return [ci] + chain if chain else []
def process_merged_chars(cls, raw_tree: PQTree,
translation: Tuple[Mapping[Tuple[ContextChar, ContextChar], MergedChar]],
perms):
assert len(translation) == 1
tree_copy = deepcopy(raw_tree)
reverse_translation = invert_dict_multi_val(translation[0])
for node, parent in tree_copy:
if not isinstance(node, LeafNode):
continue
if not isinstance(node.ci.alt_sign, MergedChar):
continue
mc: MergedChar = node.ci.alt_sign
cc1, cc2 = reverse_translation[mc][0]
first_ci_start = node.ci.first_start
first_leaf_ci = CommonInterval((first_ci_start, first_ci_start))
first_leaf_ci.sign = cc1.char
second_leaf_ci = CommonInterval((first_ci_start + 1, first_ci_start + 1))
second_leaf_ci.sign = cc2.char
first_leaf = LeafNode(first_leaf_ci)
second_leaf = LeafNode(second_leaf_ci)
new_qnode = QNode((first_ci_start, first_ci_start + 1)).with_children((first_leaf, second_leaf))
# if parent is a P node just accept children as leafs,
# same behaviour in case the chars always show up in the same order
# otherwise add an Q node as a child - and split the node as its leaf children
# if parent is a P node - transform merged char to child QNode
# in case the parent is Q - if the merged char is first or last in children order
# try lift leafs to parent -
# otherwise, the merged char in the middle of its parent - same as P case
# if isinstance(parent, PNode) or len(mc.char_orders) == 1:
# parent.replace_child(node, first_leaf, second_leaf)
# else:
# parent.replace_child(node,
# QNode((first_ci_start, first_ci_start + 1)).with_children(
# (first_leaf, second_leaf)))
if isinstance(parent, PNode):
parent.replace_child(node, new_qnode)
# elif node in [parent.children[0], parent.children[-1]] and mc.other_side_same:
# parent.replace_child(node, first_leaf, second_leaf)
# elif mc.context_same:
# parent.replace_child(node, first_leaf, second_leaf)
# else:
# parent.replace_child(node, new_qnode)
elif cls.possible_raise_merged_char(perms, node, first_leaf, second_leaf, parent):
parent.replace_child(node, first_leaf, second_leaf)
else:
parent.replace_child(node, new_qnode)
return tree_copy
def known_example():
"""
the one from:
Gene Proximity Analysis across Whole Genomes via PQ Trees1
"""
commons1 = [
# len 1
*[CommonInterval((i, i)) for i in range(9)],
# len 2
CommonInterval((0, 1), (6, 7), (0, 1)),
CommonInterval((1, 2), (7, 8), (1, 2)),
CommonInterval((3, 4), (2, 3), (5, 6)),
CommonInterval((4, 5), (3, 4), (6, 7)),
# len 3
CommonInterval((0, 2), (6, 8), (0, 2)),
CommonInterval((3, 5), (2, 4), (5, 7)),
# len 4
CommonInterval((3, 6), (2, 5), (4, 7)),
# len 5
CommonInterval((3, 7), (1, 5), (3, 7)),
# len 6
CommonInterval((3, 8), (0, 5), (3, 8)),
# len 8
CommonInterval((0, 7), (1, 8), (0, 7)),
# len 9
CommonInterval((0, 8), (0, 8), (0, 8)),
]
ir_intervals = ReduceIntervals.reduce(commons1)
s = IntervalHierarchy.from_irreducible_intervals(ir_intervals)
pqtree = PQTreeBuilder._from_s(s, None)
assert pqtree.to_parens() == "[[0 1 2] [[[3 4 5] 6] 7] 8]"
def test_common_intervals_3_strings_len_9(alg):
pi_1 = [0, 1, 2, 3, 4, 5, 6, 7, 8]
pi_2 = [8, 7, 3, 4, 5, 6, 0, 1, 2]
pi_3 = [0, 1, 2, 7, 6, 3, 4, 5, 8]
commons = [
# len 2
CommonInterval((0, 1), (6, 7), (0, 1)),
CommonInterval((1, 2), (7, 8), (1, 2)),
CommonInterval((3, 4), (2, 3), (5, 6)),
CommonInterval((4, 5), (3, 4), (6, 7)),
# len 3
CommonInterval((0, 2), (6, 8), (0, 2)),
CommonInterval((3, 5), (2, 4), (5, 7)),
# len 4
CommonInterval((3, 6), (2, 5), (4, 7)),
# len 5
CommonInterval((3, 7), (1, 5), (3, 7)),
# len 6
CommonInterval((3, 8), (0, 5), (3, 8)),
# len 8
CommonInterval((0, 7), (1, 8), (0, 7)),
# len 9
CommonInterval((0, 8), (0, 8), (0, 8)),
]
test_results(commons, alg, pi_1, pi_2, pi_3)
def test_common_intervals_len_9(alg):
sig1 = (0, 1, 2, 3, 4, 5, 6, 7, 8)
sig2 = (8, 7, 3, 4, 5, 6, 0, 1, 2)
commons = [
# len 2
CommonInterval((0, 1), (6, 7)),
CommonInterval((1, 2), (7, 8)),
CommonInterval((3, 4), (2, 3)),
CommonInterval((4, 5), (3, 4)),
CommonInterval((5, 6), (4, 5)),
CommonInterval((7, 8), (0, 1)),
# len 3
CommonInterval((0, 2), (6, 8)),
CommonInterval((3, 5), (2, 4)),
CommonInterval((4, 6), (3, 5)),
# len 4
CommonInterval((3, 6), (2, 5)),
# len 4
CommonInterval((3, 7), (1, 5)),
# len 6
CommonInterval((3, 8), (0, 5)),
# len 7
CommonInterval((0, 6), (2, 8)),
# len 8
CommonInterval((0, 7), (1, 8)),
# len 9
CommonInterval((0, 8), (0, 8)),
]
test_results(commons, alg, sig1, sig2)
def test_given_examples():
# C = {[1, 2], [2, 3], [1, 3], [4, 5], [5, 6], [4, 6], [4, 7], [4, 8], [4, 9], [1, 8], [1, 9]}
# I = {[1, 2], [2, 3], [4, 5], [5, 6], [4, 7], [4, 8], [4, 9], [1, 8] }
commons1 = [
# len 2
CommonInterval((0, 1), (6, 7), (0, 1)),
CommonInterval((1, 2), (7, 8), (1, 2)),
CommonInterval((3, 4), (2, 3), (5, 6)),
CommonInterval((4, 5), (3, 4), (6, 7)),
# len 3
CommonInterval((0, 2), (6, 8), (0, 2)),
CommonInterval((3, 5), (2, 4), (5, 7)),
# len 4
CommonInterval((3, 6), (2, 5), (4, 7)),
# len 5
CommonInterval((3, 7), (1, 5), (3, 7)),
# len 6
CommonInterval((3, 8), (0, 5), (3, 8)),
# len 8
CommonInterval((0, 7), (1, 8), (0, 7)),
# len 9
CommonInterval((0, 8), (0, 8), (0, 8)),
]
irreducible1 = {
# len 2
CommonInterval((0, 1), (6, 7), (0, 1)),
CommonInterval((1, 2), (7, 8), (1, 2)),
CommonInterval((3, 4), (2, 3), (5, 6)),
CommonInterval((4, 5), (3, 4), (6, 7)),
# len 4
CommonInterval((3, 6), (2, 5), (4, 7)),
# len 5
CommonInterval((3, 7), (1, 5), (3, 7)),
# len 6
CommonInterval((3, 8), (0, 5), (3, 8)),
# len 8
CommonInterval((0, 7), (1, 8), (0, 7)),
}
commons2 = [
# len 2
CommonInterval((0, 1)),
CommonInterval((1, 2)),
# len 3
CommonInterval((0, 2)),
# len 5
CommonInterval((2, 7)),
CommonInterval((3, 8)),
CommonInterval((4, 9)),
# len 7
CommonInterval((0, 7)),
# len 8
CommonInterval((0, 8)),
# len 9
CommonInterval((0, 9)),
]
irreducible2 = {
# len 2
CommonInterval((0, 1)),
CommonInterval((1, 2)),
# len 5
CommonInterval((2, 7)),
CommonInterval((3, 8)),
CommonInterval((4, 9)),
}
assert ReduceIntervals.reduce(commons1) == irreducible1
assert ReduceIntervals.reduce(commons2) == irreducible2
def produce_s_for_example():
commons1 = [
# len 1
*[CommonInterval((i, i)) for i in range(9)],
# len 2
CommonInterval((0, 1), (6, 7), (0, 1)),
CommonInterval((1, 2), (7, 8), (1, 2)),
CommonInterval((3, 4), (2, 3), (5, 6)),
CommonInterval((4, 5), (3, 4), (6, 7)),
# len 3
CommonInterval((0, 2), (6, 8), (0, 2)),
CommonInterval((3, 5), (2, 4), (5, 7)),
# len 4
CommonInterval((3, 6), (2, 5), (4, 7)),
# len 5
CommonInterval((3, 7), (1, 5), (3, 7)),
# len 6
CommonInterval((3, 8), (0, 5), (3, 8)),
# len 8
CommonInterval((0, 7), (1, 8), (0, 7)),
# len 9
CommonInterval((0, 8), (0, 8), (0, 8)),
]
ir_intervals = ReduceIntervals.reduce(commons1)
s = IntervalHierarchy.from_irreducible_intervals(ir_intervals)
pprint(s.nesting_levels)
def common_k(perm_id: Sequence, perms: Sequence[Sequence], k: int):
for w1 in window(perm_id, k):
chars1 = set(w1)
if in_others := find_in_others(chars1, perms[1:]):
yield CommonInterval((w1[0], w1[-1]), *in_others)