def delete(self, num: int, node: TreeNode = None):
node = node or self.root
if num < node.val:
node.left = self.delete(num, node.left)
elif num > node.val:
node.right = self.delete(num, node.right)
else:
self.size -= 1
if node.left is None and node.right is None:
if self.root == node:
self.root = None
return None
elif node.left is None:
if self.root == node:
self.root = node.right
return node.right
elif node.right is None:
if self.root == node:
self.root = node.left
return node.left
else:
min_node = node.right
while min_node.left:
min_node = min_node.left
node.val = min_node.val
node.right = self.delete(min_node.val, node.right)
return node
def delete(self, node: TreeNode, val: int):
if node is None:
return
if val < node.val:
node.left = self.delete(node.left, val)
if val > node.val:
node.right = self.delete(node.right, val)
if val == node.val:
if node.left and node.right:
predecessor = node.left
while predecessor.right:
predecessor = predecessor.right
node.val = predecessor.val
predecessor.val = val
node.left = self.delete(node.left, val)
elif node.left:
return node.left
elif node.right:
return node.right
else:
return None
return node
def deserialize(self, data):
"""Decodes your encoded data to tree.
:type data: str
:rtype: TreeNode
"""
if len(data) == 0:
return None
root = TreeNode(data[0])
idx = 1
dq = deque([root])
while dq and idx < len(data):
cur_node = dq.popleft()
if data[idx] is not None:
cur_node.left = TreeNode(data[idx])
dq.append(cur_node.left)
idx += 1
if idx < len(data) and data[idx] is not None:
cur_node.right = TreeNode(data[idx])
dq.append(cur_node.right)
idx += 1
return root
def insertNode(self, node: TreeNode, val: int):
if node.val is None or val > node.val:
if node.right:
self.insertNode(node.right, val)
else:
node.right = TreeNode(val)
elif val < node.val:
if node.left:
self.insertNode(node.left, val)
else:
node.left = TreeNode(val)
def helper(left: int, right: int):
if left > right:
return None
mid = (left + right) // 2
root = TreeNode(nums[mid])
root.left = helper(left, mid - 1)
root.right = helper(mid + 1, right)
return root
def rdeserialize(l: List[str]) -> TreeNode:
if l[0] == "None":
l.pop(0)
return None
root = TreeNode(l[0])
l.pop(0)
root.left = rdeserialize(l)
root.right = rdeserialize(l)
return root
def helper(l_idx: int, r_idx: int) -> TreeNode:
if l_idx > r_idx:
return None
val = preorder.pop(0)
node = TreeNode(val)
idx = idx_map[val]
node.left = helper(l_idx, idx - 1)
node.right = helper(idx + 1, r_idx)
return node
def insert(self, node: TreeNode, val: int):
if node is None or node.val == val:
return
if val < node.val:
if node.left:
self.insert(node.left, val)
else:
node.left = TreeNode(val)
if val > node.val:
if node.right:
self.insert(node.right, val)
else:
node.right = TreeNode(val)
def helper(in_left = 0, in_right = len(inorder)) -> TreeNode:
nonlocal pre_idx
if in_left == in_right:
return None
val = preorder[pre_idx]
node = TreeNode(val)
pre_idx += 1
idx = idx_map[val]
node.left = helper(in_left, idx)
node.right = helper(idx + 1, in_right)
return node
def helper(in_left: int, in_right: int) -> TreeNode:
if in_left > in_right:
return None
# pick up the last element as a root
val = postorder.pop()
root = TreeNode(val)
# root splits inorder list
# into left and right subtree
index = idx_map[val]
# build right subtree
root.right = helper(index + 1, in_right)
# build left subtree
root.left = helper(in_left, index - 1)
return root
def test_bst():
bst = BST()
root = TreeNode(5)
assert inorder_traverse(root) == [5]
bst.insert(root, 2)
assert inorder_traverse(root) == [2, 5]
bst.insert(root, 8)
assert inorder_traverse(root) == [2, 5, 8]
bst.insert(root, 1)
assert inorder_traverse(root) == [1, 2, 5, 8]
bst.insert(root, 6)
assert inorder_traverse(root) == [1, 2, 5, 6, 8]
bst.insert(root, 7)
assert inorder_traverse(root) == [1, 2, 5, 6, 7, 8]
bst.insert(root, 12)
assert inorder_traverse(root) == [1, 2, 5, 6, 7, 8, 12]
bst.insert(root, 9)
assert inorder_traverse(root) == [1, 2, 5, 6, 7, 8, 9, 12]
bst.insert(root, 13)
assert inorder_traverse(root) == [1, 2, 5, 6, 7, 8, 9, 12, 13]
assert bst.search(root, 5) == True
assert bst.search(root, 3) == False
# if left and right exist
bst.delete(root, 8)
assert inorder_traverse(root) == [1, 2, 5, 6, 7, 9, 12, 13]
# if left and right is None
bst.delete(root, 1)
assert inorder_traverse(root) == [2, 5, 6, 7, 9, 12, 13]
bst.delete(root, 2)
assert inorder_traverse(root) == [5, 6, 7, 9, 12, 13]
# if left is None
bst.delete(root, 7)
assert bst.search(root, 7) == False
assert inorder_traverse(root) == [5, 6, 9, 12, 13]
# if right is None
bst.insert(root, 2)
bst.insert(root, 1)
bst.delete(root, 2)
assert inorder_traverse(root) == [1, 5, 6, 9, 12, 13]
# delete root node
bst.delete(root, 5)
assert inorder_traverse(root) == [1, 6, 9, 12, 13]
def insert(self, num: int):
self.size += 1
if self.root is None:
self.root = TreeNode(num)
return
node = self.root
while node:
if num < node.val:
if not node.left:
node.left = TreeNode(num)
return
node = node.left
else:
if not node.right:
node.right = TreeNode(num)
return
node = node.right
def run_branch_and_bound(self, data):
start_time = time.time()
self.vertex_count = len(data)
data = numpy.array(data)
#ensure all the zeros or negative values are set to inifinity,
#zero has a special meaining for this algorithnm
data[data <= 0] = math.inf
#set the upper bound to infinity
self.upper_bound = math.inf
#perform the initial reduction and get the lower bound
starting_reduced_matrix, lower_bound = self.calculate_reduction(
data, 0, 0)
#created the first node in the state space tree
tree_nodes = []
working_node = TreeNode(1, 1, starting_reduced_matrix, lower_bound, 0,
[1], 0, -1)
tree_nodes.append(working_node)
while (working_node != None):
#do a deep dive to the first leaf
working_path, working_bound, working_node = self.deep_dive_from_node(
working_node.vertex_id, working_node, tree_nodes)
#if we have a loop set the upper bound else keep looking for a loop
if working_path[0] == working_path[-1]:
status = "searched"
#set upper bound to the result of the dive if lower
if working_bound < self.upper_bound:
self.upper_bound = working_bound
self.best_path = working_path
else:
status = "pruned"
while working_node.number_of_children < 2:
if working_node.vertex_id != self.start_vertex:
working_node.status = status
working_node = self.find_node_in_list_by_node_id(
working_node.parent_node_id, tree_nodes)
if (working_node == None):
break
return self.upper_bound, self.best_path, time.time() - start_time
def sortedArrayToBST(self, nums: List[int]) -> TreeNode:
def dfs(node, start: int, end: int):
mid = (start + end) // 2
self.insertNode(node, nums[mid])
if start < mid:
dfs(node, start, mid - 1)
if mid < end:
dfs(node, mid + 1, end)
dummy_node = TreeNode(None)
dfs(dummy_node, 0, len(nums) - 1)
return dummy_node.right
def buildTree(self, preorder: List[int], inorder: List[int]) -> TreeNode:
if not inorder:
return None
idx_map = {val: idx for idx, val in enumerate(inorder)}
dq = deque(preorder)
root = TreeNode(dq.popleft())
stack = [root]
while dq:
val = dq.popleft()
node = TreeNode(val)
# left
if idx_map[val] < idx_map[stack[-1].val]:
stack[-1].left = node
else:
# right
while stack and idx_map[stack[-1].val] < idx_map[val]:
u = stack.pop()
u.right = node
stack.append(node)
return root
def find_lowest_cost_next_vertex(self, parent_node, nodes_list):
available_vertices = self.get_possible_edges_for_vertex(parent_node)
lowest_cost = math.inf
least_cost_vertex = 0
least_cost_node = None
for vertex in available_vertices:
# need to copy this every loop to prevent carry over changes
path = copy.deepcopy(parent_node.path_so_far)
matrix = copy.deepcopy(parent_node.reduced_matrix)
path.append(vertex)
# check for existing node in the list by path
exiting_node = self.find_node_in_list_by_path(path, nodes_list)
if (exiting_node != None):
if (exiting_node.status != "pruned"
and exiting_node.status != "searched"):
tree_node = exiting_node
else:
#reduce the number of children of the parent node
parent_node.number_of_children = parent_node.number_of_children - 1
continue
else:
# actual cost of going between current vertices
actual_cost = matrix[parent_node.index_in_matrix, vertex - 1]
# reduced matrix and cost for node being calculated
reduced_matrix, reduction_cost = self.calculate_reduction(
matrix, parent_node.vertex_id, vertex)
cost = reduction_cost + parent_node.cost + actual_cost
# add node to the list since we have calculated it
tree_node = TreeNode(
len(nodes_list) + 1, vertex, reduced_matrix, cost,
vertex - 1, path, parent_node.node_id, -1)
nodes_list.append(tree_node)
if tree_node.cost < lowest_cost:
lowest_cost = tree_node.cost
least_cost_vertex = tree_node.vertex_id
least_cost_node = tree_node
return least_cost_vertex, least_cost_node, nodes_list
