def __init__(self, arr=None, capacity=None):
if isinstance(arr, Array):
self._data = arr
for i in range(self._parent(arr.get_size() - 1), -1, -1):
self._sift_down(i)
return
if not capacity:
self._data = Array()
else:
self._data = Array(capacity=capacity)
def __init__(self, arr=None, capacity=None):
# heapify Array to Heap and save into _data,
if isinstance(arr, Array):
self._data = arr
# find the parent of last leaf index, till the end of index=0, but range need to have extra 开区间，所以要用-1
# 逐个下沉
for i in range(self._parent(arr.get_size() - 1), -1, -1):
self._sift_down(i)
return
if not capacity:
self._data = Array()
else:
self._data = Array(capacity=capacity)
class ArrayStack(StackBase):
"""使用自定义的Array实现"""
def __init__(self, capacity=0):
self._array = Array(capacity=capacity)
def push(self, e):
self._array.add_last(e)
def pop(self):
return self._array.remove_last()
def peek(self):
return self._array.get_last()
def get_size(self):
return self._array.get_size()
def is_empty(self):
return self._array.is_empty()
def get_capacity(self):
return self._array.get_capacity()
def __str__(self):
return str('<chapter_03_Stack_Queue.stack.ArrayStack> : {}'.format(
self._array))
def __repr__(self):
return self.__str__()
class ArrayQueue(QueueBase):
def __init__(self, capacity=0):
self._array = Array(capacity)
def get_size(self):
return self._array.get_size()
def is_empty(self):
return self._array.is_empty()
def get_capacity(self):
return self.get_capacity()
def enqueue(self, e):
self._array.add_last(e)
def dequeue(self):
return self._array.remove_first()
def get_front(self):
return self._array.get_first()
def __str__(self):
return str('<chapter_03_Stack_Queue.queue.ArrayQueue> : {}'.format(
self._array))
def __repr__(self):
return self.__str__()
def __init__(self, capacity=0):
self._array = Array(capacity)
class MaxHeap:
def __init__(self, arr=None, capacity=None):
# heapify Array to Heap and save into _data,
if isinstance(arr, Array):
self._data = arr
# find the parent of last leaf index, till the end of index=0, but range need to have extra 开区间，所以要用-1
# 逐个下沉
for i in range(self._parent(arr.get_size() - 1), -1, -1):
self._sift_down(i)
return
if not capacity:
self._data = Array()
else:
self._data = Array(capacity=capacity)
def size(self):
return self._data.get_size()
def is_empty(self):
return self._data.is_empty()
# (1) return parent and left/right child
# 返回完全二叉树数组表示中，一个索引所表示的元素的父亲节点的索引 i-1 // 2, index start from 0
def _parent(self, index):
if index == 0:
raise ValueError('index-0 doesn\'t have parent.')
return (index - 1) // 2
# 返回完全二叉树数组表示中，一个索引所表示的元素的左孩子节点的索引 2 * i + 1
def _left_child(self, index):
return index * 2 + 1
# 返回完全二叉树数组表示中，一个索引所表示的元素的右孩子节点的索引 2 * i + 2
def _right_child(self, index):
return index * 2 + 2
# (2) add and find/extract element
# add (last) 和父亲节点对比逐层上浮, add the most recent index shift_up
def add(self, e):
self._data.add_last(e)
self._sift_up(self._data.get_size() - 1)
def _sift_up(self, k):
while k > 0 and self._data.get(k) > self._data.get(self._parent(k)):
self._data.swap(k, self._parent(k))
k = self._parent(k)
# k sift up to 0 stop compare
# extract max[0]: sway the last one with 0 and remove, then sift down
def find_max(self):
if self._data.get_size() == 0:
raise ValueError('Can not find_max when heap is empty.')
return self._data.get(0)
def extract_max(self):
ret = self.find_max()
self._data.swap(0, self._data.get_size() - 1)
self._data.remove_last()
self._sift_down(0)
return ret
# extra element, parent compare to child, swap the maximum child
def _sift_down(self, k):
# k End Condition:k not stay in last leaf layer, no left child
while self._left_child(k) < self._data.get_size():
j = self._left_child(k)
# Compare right and left child to find the max one, k has right child, j+1<size
if j + 1 < self._data.get_size() and self._data.get(j + 1) > self._data.get(j):
# 说明右孩子的值比左孩子的值大 j+=1
j = self._right_child(k)
# k 没有右孩子或者右孩子还要小
# do not have right child or right child<left child, j is larger index, otherwise, update with right index
# 此时self._data.get(j)是左孩子和右孩子中的最大值
# no need to sift down
if self._data.get(k) > self._data.get(j):
break
# By default otherwise swap, then evaluate sift down updated with new k (max child in j)
self._data.swap(k, j)
k = j
def replace(self, e):
ret = self.find_max()
# 这样可以一次logn完成
self._data.set(0, e)
self._sift_down(0)
return ret
for i in range(n):
max_heap.add(randint(0, 1000))
print('heap add: ', time() - start_time1) # head add: 57.48132228851318
# check results must in sequence bigger and smaller
# arr_pop = []
# for i in range(n):
# arr_pop.append(max_heap.extract_max())
# for i in range(n-1):
# if arr_pop[i] < arr_pop[i+1]:
# raise ValueError('Error')
#
# print("Test MaxHeap completed.")
start_time2 = time()
arr = Array()
from random import randint
for i in range(n):
arr.add_last(randint(0, 1000))
max_heap = MaxHeap(arr)
print('heapify: ', time() - start_time2) # heapify: 46.80660963058472
def testHeap(arr, isheapify):
start__time = time()
if(isheapify):
max__heap = MaxHeap(arr)
else:
max__heap = MaxHeap()
for i in range(len(arr)):
max__heap.add(i)
class MaxHeap:
#heapify，将任意数组整理成最大堆形式O(n)
def __init__(self, arr=None, capacity=None):
if isinstance(arr, Array):
self._data = arr
#从最后一个非叶子节点开始，就是最后叶子节点的父节点，进行下沉操作
for i in range(self._parent(arr.get_size() - 1), -1, -1):
self._sift_down(i)
return
if not capacity:
self._data = Array()
else:
self._data = Array(capacity=capacity)
def size(self):
return self._data.get_size()
def is_empty(self):
return self._data.is_empty()
# 返回完全二叉树数组表示中，一个索引所表示的元素的父亲节点的索引 i // 2
def _parent(self, index):
if index == 0:
raise ValueError('index-0 doesn\'t have parent.')
return (index - 1) // 2
# 返回完全二叉树数组表示中，一个索引所表示的元素的左孩子节点的索引 2 * i + 1
def _left_child(self, index):
return index * 2 + 1
# 返回完全二叉树数组表示中，一个索引所表示的元素的右孩子节点的索引 2 * i + 2
def _right_child(self, index):
return index * 2 + 2
#向堆里添加元素
def add(self, e):
self._data.add_last(e)
#传入的参数是上浮所对应的索引：最后一个元素
self._sift_up(self._data.get_size() - 1)
def _sift_up(self, k):
while k > 0 and self._data.get(k) > self._data.get(self._parent(k)):
self._data.swap(k, self._parent(k))
k = self._parent(k)
#看堆中最大元素
def find_max(self):
if self._data.get_size() == 0:
raise ValueError('Can not find_max when heap is empty.')
return self._data.get(0)
#取出最大元素
def extract_max(self):
ret = self.find_max()
self._data.swap(0, self._data.get_size() - 1)
self._data.remove_last()
self._sift_down(0)
return ret
def _sift_down(self, k):
#右孩子的索引比左孩子大
while self._left_child(k) < self._data.get_size():
j = self._left_child(k)
if j + 1 < self._data.get_size() and self._data.get(j + 1) > \
self._data.get(j):
# 说明右孩子的值比左孩子的值大
j = self._right_child(k)
# 此时self._data.get(j)是左孩子和右孩子中的最大值
if self._data.get(k) > self._data.get(j):
break
self._data.swap(k, j)
k = j
#去除堆中最大元素，并且替换成元素e
def replace(self, e):
ret = self.find_max()
# 这样可以一次logn完成
self._data.set(0, e)
self._sift_down(0)
return ret
