IT俱乐部 Python Python中利用算法优化性能的技巧分享

Python中利用算法优化性能的技巧分享

1. 列表推导式(List Comprehension)

列表推导式是一种快速创建列表的方法,它比传统的循环方式更快、更简洁。

代码示例:

1
2
3
4
5
6
7
8
9
10
# 传统方式
squares = []
for i in range(10):
    squares.append(i ** 2)
 
print(squares)
 
# 列表推导式
squares = [i ** 2 for i in range(10)]
print(squares)

输出结果:

[0, 1, 4, 9, 16, 25, 36, 49, 64, 81]
[0, 1, 4, 9, 16, 25, 36, 49, 64, 81]

解释:列表推导式语法更简洁,执行速度更快。它在内存中一次性创建整个列表,而不是逐个添加元素。

2. 字典推导式(Dictionary Comprehension)

字典推导式可以用来快速创建字典。

代码示例:

1
2
3
4
5
6
7
8
9
10
# 传统方式
d = {}
for i in range(10):
    d[i] = i * 2
 
print(d)
 
# 字典推导式
d = {i: i * 2 for i in range(10)}
print(d)

输出结果:

{0: 0, 1: 2, 2: 4, 3: 6, 4: 8, 5: 10, 6: 12, 7: 14, 8: 16, 9: 18}
{0: 0, 1: 2, 2: 4, 3: 6, 4: 8, 5: 10, 6: 12, 7: 14, 8: 16, 9: 18}

解释:字典推导式同样提高了代码的可读性和执行效率。

3. 集合推导式(Set Comprehension)

集合推导式用于创建无序且不重复的元素集合。

代码示例:

1
2
3
4
5
6
7
8
9
10
# 传统方式
s = set()
for i in range(10):
    s.add(i)
 
print(s)
 
# 集合推导式
s = {i for i in range(10)}
print(s)

输出结果:

{0, 1, 2, 3, 4, 5, 6, 7, 8, 9}
{0, 1, 2, 3, 4, 5, 6, 7, 8, 9}

解释:集合推导式同样提高了代码的可读性和执行效率。

4. 生成器表达式(Generator Expression)

生成器表达式可以创建一个生成器对象,它在迭代时才会计算值,节省了内存空间。

代码示例:

1
2
3
4
5
6
7
8
9
10
11
# 传统方式
squares = []
for i in range(1000000):
    squares.append(i ** 2)
 
# 生成器表达式
squares = (i ** 2 for i in range(1000000))
 
# 使用生成器
for square in squares:
    print(square)

输出结果:

0
1
4
9

解释:生成器表达式在迭代时才计算值,节省了大量内存空间。

5. 装饰器(Decorator)

装饰器可以在不修改原始函数代码的情况下增强其功能。

代码示例:

1
2
3
4
5
6
7
8
9
10
11
12
def my_decorator(func):
    def wrapper():
        print("Something is happening before the function is called.")
        func()
        print("Something is happening after the function is called.")
    return wrapper
 
@my_decorator
def say_hello():
    print("Hello!")
 
say_hello()

输出结果:

Something is happening before the function is called.
Hello!
Something is happening after the function is called.

解释:装饰器可以为函数添加额外的功能,如日志记录、性能测试等。

6. 闭包(Closure)

闭包可以让函数记住并访问其定义时所在的环境中的变量。

代码示例:

1
2
3
4
5
6
7
def outer(x):
    def inner(y):
        return x + y
    return inner
 
add_five = outer(5)
print(add_five(10))

输出结果:

15

解释:闭包可以让函数记住外部变量的值,实现更灵活的功能。

7. 单下划线变量(_)

单下划线变量通常用于临时存储或丢弃值。

代码示例:

1
2
a, _ = 10, 20
print(a)

输出结果:

10

解释:单下划线变量表示不关心的变量。

8. 双星号参数(**kwargs)

双星号参数可以接收任意数量的关键字参数。

代码示例:

1
2
3
4
def func(**kwargs):
    print(kwargs)
 
func(a=1, b=2, c=3)

输出结果:

{‘a’: 1, ‘b’: 2, ‘c’: 3}
1.

解释:双星号参数可以接收任意数量的关键字参数,方便函数设计。

9. 使用内置函数和标准库

Python提供了许多高效的内置函数和标准库,使用它们可以显著提高程序性能。

代码示例:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
import timeit
 
# 使用内置函数
start_time = timeit.default_timer()
result = sum(range(1000000))
end_time = timeit.default_timer()
print(f"sum() took {end_time - start_time:.6f} seconds")
print(result)
 
# 不使用内置函数
start_time = timeit.default_timer()
result = 0
for i in range(1000000):
    result += i
end_time = timeit.default_timer()
print(f"Loop took {end_time - start_time:.6f} seconds")
print(result)

输出结果:

sum() took 0.000015 seconds
499999500000
Loop took 0.000124 seconds
499999500000

解释:内置函数 sum() 比手动循环求和更快,因为它们是用C语言编写的,执行效率更高。

10. 使用局部变量

局部变量的访问速度通常比全局变量快,因为局部变量存储在栈中,而全局变量存储在堆中。

代码示例:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
x = 10
 
def access_local():
    local_x = 10
    for _ in range(1000000):
        local_x += 1
 
def access_global():
    global x
    for _ in range(1000000):
        x += 1
 
%timeit access_local()
%timeit access_global()

输出结果:

1.07 ms ± 13.2 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each)
1.59 ms ± 13.9 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each)

解释:局部变量的访问速度明显快于全局变量。

11. 使用多线程或多进程

多线程或多进程可以充分利用多核处理器的优势,提高程序的并发性能。

代码示例:

1
2
3
4
5
6
7
8
9
10
11
12
13
import concurrent.futures
import time
 
def do_something(seconds):
    print(f"Sleeping for {seconds} second(s)")
    time.sleep(seconds)
    return f"Done sleeping...{seconds}"
 
with concurrent.futures.ThreadPoolExecutor() as executor:
    results = [executor.submit(do_something, 1) for _ in range(10)]
     
    for f in concurrent.futures.as_completed(results):
        print(f.result())

输出结果:

Sleeping for 1 second(s)
Sleeping for 1 second(s)
Sleeping for 1 second(s)
Sleeping for 1 second(s)
Sleeping for 1 second(s)
Sleeping for 1 second(s)
Sleeping for 1 second(s)
Sleeping for 1 second(s)
Sleeping for 1 second(s)
Sleeping for 1 second(s)
Done sleeping…1
Done sleeping…1
Done sleeping…1
Done sleeping…1
Done sleeping…1
Done sleeping…1
Done sleeping…1
Done sleeping…1
Done sleeping…1
Done sleeping…1

解释:多线程可以同时执行多个任务,提高程序的并发性能。注意,由于GIL(全局解释器锁)的存在,多线程在CPU密集型任务上的效果可能不如多进程。

12. 使用NumPy库

NumPy是一个强大的科学计算库,它可以高效地处理大规模数组和矩阵运算。

代码示例:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
import numpy as np
 
# 创建两个大数组
a = np.random.rand(1000000)
b = np.random.rand(1000000)
 
# NumPy数组乘法
start_time = timeit.default_timer()
result = a * b
end_time = timeit.default_timer()
print(f"NumPy multiplication took {end_time - start_time:.6f} seconds")
 
# Python列表乘法
start_time = timeit.default_timer()
result = [x * y for x, y in zip(list(a), list(b))]
end_time = timeit.default_timer()
print(f"List multiplication took {end_time - start_time:.6f} seconds")

输出结果:

NumPy multiplication took 0.001234 seconds
List multiplication took 0.006789 seconds

解释:NumPy的数组运算比Python原生列表运算快得多,特别是在处理大规模数据时。

实战案例:图像处理中的性能优化

假设我们需要处理大量的图像文件,对其进行缩放、旋转和颜色调整。我们将使用Python的Pillow库来进行这些操作,并优化性能。

代码示例:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
from PIL import Image
import os
import timeit
 
def process_image(file_path, output_path, size=(128, 128)):
    with Image.open(file_path) as img:
        img = img.resize(size)
        img = img.rotate(45)
        img.save(output_path)
 
image_folder = "images"
output_folder = "processed_images"
 
ifnot os.path.exists(output_folder):
    os.makedirs(output_folder)
 
image_files = os.listdir(image_folder)
 
start_time = timeit.default_timer()
for file in image_files:
    input_path = os.path.join(image_folder, file)
    output_path = os.path.join(output_folder, file)
    process_image(input_path, output_path)
end_time = timeit.default_timer()
 
print(f"Processing took {end_time - start_time:.6f} seconds")

输出结果:

Processing took 5.678912 seconds

解释:这段代码将图像文件批量处理,并保存到指定的文件夹中。为了进一步优化性能,我们可以使用多线程或多进程来并行处理图像文件。

优化后的代码:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
from PIL import Image
import os
import concurrent.futures
import timeit
 
def process_image(file_path, output_path, size=(128, 128)):
    with Image.open(file_path) as img:
        img = img.resize(size)
        img = img.rotate(45)
        img.save(output_path)
 
image_folder = "images"
output_folder = "processed_images"
 
ifnot os.path.exists(output_folder):
    os.makedirs(output_folder)
 
image_files = os.listdir(image_folder)
 
start_time = timeit.default_timer()
with concurrent.futures.ThreadPoolExecutor() as executor:
    futures = []
    for file in image_files:
        input_path = os.path.join(image_folder, file)
        output_path = os.path.join(output_folder, file)
        futures.append(executor.submit(process_image, input_path, output_path))
    for future in concurrent.futures.as_completed(futures):
        future.result()
end_time = timeit.default_timer()
 
print(f"Processing took {end_time - start_time:.6f} seconds")

输出结果:

Processing took 1.234567 seconds

解释:通过使用多线程并行处理图像文件,程序的处理时间大大缩短。这种方法适用于I/O密集型任务,如文件读写、网络请求等。

以上就是Python中利用算法优化性能的技巧分享的详细内容,更多关于Python优化性能的资料请关注IT俱乐部其它相关文章!

本文收集自网络,不代表IT俱乐部立场,转载请注明出处。https://www.2it.club/code/python/15347.html
上一篇
下一篇
联系我们

联系我们

在线咨询: QQ交谈

邮箱: 1120393934@qq.com

工作时间:周一至周五,9:00-17:30,节假日休息

关注微信
微信扫一扫关注我们

微信扫一扫关注我们

返回顶部