登录
首页大数据时代一行代码干掉 debug 和 print,助力算法学习(CDA干货分享)
一行代码干掉 debug 和 print,助力算法学习(CDA干货分享)
2021-12-29
收藏

作者:某某白米饭

来源:Python 技术

在写算法的时候,总是要每行每个变量一个个的 debug,有时候还要多写几个 print,一道算法题要花好长时间才能理解。pysnooper 模块可以把在运行中变量值都给打印出来。

模块安装

pip3 install pysnooper

简单例子

下面是道简单的力扣算法题作为一个简单的例子

import pysnooper @pysnooper.snoop() def longestCommonPrefix(strs): res = '' for i in zip(*strs):
        print(i) if len(set(i)) == 1:
            res += i[0] else break return res if __name__ == 'main':
    longestCommonPrefix(["flower","flow","flight"])

结果:

3:38:25.863579 call 4 def longestCommonPrefix(strs): 23:38:25.864474 line 5 res = '' New var:....... res = '' 23:38:25.864474 line 6 for i in zip(*strs): New var:....... i = ('f', 'f', 'f') 23:38:25.865479 line 7 print(i)
('f', 'f', 'f') 23:38:25.866471 line 8 if len(set(i))==1: 23:38:25.866471 line 9 res+=i[0]
Modified var:.. res = 'f' 23:38:25.866471 line 6 for i in zip(*strs):
Modified var:.. i = ('l', 'l', 'l') 23:38:25.866471 line 7 print(i)
('l', 'l', 'l') 23:38:25.867468 line 8 if len(set(i))==1: 23:38:25.867468 line 9 res+=i[0]
Modified var:.. res = 'fl' 23:38:25.868476 line 6 for i in zip(*strs):
Modified var:.. i = ('o', 'o', 'i') 23:38:25.868476 line 7 print(i)
('o', 'o', 'i') 23:38:25.869463 line 8 if len(set(i))==1: 23:38:25.869463 line 11 break 23:38:25.869463 line 12 return res 23:38:25.869463 return 12 return res Return value:.. 'fl' Elapsed time: 00:00:00.008201 

我们可以看到 pysnooper 把整个执行程序都记录了下来,其中包括行号, 行内容,变量的结果等情况,我们很容易的就看懂了这个算法的真实情况。并且不需要再使用 debug 和 print 调试代码。很是省时省力,只需要在方法上面加一行 @pysnooper.snoop()。

复杂使用

pysnooper 包含了多个参数,一起来看看吧

output

output 默认输出到控制台,设置后输出到文件,在服务器中运行的时候,特定的时间出现代码问题就很容易定位错误了,不然容易抓瞎。小编在实际中已经被这种问题困扰了好几次,每次都掉好多头发。

@pysnooper.snoop('D:pysnooper.log') def longestCommonPrefix(strs):

示例结果:

一行代码干掉 debug 和 print,助力算法学习

watch 和 watch_explode

watch 用来设置跟踪的非局部变量,watch_explode 表示设置的变量都不监控,只监控没设置的变量,正好和 watch 相反。

index = 1 @pysnooper.snoop(watch=('index')) def longestCommonPrefix(strs):

示例结果

没有加 watch 参数

Starting var:.. strs = ['flower', 'flow', 'flight'] 00:12:33.715367 call 5 def longestCommonPrefix(strs):
00:12:33.717324 line         7     res = '' New var:....... res = '' 

加了watch 参数,就会有一个 Starting var:.. index

Starting var:.. strs = ['flower', 'flow', 'flight']
Starting var:.. index = 1 00:10:35.151036 call 5 def longestCommonPrefix(strs):
00:10:35.151288 line         7     res = '' New var:....... res = '' 

depth

depth 监控函数的深度

@pysnooper.snoop(depth=2) def longestCommonPrefix(strs): otherMethod()

示例结果

Starting var:.. strs = ['flower', 'flow', 'flight'] 00:20:54.059803 call 5 def longestCommonPrefix(strs): 00:20:54.059803 line 6 otherMethod() 00:20:54.060785 call 16 def otherMethod(): 00:20:54.060785 line 17 x = 1 New var:....... x = 1 00:20:54.060785 line 18 x = x + 1 Modified var:.. x = 2 00:20:54.060785 return 18 x = x + 1 Return value:.. None 00:20:54.061782 line 7 res = '' 

监控的结果显示,当监控到调用的函数的时候,记录上会加上缩进,并将它的局部变量和返回值打印处理。

prefix

prefix 输出内容的前缀

@pysnooper.snoop(prefix='-------------') def longestCommonPrefix(strs):

示例结果

-------------Starting var:.. strs = ['flower', 'flow', 'flight']
-------------00:39:13.986741 call 5 def longestCommonPrefix(strs): -------------00:39:13.987218 line 6 res = '' 

relative_time

relative_time 代码运行的时间

@pysnooper.snoop(relative_time=True) def longestCommonPrefix(strs): 

示例结果

Starting var:.. strs = ['flower', 'flow', 'flight'] 00:00:00.000000 call 5 def longestCommonPrefix(strs):
00:00:00.001998 line         6     res = '' New var:....... res = '' 00:00:00.001998 line 7 for i in zip(*strs): 

max_variable_length

max_variable_length 输出的变量和异常的最大长度,默认是 100 个字符,超过 100 个字符就会被截断,可以设置为 max_variable_length=None 不截断输出

@pysnooper.snoop(max_variable_length=5) def longestCommonPrefix(strs): 

示例结果

Starting var:.. strs = [...]
00:56:44.343639 call 5 def longestCommonPrefix(strs): 00:56:44.344696 line 6 res = '' New var:....... res = '' 00:56:44.344696 line 7 for i in zip(*strs): New var:....... i = (...)

总结

本文介绍了怎么使用 pysnooper 工具,pysnooper 不仅可以少一些 debug 和 print,更能帮助理解算法题。

一行代码干掉 debug 和 print,助力算法学习
    

数据分析咨询请扫描二维码

客服在线
立即咨询