所有代码运行在jupyter notebook中
1、利用%lprun进行逐行分析
用 %prun 对代码中的每个函数进行分析非常有用,但有时逐行代码分析报告更方便。首先利用 Python 的包管理工具 pip 安 装 line_profiler 包:
pip install line_profiler
接下来可以 line_profiler 包提供的 IPython 扩展:
In [1]:
%load_ext line_profiler
In [2]:
def sum_of_lists(N):
total = 0
for i in range(5):
L = [j^(j >> i) for j in range(N)]
total +=sum(L)
return total
现在 %lprun 命令就可以对所有函数进行逐行分析了。在下面的例子中,我们需要明确指出要分析哪些函数:
In [3]:
%prun sum_of_lists(1000000)
Out [3]:
Timer unit: 1e-07 s
Total time: 0.0108332 s
File: <ipython-input-2-54cf9a3b9717>
Function: sum_of_lists at line 1
Line # Hits Time Per Hit % Time Line Contents
==============================================================
1 def sum_of_lists(N):
2 1 19.0 19.0 0.0 total = 0
3 6 109.0 18.2 0.1 for i in range(5):
4 5 104182.0 20836.4 96.2 L = [j^(j >> i) for j in range(N)]
5 5 4015.0 803.0 3.7 total +=sum(L)
6 1 7.0 7.0 0.0 return total
我们可以看到程序中哪些地方最耗时。可以通过这些信息修改代码,使其更高效地实现我们的目的。
2、利用%memit和%mprun进行内存分析
另一种分析是分析一个操作所用的内存量,首先 用 pip 安装这个扩展:
pip install memory_profiler
然后导入该扩展:
In [1]:
%load_ext memory_profiler
内存分析扩展包括两个有用的魔法函数:%memit 魔法函数(它提供的 内存消耗计算功能类似于 %timeit)和 %mprun 魔法函数(它提供的内 存消耗计算功能类似于 %lprun)。%memit 函数用起来很简单:
In [2]:
%memit sum_of_lists(1000000)
Out [2]:
peak memory: 125.06 MiB, increment: 73.29 MiB
可以看到,这个函数大概消耗了 125.86MB 的内存。
对于逐行代码的内存消耗描述,首先用 %%file 魔法函数创建一个简单的模块,将该模块命名为 mprun_demo.py。它包含 sum_of_lists 函数,该函数 中包含一次加法,能使内存分析结果更清晰:
In [3]:
%%file mprun_demo.py
def sum_of_lists(N):
total = 0
for i in range(5):
L = [j^(j >> i) for j in range(N)]
total +=sum(L)
del L
return total
Out [3]:
Writing mprun_demo.py
重新导入函数,并运行逐行内存分析器:
In [4]:
from mprun_demo import sum_of_lists
%mprun -f sum_of_lists sum_of_lists(100000)
Out [4]:
Filename: F:\代码01Data\数据分析学习代码\Python数据科学手册\练习代码\mprun_demo.py
Line # Mem usage Increment Line Contents
================================================
1 81.0 MiB 81.0 MiB def sum_of_lists(N):
2 81.0 MiB 0.0 MiB total = 0
3 82.8 MiB 0.0 MiB for i in range(5):
4 83.3 MiB 0.6 MiB L = [j^(j >> i) for j in range(N)]
5 83.3 MiB 0.0 MiB total +=sum(L)
6 82.8 MiB 0.0 MiB del L
7 82.8 MiB 0.0 MiB return total
可以更改代码,把这利用到分析其他代码的内存上,这样分析可以提高算法的效率。
Increment 列告诉我们每行代码对总内存预算的影响。如果想了解这个函数怎么用 ,在jupyter notebook里输入:(用符号?获取文档,通过符号??获取源代码)
%memit?
# %memit??
结果为:
Docstring:
Measure memory usage of a Python statement
Usage, in line mode:
%memit [-r<R>t<T>i<I>] statement
Usage, in cell mode:
%%memit [-r<R>t<T>i<I>] setup_code
code...
code...
This function can be used both as a line and cell magic:
- In line mode you can measure a single-line statement (though multiple
ones can be chained with using semicolons).
- In cell mode, the statement in the first line is used as setup code
(executed but not measured) and the body of the cell is measured.
The cell body has access to any variables created in the setup code.
Options:
-r<R>: repeat the loop iteration <R> times and take the best result.
Default: 1
-t<T>: timeout after <T> seconds. Default: None
-i<I>: Get time information at an interval of I times per second.
Defaults to 0.1 so that there is ten measurements per second.
-c: If present, add the memory usage of any children process to the report.
-o: If present, return a object containing memit run details
-q: If present, be quiet and do not output a result.
Examples
--------
::
In [1]: %memit range(10000)
peak memory: 21.42 MiB, increment: 0.41 MiB
In [2]: %memit range(1000000)
peak memory: 52.10 MiB, increment: 31.08 MiB
In [3]: %%memit l=range(1000000)
...: len(l)
...:
peak memory: 52.14 MiB, increment: 0.08 MiB