python The efficiency of program execution is significantly lower than c Of .
Optimize python One idea of the code is ： take python Core code （ It usually needs to be called many times , Or the core loop ） convert to c Language .
There is no need to rewrite the program to c, Only need to use Cython Bao Jiang python The program can be converted .
1.Cython package
Cython For package installation, see https://pypi.org/project/Cython/ pip install Cython conda install Cython
General installation anaconda When it comes to Cython package ; You can check to see if you already have this package .（ Directly in the program import Cython, If no error is reported, it means that there is ; It can also be done through pip list Etc ）
2.python–>c
2.1
Rename the program to be converted to xxxx.pyx; For example, I have an encapsulated class .py file word_frequency.py, Need to be renamed to word_frequency.pyx
2.2
Create a file in the same directory setup.py; effect ： Use cython The compiler compiles pyx file
from distutils.core import setup
from Cython.Build import cythonize
setup(ext_modules=cythonize("word_frequency.pyx"))
2.3
Start compilation , On the command line ： python setup.py build_ext --inplace (# inplace Indicates that the source code is in the same directory )
If compilation is successful , Then two files will appear in the directory ：word_frequency.c,word_frequency.cp38-win_amd64.pyd（ If in Linux Under the platform word_frequency.cp38-win_amd64.so）, here .so Documents or .pyd Files can be like ordinary python file , By import.
In the above documents word_frequency.c It's no use , You can delete ;word_frequency.cp38-win_amd64.pyd The file can be renamed to word_frequency.pyd.
2.4
Call normally in other files word_frequency.pyd that will do （ Can be word_frequency.py Change it to another name , So as not to call wrong ）.
from word_frequency import word_frequency # Same import
wordFreq = word_frequency(dict_fullpath) # The same way of using
2.5
contrast
Before conversion ：130.40097641944885 s
After the transformation ：68.84645509719849 s
The speed has only doubled .
Cython Can compare with conventional Python How much faster the code actually depends on the code itself . for example , If you run a multivariable cycle with a high cost ,Cython Obviously superior to conventional Python Code . Recursive functions also make Cython Faster than Python.
3 About .pyd file
If you want to be right python File encryption , Will be able to python The module is compiled into .pyd file
pyd File is made up of D A language written in dll file , At present, there is no way to decompile , Can only be disassembled . to python Files provide high security . And you can also write some python plug-in unit , convenient python Development .
