We have summarized the points for those who can use C ++ but do not understand python to acquire the minimum programming notation that allows quick machine learning. If you know this, you can use the minimum functions.
Since the library is extensive, you can easily write complicated code by writing in C ++. Web cooperation is relatively easy. There is a lot of information on the net.
The following all libraries exist. Mastering the library means mastering python. ・ Machine learning ・ Web scraping ・ Excel operation ・ Data preprocessing ・ Visualize
python 3.8.2 Ubuntu 18.04
In the case of linux, python3 is included as standard, so there is no need to build an environment. If you want to run it on Anaconda, please refer to here. Anaconda is explained here [https://py-prog.com/whats-anaconda-and-its-merit-for-python-programming/) Anaconda makes it easy to create virtual environments. In python, which installs a lot of packages, you often want to change the environment, so if you want to use it seriously, Anaconda may be indispensable.
IDE ・ PyCharm ・ Spyder editor ・ Atom ・ VS code
| C++ | Python | |
|---|---|---|
| Variable type | Yes | None |
| Variable assignment | copy | Pointer |
| Function delimiter | {}Braces | :And indent |
| Arithmetic operator | C++In addition to**Add exponentiation |
|
| Logical operator | &&,ll,! | and,or,not |
| Code break | ; | new line |
| Pointer | Clear | 不Clear |
| Comment out | // | # |
| Importing external libraries | #include | import |
| compile | necessary | Unnecessary |
| main function | Yes | None |
| Matrix calculation(standard) | impossible | Possible with numpy library |
Since python is a dynamically typed language, the variable types that are commonplace in C ++ do not exist. Therefore, the variable is defined only by the variable name as it is. The type is determined by the assigned value.
a=1234 #Integer type
b=12.345 #Floating point type
c="hello world" #String type
d=[1,2,3,4] #List type
There are no explicit pointers in python. Instead, the assignment to the variable is passing a reference rather than a copy, because the variable is both a value holder and a pointer. ** ** If you change the value of the assignment destination, the value of the variable of the assignment source will also change, so be careful of the difference in handling. (For list) Refer to the following https://qiita.com/maruman029/items/21ad04e326bb4fb6f71a
In the case of C ++, 1 code is recognized no matter how many line breaks are made unless the delimiter ; is added, but in the case of python, the line break is a clear delimiter.
In python, the indent indicates the scope range. Therefore, if even one character is out of indentation, it will not be recognized as a code in the same range. It's a very indentable language, so coding can be harsh if you don't use the editor's indentation formatting feature. For example, the following code will result in an indentation error. Be aware that indentation often shifts and bugs occur in the loop.
a=1234
b=12.345
c="hello world" #Not in the same range
d=[1,2,3,4]
Global variables in C ++ can be accessed as they are from functions, but in the case of python, they cannot be assigned unless they are declared locally as global variable name in the function.
a = 123
def Global():
global a
a = 456
print("global variable:", a)
Global()
Some main functions are omitted for readability.
You don't need \ n because python is an automatic line break.
c.c
int a=123;
printf("%d\n",a);
python.py
a=123
print(a)
c++.cpp
#include "hogohoge.h"
python.py
import hogehoge
In python it is called a "list". The memory area of the python array is dynamically allocated, so you can add elements later. Use append () to add an element. If you substitute with ʻa [3] = 4` like C ++, an error will occur, so be careful.
c++.cpp
int a[4]={1,2,3,};
int b = a[2];
printf("%d\n", b);
a[3] = 4;
printf("%d\n", a);
python.py
a=[1,2,3]
b=a[2]
print(b)
a.append(4)
print(a)
Looking at the python code, I find ʻa.append (4) . This is a list method, a function used to manipulate lists. In python, variables and external libraries are treated as a kind of object, and you can access the methods provided in python with . method`. There are various other methods to insert in the middle, so here will be helpful.
An array whose value cannot be changed later. In c ++, it is declared as const, but in pyton, it is called a tuple and is written differently from the list.
tupl.cpp
const int a[4] = {1, 2, 3, 4};
b = a[2];
printf("%d", b);
tupl.py
a = (1, 2, 3, 4)
b = a[2]#Get element
print(b)
Python has a feature called a dictionary. It's like an array that can store data with a combination of keys and values. You can call it by specifying the key. Therefore, readability is improved. It may be easier to understand if you imagine it as something similar to a c ++ enum.
python.py
lang={"c":1,"python":2,"ruby":3}#Creating a dictionary
print(lang["python"])
lang["ruby"]=4#Swap elements
print(lang["ruby"])
lang["java"]=3 #Add element
print(lang)
Python does not have a delimiter such as {} like c ++. Therefore, by opening the next line of ʻif hogehoge:` by one tab, * the indent is changed to clearly indicate that it is an if statement process. If you bring pirnt to the left end, it will be indented the same as the if statement, so it will be the process executed at the outside of the if. Also shift the indent when nesting.
c++.cpp
int a = 7;
int b = 12;
if (a <= 10 && b <= 10)
{
printf("Both a and b are 10 or less");
}
else if (a % 2 == 0)
{
printf("a is even");
}
else
{
printf("Does not meet the conditions")
}
python.py
a = 7
b = 12
if a <= 10 and b <= 10:
print("Both a and b are 10 or less")
elif a % 2 == 0:
print("a is even")
else:
print("Does not meet the conditions")
In the case of c ++, if you want to retrieve the contents of the array one by one, you have to specify the number of elements in the array and loop. In the case of python, all the elements are fetched one by one just by throwing the list.
for.cpp
int ls[3] = {1, 2, 5};
for (int i = 0; i < sizeof ls / sizeof ls[0]; i++)
{
printf("%d\n", ls[i]);
}
for (int i = 0; i < 3; i++)
{
printf("%d\n", i);
}
for.py
ls = [1, 2, 5]
for i in ls:
print(i)
for i in range(3):
print(i)
1 2 5 0 1 2
while.cpp
int a = 0;
while(a < 3)
{
printf("%d", a);
a += 1;
}
while.py
a = 0
while a < 3:
print(a)
a += 1
0 1 2
When receiving a variadic argument, prefix the formal argument with *.
When expanding the contents of a list or tuple to an argument and passing it, add * before the actual argument.
function.cpp
int Add(a, b)
{
c = a + b;
return c;
}
int Add3(int *c)
{
int d = 0;
d = c[0] + c[1] + c[2];
return d;
}
int main()
{
printf("%d\n", Add(1, 2));
const int e[3] = {1, 2, 3};
printf("%d\n", Add3(e));
}
function.py
def Add(a, b):
c = a + b
return c
def Add3(*c): #Receive all arguments as variable super arguments
d = c[0] + c[1] + c[2]
return d
def main():
print(Add(1, 2))
print(Add3(1, 2, 3))
e = (4, 5)
print(Add(*e)) #It is also possible to expand the argument of a list or tuple and pass it
return
if __name__ == "__main__":
main()
classdefine.cpp
class Hoge
{
public:
int a;
Hoge(int a)
{
this->a = a;
}
void add(int b)
{
printf("%d\n", a + b);
}
void mutipl(int b)
{
printf("%d\n", a * b);
}
}
classdefine.py
class Hoge:
def __init__(self, a):
self.a = a
def add(self, b):
print(self.a + b)
def mutipl(self, b):
print(self.a * b)
In c ++, the constructor has the same name as the class, but in the "class" of python, the constructor is represented by def __init __ (self, a):, and the variables in the class are declared as self.a in the constructor. To do. The definition itself can be done from anywhere. Access the instance variable via "self" by passing "self" which is a reference of the class itself to the argument of the member function. Be sure to write self as an argument.
classcntrol.cpp
Hoge hoge(3);
hoge.add(4);
hoge.mutipl(4);
classcontrol.py
hoge = Hoge(3)
hoge.add(4)
hoge.mutipl(4)
Inheritance.cpp
class Fuga : public Hoge
{
public:
void subtract(int b)
{
printf("%d\n", a - b);
}
void divide(int b)
{
printf("%d\n", a / b);
}
}
Fuga fuga(3);
fuga.add(4);
fuga.mutipl(4);
fuga.subtract(4);
fuga.divide(4);
Inheritance.py
class Fuga(Hoge): #Inherit Hoge
def subtract(self, b):
print(self.a - b)
def divide(self, b):
print(self.a/b)
fuga = Fuga(3)
fuga.add(4)
fuga.mutipl(4)
fuga.subtract(4)
fuga.divide(4)
Matrix operations can be done very easily by using this. Required for machine learning. This person describes how to use it in detail, so please refer to it.
Please refer to this article for how to write.
In python, external libraries are treated as objects, but you can replace them with ʻas` because it's a hassle to write long names.
import numpy as np
pip3 install module_name
python3 file_name
anaconda
python file_name
File operations are fairly easy in python. https://qiita.com/hiroyuki_mrp/items/8bbd9ab6c16601e87a9c
https://note.nkmk.me/python-import-usage/
It's still just a small part of python, so it's a good idea to find out if there are any features you're interested in.
Yukinaga Gazuma (Author) "First Deep Learning: Neural Network and Back Propagation Learned with Python"
Recommended Posts