How to enter Japanese with Python curses

environment

Mint 19.10
Python 3.7.5

Introduction

When I used getch () of curses, I couldn't receive Japanese (multibyte characters) well when encoding with UTF-8, so I made a processing part.

Symptoms

For example, if you type'a', it will be divided into 3 bytes of 0xe3 0x81 0x82 and will be input 3 times. However, this is inconvenient. (What I really want is 0x3042)

Cause

Since getch () looks at the input byte by byte, UTF-8 cannot correctly receive Japanese etc. which is 3 bytes. So, I will make my own UTF-8 encoder.

solution

Looking at the first byte, the number of bytes in the character string is fixed, so conditional branching is performed as follows.

Pseudo code. The stdscr part etc. are omitted. Declare the window object accordingly

import curses

window = curses.stdscr()
key = window.getch()

#Processing of multi-byte characters
#In Japanese, it's 3 bytes, so you need to pool it. Look at the first byte
#Since the remaining vices and the number are fixed, the process is performed.
text_pool = [key]
if 0x00 <= key <= 0x7f:
     #1B so you don't have to do anything
     #ascii compatible area
     pass
elif 0x80 <= key <= 0xbf:
     #This should be after the second character, so it would be strange if it came in
     print(key)
     exit(1)
elif 0xc0 <= key <= 0xdf:
     #2B characters with umlauts
     text_pool.append(self.window.getch())
     # text_pool => [0dAAA, 0dBBB]
     # 110a aabb 10bb bbbb <=This is text_Contents of pool(Decimal version)
     #0b00000aaa bbbbbbbb is taken out and char c= (char) (data[i] & 0xff);
     #Convert to decimal number and assign to key
     a, b = text_pool
     tmp = map(lambda x: bin(x)[2:], [0b00011111 & a, 0b00111111 & b])
     tmp = ''.join(item.zfill(6) for item in tmp)
     key = int(tmp,2)
elif 0xe0 <= key <= 0xef:
     #3B Japanese is here
     for _ in range(2):
         text_pool.append(self.window.getch())
         a, b, c = text_pool
         # 0b 1110xxxx 10xxyyyy 10yyyyyy
         # 0d a        b        c
         tmp = map(lambda x: bin(x)[2:], [0b00001111 & a, 0b00111111 & b, 0b00111111 & c])
         tmp = ''.join([item.zfill(6) for item in tmp])
         key = int(tmp,2)
elif 0xf0 <= key <= 0xff:
#4B I've never seen it, but bug fixing
     for _ in range(3):
         text_pool.append(self.window.getch())
         a, b, c ,d = text_pool
         # 11110xxx 10xxyyyy 10yyyyzz 10zzzzzz
         tmp = map(lambda x: bin(x)[2:], [0b00000111 & a, 0b00111111 & b, 0b00111111 & c, 0b00111111 & d])
         tmp = ''.join([item.zfill(6) for item in tmp])
         key = int(tmp,2)
else:
    #Special key
    pass


print(chr(key))

Relation

About mouse wheel judgment https://qiita.com/t4t5u0/items/ae6d25e05b7f7094330e

reference

https://seiai.ed.jp/sys/text/cs/mcodes/ucodeutf8.html