Format input (include pipeline, clipboard and arguments) text to improve its readability, with a focus on Chinese text.

format_text.py

#!/usr/bin/env python
"""
Format input (include pipeline, clipboard and arguments) text 
to improve its readability, with a focus on Chinese text.
"""
import platform
import re
import shutil
import subprocess
import sys
from os import isatty
from typing import Callable, List

BASIC_TRANSLATION_TABLE = str.maketrans("％＃＠＆｜１２３４５６７８９０", "%#@&|1234567890")

ENGLISH_TO_CHINESE_TABLE = str.maketrans(",.:;!?()", "，。：；！？（）")

NON_SPACE_AROUND_PUNCTUATION = "，。：；！？、【】（）「」"

NON_PREFIX_SPACE_PUNCTUATION_EN = ",.:;!?”’)"

NON_PREFIX_SPACE_PUNCTUATION = (
    NON_PREFIX_SPACE_PUNCTUATION_EN + NON_SPACE_AROUND_PUNCTUATION
)
NON_SUFFIX_SPACE_PUNCTUATION = NON_SPACE_AROUND_PUNCTUATION


def remove_spaces_around_chinese(text: str):
    """Remove all spaces between Chinese characters
    For example:
        input: "你 好 世 界 , Hello world ! "
        output: "你好世界 , Hello world ! "
    """
    # NOTE: We need to perform two substitutions in this case, and here's why:
    # our first pattern, ([\u4e00-\u9fff])\\s+([\u4e00-\u9fff]),
    # matches all instances where a Chinese character is followed by spaces and ends with another Chinese character.
    # If we have matched matchobj1 and matchobj2, and there are only spaces between them,
    # making only one substitution will leave a space between matchobj1 and matchobj2.
    # Therefore, we need to perform a second substitution.
    # If there's a more simple solution, please let tell us.
    text = re.sub(f"([\u4e00-\u9fff])([ \u3000]+|\n)([\u4e00-\u9fff])", r"\1\3", text)
    text = re.sub(f"([\u4e00-\u9fff])([ \u3000]+|\n)([\u4e00-\u9fff])", r"\1\3", text)

    return text


def remove_redundant_sapces(text: str) -> str:
    """Remove unnecessary spaces around punctuation"""
    # Remove leading spaces for some punctuation
    text = re.sub(f" +([{NON_PREFIX_SPACE_PUNCTUATION}])", r"\1", text)

    # Remove trailing space for some punctuation
    text = re.sub(f"([{NON_SUFFIX_SPACE_PUNCTUATION}]) +", r"\1", text)

    # Remove multiple leading spaces from some punctuation
    text = re.sub(f"""([{NON_PREFIX_SPACE_PUNCTUATION_EN}#'"]) +""", r"\1 ", text)

    # Remove spaces between `‘’“”` punctuation and Chinese character
    text = re.sub("([\u4e00-\u9fff])( *)([‘’“”])", r"\1\3", text)
    text = re.sub("([‘’“”])( *)([\u4e00-\u9fff])", r"\1\3", text)

    return text.rstrip()


def _standardlize_repl(matchobj: re.Match):
    """Replace English punctuation to Chinese punctuation"""
    return matchobj.group().translate(ENGLISH_TO_CHINESE_TABLE)


def fix_mixed_punctuation(text: str) -> str:
    """Convert English punctuation between Chinese characters to Chinese punctuation"""
    # Basic translation
    text = text.translate(BASIC_TRANSLATION_TABLE)
    # Replace English punctuation to Chinese punctuation around Chinese characters or Chinese punctuation
    text = re.sub("([\u4e00-\u9fff，。：；！？])(\\s*)([,.:;!?()])", _standardlize_repl, text)
    # Another substitution for the case `你好: (世界)` and so on
    text = re.sub("([\u4e00-\u9fff，。：；！？])(\\s*)([,.:;!?()])", _standardlize_repl, text)

    return text


def format_text(text: str) -> str:
    """Format text using series of formatters"""
    formatters: List[Callable[[str], str]] = [
        fix_mixed_punctuation,
        remove_redundant_sapces,
        remove_spaces_around_chinese,
    ]
    for f in formatters:
        text = f(text)
    return text


def get_clipboard_tk():
    """Access system clipboard using `tkinter`"""
    import tkinter as tk

    root = tk.Tk()
    root.withdraw()
    return root.clipboard_get()


def get_clipboard_macos() -> str:
    """Access macOS system clipboard

    Use `pbpaste` command to capture output,
    alternatively `osascript -e 'get the clipboard as text'`.
    """
    return subprocess.getoutput("pbpaste")


def get_clipboard_linux() -> str:
    """Access Linux system clipboard

    There's three tools we can use: `xsel`, `xclip` and `wl-paste`,
    if none of them installed, we use tkinter as final choice.
    """
    commands = {
        "wl-paste": "wl-paste",
        "xclip": "xclip -selection clipboard -o",
        "xsel": "xsel -b",
    }
    for cmd in commands:
        if shutil.which(cmd):
            return subprocess.getoutput(commands[cmd])

    return get_clipboard_tk()


def get_clipboard() -> str:
    """Access system clipboard context"""
    if platform.system() == "Linux":
        return get_clipboard_linux()
    elif platform.system() == "Darwin":
        return get_clipboard_macos()
    else:
        return get_clipboard_tk()


def main():
    is_pipe = not isatty(sys.stdin.fileno())
    if is_pipe:
        text = sys.stdin.read()
    elif len(sys.argv) > 1:
        text = " ".join(sys.argv[1:])
    else:
        text = get_clipboard()
    print(format_text(text))


if __name__ == "__main__":
    main()