その透明な文字に混じらず、見つけ出すんだ。 ref: http://qiita.com/kitsuyui/items/12db383f5e5971f32b08

file4.txt

U+20    SPACE
U+A0    NO-BREAK SPACE
U+2000  EN QUAD
U+2001  EM QUAD
U+2002  EN SPACE
U+2003  EM SPACE
U+2004  THREE-PER-EM SPACE
U+2005  FOUR-PER-EM SPACE
U+2006  SIX-PER-EM SPACE
U+2007  FIGURE SPACE
U+2008  PUNCTUATION SPACE
U+2009  THIN SPACE
U+200A  HAIR SPACE
U+205F  MEDIUM MATHEMATICAL SPACE
U+3000  IDEOGRAPHIC SPACE
U+FFA0  HALFWIDTH HANGUL FILLER

file5.txt

U+FFA0  HALFWIDTH HANGUL FILLER

file6.txt

U+3164  HANGUL FILLER

find_invisible.py

import unicodedata

from PIL import Image
from PIL import ImageDraw
from PIL import ImageFont


FONT = ImageFont.truetype('./unifont-7.0.06.ttf', 32)


def main():
    assert_principles()

    for char in invisible_characters():
        try:
            name = unicodedata.name(char)
        except ValueError:
            name = ''

        line = 'U+{:X}\t{}'.format(ord(char), name)
        print(line)


def invisible_characters():
    '''空白として描画された文字のみを返却する'''
    empty_draw_bytes = letter_draw_bytes(' ')
    for char in unicode_characters():
        draw_bytes = letter_draw_bytes(char)
        if draw_bytes == empty_draw_bytes:
            yield char


def unicode_characters():
    '''Unicode コードポイントに含まれる文字全てを返すジェネレータ
ただし、 PUA (Private Use Area) は含めない。
    '''

    for codepoint in range(0, 0xE000):
        yield chr(codepoint)

    for codepoint in range(0xF8FF + 1, 0xF0000):
        yield chr(codepoint)

    for codepoint in range(0xFFFFD + 1, 0x100000):
        yield chr(codepoint)

    for codepoint in range(0x10FFFD + 1, 0x10FFFF):
        yield chr(codepoint)


def letter_draw_bytes(char):
    '''空白の画像に文字を描画した際の bytes を取得する'''

    image = Image.new('RGB', (32, 32), (255, 255, 255))
    draw = ImageDraw.Draw(image)
    draw.text((0, 0), char, font=FONT, fill='#000')
    return image.tobytes()


def assert_principles():
    '''以下の原則が成立することをいくつかの例でチェック'''
    f = letter_draw_bytes

    assert f('あ') == f('あ')  # 同じ文字の bytes は同じ

    assert f('') == f(' ') == f('　')  # スペースは空文字列と同じ bytes になる

    assert f('あ') != f('い')  # 見た目が異なる文字の bytes は異なる


if __name__ == '__main__':
    main()

tsv

U+9 
U+A 
U+B 
U+C 
U+D 
U+20    SPACE
U+85    
U+A0    NO-BREAK SPACE
U+2000  EN QUAD
U+2001  EM QUAD
U+2002  EN SPACE
U+2003  EM SPACE
U+2004  THREE-PER-EM SPACE
U+2005  FOUR-PER-EM SPACE
U+2006  SIX-PER-EM SPACE
U+2007  FIGURE SPACE
U+2008  PUNCTUATION SPACE
U+2009  THIN SPACE
U+200A  HAIR SPACE
U+2028  LINE SEPARATOR
U+2029  PARAGRAPH SEPARATOR
U+202F  NARROW NO-BREAK SPACE
U+205F  MEDIUM MATHEMATICAL SPACE
U+3000  IDEOGRAPHIC SPACE
U+180E  MONGOLIAN VOWEL SEPARATOR
U+200B  ZERO WIDTH SPACE
U+200C  ZERO WIDTH NON-JOINER
U+200D  ZERO WIDTH JOINER
U+2060  WORD JOINER
U+2061  FUNCTION APPLICATION
U+2062  INVISIBLE TIMES
U+2063  INVISIBLE SEPARATOR
U+2064  INVISIBLE PLUS
U+3164  HANGUL FILLER
U+FFA0  HALFWIDTH HANGUL FILLER
U+FEFF  ZERO WIDTH NO-BREAK SPACE

どうみてもAB

A⁠B

実際のところ

A{U+2060}B