[Python] Writing upside-down characters

[unutbu]

I would like to make a command-line python program to translate [a-zA-Z] strings into their "upside-down" equivalent:

˙ʇuǝlɐʌınbǝ "uʍop-ǝpısdn" ɹıǝɥʇ oʇuı sƃuıɹʇs [z-ɐz-ɐ] ǝʇɐlsuɐɹʇ oʇ ɯɐɹƃoɹd uoɥʇʎd ǝuıl-puɐɯɯoɔ ɐ ǝʞɐɯ oʇ ǝʞıl plnoʍ ı

I used this to list all the utf8 characters:

#!/usr/bin/env python
codec='utf8'
for num in range(32,65536):
    achr=unichr(num)
    print("'':%s, # %s"%(num,achr.encode(codec))) 

Using that list, I was able to come up with this:

#!/usr/bin/env python
# -*- encoding: utf-8 -*-

import sys
rot180={
    'a':592, # ɐ
    'b':5227, # ᑫ
    'd':1088, # р
    'e':600, # ɘ
    'h':1063, # Ч
    'l':645, # ʅ    
    'p':599, # ɗ
    'r':633, # ɹ
    't':647, # ʇ
    'v':8743, # ∧
    'w':653, # ʍ
    'y':654, # ʎ
    'A':5578, # ᗊ
    'B':5626, # ᗺ
    'C':390, # Ɔ
    'D':5613, # ᗭ
    'E':8707, # ∃
    'O':2848, # ଠ
    'R':5512, # ᖈ
    'T':639, # ɿ
    'U':1055, # П
    'V':923, # Λ
    '3':5620, # ᗴ
    }

input_str=sys.argv[1]
print(''.join([unichr(rot180.get(achr,ord(achr))).encode('utf8') for achr in input_str])) 

In so far as it goes, it works okay:

% upsidedown.py Howdy
Hoʍрʎ

The problem is, I don't have a complete set of upside-down characters.

According to http://www.sevenwires.com/play/UpsideDownLetters.html, the characters come from the "Latin Extended" and "International Phonetic Alphabet" character sets.

I've looked these character sets here: file:///usr/share/doc/python2.5/html/lib/standard-encodings.html but couldn't find them.

Do you know what encoding I should be using to access upside-down characters?

[simeon87]

I don't think that every arbitrary character has an upside-down equivalent in some character set so I think the best you can do is to manually create a mapping between characters and their upside-down equivalent.

For example, in the source of the webpage that you mentioned, it also uses a table to lookup the upside-down character.

[unutbu]

Thanks simeon87. Indeed, I don't mind making a dictionary by hand, even though it feels icky

I've only made a cursory run through the utf-8 character set to build the rot180 dictionary as it stands so far.

The problem is, I'm not sure I'm going to find a complete set of upside down characters even for [a-z] in utf8.

http://www.sevenwires.com/play/UpsideDownLetters.html demonstrates that javascript can do it. I'd like to be able to do it in Python.

Should I just keep looking through my list of utf8, or am I looking in the wrong place?

[unutbu]

Oh, I'm so silly. The javascript code is visible if you click View>Page Source... LOL

[unutbu]

Here is the finished script. It can do everything http://www.sevenwires.com/play/UpsideDownLetters.html can do, but from the command-line.
(I also added a few capital letters.)

Many thanks to simeon87 for his help.

#!/usr/bin/env python
# -*- encoding: utf-8 -*-
import sys

__usage__='''
upsidedown.py "The quick brown fox jumps over the lazy dog."
˙ƃop ʎzɐʅ ɘɥʇ ɹɘʌo sdɯuɾ xoɟ uʍoɹq ʞɔıuq ɘɥﻠ
'''

rot180={
    'a' : u'\u0250', # ɐ
    'b' : u'q',
    'c' : u'\u0254', # ɔ
    'd' : u'p',
    'e' : u'\u0258', # ǝ
    'f' : u'\u025F', # ɟ
    'g' : u'\u0183', # ƃ
    'h' : u'\u0265', # ɥ
    'i' : u'\u0131', # ı
    'j' : u'\u027E', # ɾ
    'k' : u'\u029E', # ʞ
    'l' : u'\u0285', # ʅ    
    'm' : u'\u026F', # ɯ
    'n' : u'u', # u
    'p' : u'd',
    'r' : u'\u0279', # ɹ
    't' : u'\u0287', # ʇ
    'u' : u'n',
    'v' : u'\u028C', # ʌ
    'w' : u'\u028D', # ʍ
    'y' : u'\u028E', # ʎ
    '.' : u'\u02D9', # ˙
    '[' : u']',
    '(' : u')',
    '{' : u'}',
    '?' : u'\u00BF', # ¿
    '!' : u'\u00A1', # ¡
    "\'" : u',', 
    '<' : u'>',
    '_' : u'\u203E', # ‾
    '"' : u'\u201E', # „
    '\\' : u'\\',
    ';' : u'\u061B', # ؛
    '\u203F' : u'\u2040', # ‿ --> ⁀
    '\u2045' : u'\u2046', # ⁅ --> ⁆
    '\u2234' : u'\u2235', # ∴ --> ∵
#     'A':u'\u15CA', # ᗊ (viewable in emacs, but not in terminal)
#     'B':u'\u15FA', # ᗺ (viewable in emacs, but not in terminal)
    'C':u'\u0186', # Ɔ (viewable in emacs, but not in terminal)
#     'D':u'\u15ED', # ᗭ (viewable in emacs, but not in terminal)
    'E':u'\u2203', # ∃ (viewable in emacs, but not in terminal)
#     'N':u'\u2d4d', # ⵍ (viewable in terminal, but not in emacs)
#     'O':u'\u0B20', # ଠ (viewable in emacs, but not in terminal)
#     'R':u'\u1588', # ᖈ (viewable in emacs, but not in terminal)
    'T':u'\uFEE0', # ﻠ
    'U':u'\u041F', # П
    'V':u'\u039B', # Λ
#     'V':u'\u2d37', # ⴷ (viewable in terminal, but not in emacs)
#     'Y':u'\u2d43', # ⵃ (viewable in terminal, but not in emacs)
#     '2':u'\u2D52', # ⵒ (viewable in terminal, but not in emacs)
#     '3':u'\u15F4', # ᗴ (viewable in emacs, but not in terminal)
    }

input_str=sys.argv[1]
print(''.join(reversed([rot180.get(achr,achr) for achr in input_str]))) 

If anyone can explain why certain unicode characters are viewable in emacs, but not from a gnome-terminal, and vice versa, I'd really like to know. (See the commented-out source code for examples.)