正则表达式：将子字符串替换为多种单词形式

Question

Some human languages are distinguished by a reach morphology and developed system of grammatical genders. For instance, in Slavic languages almost every adjective in single grammatical number has three differing forms, according to the amount of genders ( masculine , feminine and neuter ):

white <-> белый (m), белая (f), белое (n)     

In some cases it would be useful to get list of possible word forms while using regular expressions for substring replacement.

Now I'm curious, if Python (or any other scripting language) allows to do things like below ( WARNING : the code snippets below is a Python-like pseudocode , but not a working Python code):

# I would like to handle russian genders like that:  
>>> re.sub(r"Бел.", r"Бел[ый|ая|ое]", "Бел. Берель")
["Белый Берель", "Белая Берель", "Белое Берель"]

# A very artifical example for those who prefer latin:
>>> re.sub(r"Go.", r"Go[ld|lden]", "Go. Ochre")
["Gold Ochre", "Golden Ochre"] 

So can I use regular expressions to get list of matching combinations of words?

Answer 1

No, but you can iterate over the list of suffixes with a list comprehension like this (for Python 2):

>>> suffixes = u'ый|ая|ое'.split('|')
>>> suffixes
[u'\u044b\u0439', u'\u0430\u044f', u'\u043e\u0435']
>>> replacements = [re.sub(u"Бел.", u"Бел%s" % s, u"Бел. Берель") for s in suffixes]
>>> replacements
[u'\u0411\u0435\u043b\u044b\u0439 \u0411\u0435\u0440\u0435\u043b\u044c', u'\u0411\u0435\u043b\u0430\u044f \u0411\u0435\u0440\u0435\u043b\u044c', u'\u0411\u0435\u043b\u043e\u0435 \u0411\u0435\u0440\u0435\u043b\u044c']
>>> for s in replacements:
...     print s
... 
Белый Берель
Белая Берель
Белое Берель

It's a somewhat clearer in Python 3:

>>> suffixes = u'ый|ая|ое'.split('|')
>>> suffixes
['ый', 'ая', 'ое']
>>> >>> suffixes = u'ый|ая|ое'.split('|')
>>> suffixes
[u'\u044b\u0439', u'\u0430\u044f', u'\u043e\u0435']
>>> replacements = [re.sub("Бел.", "Бел%s" % s, "Бел. Берель") for s in suffixes]
>>> replacements
['Белый Берель', 'Белая Берель', 'Белое Берель']

Answer 2

Thanks for your latin example!

I'd use a loop for all possible conditions:

import re 

str = "Go. Ochre"

find_list = [r'Go.']
replace_list = ['Gold', 'Golden']

for value in find_list:
    for item in replace_list:
        print re.sub(value, item, str)

I'm not sure if thats really what you want - and i don't know much about the efficiency of this method.

But it's very readable code, simple to maintain - and could easily be written as a re-usable function for general purpose.

---

...when it comes to this example, it would be better to just replace() the strings - no need for a regex.