匹配任意分割多行的字符串

Question

Is there a way in regex's to match a string that is arbitrarily split over multiple lines - say we have the following format in a file:

msgid "This is "
"an example string"
msgstr "..."

msgid "This is an example string"
msgstr "..."

msgid ""
"This is an " 
"example" 
" string"
msgstr "..."

msgid "This is " 
"an unmatching string" 
msgstr "..."

So we would like to have a pattern that would match all the example strings, ie: match the string regardless of how it's split across lines. Notice that we are after a specific string as shown in the sample, not just any string. So in this case we would like to match the string "This is an example string" .

Of course we can can easily concat the strings then apply the match, but got me wondering if this is possible. I'm talking Python regex's but a general answer is ok.

Answer 1

Do you want to match a series of words? If so, you could look for words with just spaces (\\s) in between, since \\s matches newlines and spaces alike.

import re

search_for = "This is an example string"
search_for_re = r"\b" + r"\s+".join(search_for.split()) + r"\b"
pattern = re.compile(search_for_re)
match = lambda s: pattern.match(s) is not None

s = "This is an example string"
print match(s), ":", repr(s)

s = "This is an \n example string"
print match(s), ":", repr(s)

s = "This is \n an unmatching string"
print match(s), ":", repr(s)

Prints:

True : 'This is an example string'
True : 'This is an \n example string'
False : 'This is \n an unmatching string'

Answer 2

This is a bit tricky with the need for quotes on every line, and the allowance of empty lines. Here's a regex that matches the file you posted correctly:

'(""\n)*"This(( "\n(""\n)*")|("\n(""\n)*" )| )is(( "\n(""\n)*")|("\n(""\n)*" )| )an(( "\n(""\n)*")|("\n(""\n)*" )| )example(( "\n(""\n)*")|("\n(""\n)*" )| )string"'

That's a bit confusing, but all it is is the string you want to match, but it starts with:

(""\n)*"

and has replaces the spaces between each word with:

(( "\n(""\n)*")|("\n(""\n)*" )| )

which checks for three different possibilities after each word, either a "space, quote, newline, (unlimited number of empty strings) quote", or that same sequence but more the space to the end, or just a space.

A much easier way to get this working would be to write a little function that would take in the string you are trying to match and return the regex that will match it:

def getregex(string):
    return '(""\n)*"' + string.replace(" ", '(( "\n(""\n)*")|("\n(""\n)*" )| )') + '"'

So, if you had the file you posted in a string called "filestring", you would get the matches like this:

import re

def getregex(string):
    return '(""\n)*"' + string.replace(" ", '(( "\n(""\n)*")|("\n(""\n)*" )| )') + '"'

matcher = re.compile(getregex("This is an example string"))

for i in matcher.finditer(filestring):
    print i.group(0), "\n"

>>> "This is "
    "an example string"

    "This is an example string"

    ""
    "This is an "
    "example"
    " string"

This regex doesn't take into account the space you have after "example" in the third msgid, but I assume this is generated by a machine and that's a mistake.