正则表达式：捕获成对的花括号

Question

I want to capture matched curly braces.

For example :

Some example text with \\added[author]{text with curly braces{some text}..}

Some example text with \\added[author]{text without curly braces}

Some example text with \\added[author]{text with {}and {} and {}curly braces{some text}..}

Some example text with \\added[author]{text with {}and {} and {}curly braces{some text}..} and extented text with curly braces {}

Expected output:

Some example text with text with curly braces{some text}..

Some example text with text without curly braces

Some example text with text with {}and {} and {}curly braces{some text}..

Some example text with text with {}and {} and {}curly braces{some text}.. and extented text with curly braces {}

ie I want to capture the text between \\added[]{ and } (its relative closing curly braces).Problem with my regex is, I don't know how to capture between the related curly braces.

I tried,

       "/\\\\added\\[.*?\\]{(.[^{]*?)}/s"

I know it ignores if { present in the text. But I don't get an idea how to create a regex to get matched curly braces alone.

Answer 1

To match paired braces you'll want to use a recursive subpattern .

---

Example:

$regex = <<<'REGEX'
/
\\added\[.*?\]                # Initial \added[author]

(                             # Group to be recursed on.
    {                         # Opening brace.

    (                         # Group for use in replacement.

        ((?>[^{}]+)|(?1))*    # Any number of substrings which can be either:
                              # - a sequence of non-braces, or
                              # - a recursive match on the first capturing group.
    )

    }                         # Closing brace.
)
/xs
REGEX;

$strings = [
    'Some example text with \added[author]{text with curly braces{some text}..}',
    'Some example text with \added[author]{text without curly braces}',
    'Some example text with \added[author]{text with {}and {} and {}curly braces{some text}..}',
    'Some example text with \added[author]{text with {}and {} and {}curly braces{some text}..} and extented text with curly braces {}'
];

foreach ($strings as $string) {
    echo preg_replace($regex, '$2', $string), "\n";
}

Output:

Some example text with text with curly braces{some text}..
Some example text with text without curly braces
Some example text with text with {}and {} and {}curly braces{some text}..
Some example text with text with {}and {} and {}curly braces{some text}.. and extented text with curly braces {}

Answer 2

Here, should work

/\\added\[.*\]\{(.*(?:.*\{.*\}.*)*)\}/gU

Explanation

/\\\\added\\ is a Latex tag,

\\[.*\\] is an option of Latex tag,

\\{ open bracket,

(.*(?:.*\\{.*\\}.*)*) is captured text which here we also prevent for recursive {...} or multiple {...} inside our target tag,

\\} close bracket.

Strategy

I do not consider pair of bracket as a recursive form

{ { {...} } }
c b a   a b c

where we have pair a , b and c ,

but I consider them like this!

{ { {...} } }   
a b c   a b c

see: DEMO

The last two examples in my demo also prove that it work correctly.

IMPORTANT: the modifier U suppose to be used here for a purpose of non-greedy quantifier otherwise my regex will not work correctly.

Answer 3

使用以下正则表达式：

\\\\added\\[[^\\]]\*][^\\{]\*{((?:(?:[^\\{\\}]\*\\{[^\\}\\{]\*\\})\*||[^\\}]\*)\*)}

Answer 4

Use this regex

/\\added[^]]*]{([^}]*}[^}]*)}/s

Demo here