PHP/PCRE regex starts with on recursive pattern

Question

I'm trying to match markdown tags with recursive.

Input Syntax

(TYPE: VALUE ATTR_KEY: ATTR_VALUE)

Note that syntax should be starts with: [a-z0-9_-]+:

Sample Inputs:

(image: sky.jpg)
(image: sky.jpg caption: Sky (Issue This) View)
(link: https://stackoverflow.com text: Stack Overflow)
(link: https://stackoverflow.com text: Stack Overflow rel=nofollow)
(video: http://www.youtube.com/watch?v=49Kh1mS4Fhs)

Currently using following regex:

(?=[^\]])\([a-z0-9_-]+:.*?\)

But issue coming from here, because match with:

(image: sky.jpg caption: Sky (Issue This)

Expected match:

(image: sky.jpg caption: Sky (Issue This) View)

If parentheses are used again in parentheses, it does not match exactly.

I tried following recursive patterns and works but i need to restrict starts with characters.

(?s)\((?:[^()]+|(?R))*+\)
\((?:[^)(]+|(?R))*+\)

Can't you just match whole tag and then process it with PHP splitting string? — Justinas
– Justinas, Commented Mar 4, 2020 at 13:37
@Justinas I process only the tag ones, not inside every parenthesis. — Bora
– Bora, Commented Mar 4, 2020 at 13:39

Wiktor Stribiżew · Accepted Answer · 2020-03-05 10:50:32Z

2

You should use a positive lookahead to match sure the match starts with that pattern, but you will have to wrap the whole parentheses matching pattern within another capturing group and use a (?1) subroutine instead of (?R) to only recurse that pattern, not the whole regex:

(?=\([a-z0-9_-]+:)(\((?:[^()]+|(?1))*+\))
^^^^^^^^^^^^^^^^^^^            ^^^^     ^

See the regex demo.

Details

(?=\([a-z0-9_-]+:) - a positive lookahead that requires (, 1+ lowercase ASCII letters, digits, underscores or hyphens followed with : immediately to the right of the current location
(\((?:[^()]+|(?1))*+\)) - Capturing group 1 (it will be recursed later):
- \( - (
- (?:[^()]+|(?1))*+ - 1+ repetitions of 1+ any chars other than ( and ) or the whole Group 1 pattern (recursed)
- \) - )

In case you want to also support smileys, you may add their specific patterns in the alternation group where the regex subroutine resides, as the first alternative:

(?=\([a-z0-9_-]+:)(\((?::[)(]|[^()]|(?1))*+\))
                        ^^^^^

I add :[)(] that matches :) or :( and removed + from after [^()] so as to be able to check the string inside nested parentheses character by character.

Feel free to adjust it to your needs, or add more smiley patterns.

See this regex demo with the (?=\([a-z0-9_-]+:)(\((?::(?:[()pPDd*oO]|'\()|<3|;\)|[^()]|(?1))*+\)) regex.

edited Mar 5, 2020 at 10:50

answered Mar 4, 2020 at 13:38

Wiktor Stribiżew

631k41 gold badges502 silver badges632 bronze badges

Sign up to request clarification or add additional context in comments.

5 Comments

Bora Over a year ago

Naturally, we use missing parentheses do not match. But it doesn't match when we use emoji, though it's not very important. Do you think there is a simple solution to this? regex101.com/r/1Tti14/4

Bora Over a year ago

Thank you for your great help.This was also very good. But it was true that it didn't match when there were actually missing brackets. It should be able to use emoji in parentheses, but should not match when missing parentheses. Is this hard?

Bora Over a year ago

Hmm, still matching with Missing bracket usage :( An error can occur when I try to parse it.

Wiktor Stribiżew Over a year ago

I misunderstood. So, Matching bracket should not match and Emoji should? Like in (?=\([a-z0-9_-]+:)(\((?::[)(]|[^()]|(?1))*+\)) (demo)?

Bora Over a year ago

That's exactly it. How can you write so great. I really thank you again.

Collectives™ on Stack Overflow

PHP/PCRE regex starts with on recursive pattern

1 Answer 1

5 Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

5 Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related