Find and replace with variable text - regex

Question

Trying to batch a bunch of conversions with a regex find and replace and I'm not actually sure it's possible, let alone how to achieve this.
Using Sublime text as an editor, open to other tools to accomplish this if possible.
Two sample lines :
Session::flash( 'error', 'Only users with permission may view the directory user.' );
Session::flash( 'error', 'System user ID does not exist.' );
** Desired outcome: **
flash('Only users with permission may view the directory user.')->error();
flash('System user ID does not exist.')->error();
Current Regex that matches:
Session::flash(\s*'error',.* )
Is it possible, that the text lines can be saved and reused in the replace lines? Hoping for a solution along the lines of $variable so that I may replace the strings with something like
** Wishful line: **
flash('$variable')->error();
Thanks folks!

Answer 1

You could use 2 capturing groups and in the replacement referer to those capturing groups.
\bSession::flash\(\s*'([^']+)',\s*('[^']+')\s*\);
In the replacement use:
flash($2)->$1;
Explanation
\bSession::flash\(\s* Match a wordboundary to prevent Session being part of a longer word, then match Session::flash( followed by 0+ times a whitespace char
'([^']+)' Match ', then capture in group 1 matching not a ' using a negated character class, then match ' again
,\s* Match a comma followed by 0+ times a whitespace char
('[^']+') Capture in group 2 matching ', then not ' and again '
\s*\); Match 0+ times a whitespace char followed by );
Regex demo
Result:
flash('Only users with permission may view the directory user.')->error;
flash('System user ID does not exist.')->error;

Answer 2

What you're looking for here is a capture group and a backreference.
In a regular expression anything wrapped in ( and ) is captured for later use by whatever performed the regular expression match, which in this case is Sublime Text. The number of capture groups that are supported varies depending on the regular expression library in use, but you generally get at least 10.
In use, every incidence of () creates a capture, with the first capture being numbered as 1, the second one 2 and so on (generally also the entire match is capture 0). Using the sequence \1 or $1 means "use the contents of the first capture group".
As an example, consider the regular expression ^([a-z]).\1. Breaking it down:
^ - match starting at the start of a line
( - start a capture
[a-z] - match a single lower case letter
) - end a capture
. - match any character
\1 - match whatever the contents of the first capture was
Given this input:
abc
aba
bab
This regular expression matches aba and bab because the first character in both cases is captured as \1 and needs to match later. However abc doesn't match because in that case \1 is a, but the third character is c.
The result of the capture can also be used in the replacement text as well the same way. If you modify your regular expression, you can capture the text you want to keep and use it in the replacement.
As a note, your regex as outlined in your question above doesn't match in Sublime because ( starts a capture group, and thus does not match the ( that's actually in the text. If you're using Sublime and you turn on the Highlight Matches option in the Find/Replace panel, you'll see that your regex is not considered a match.
Find:
Session::flash\(\s*'error'\s*,(.*)\);
Replace:
flash(\1)->error();
Result:
flash( 'Only users with permission may view the directory user.' )->error();
flash( 'System user ID does not exist.' )->error();
This is more or less the regex outlined in your question, except:
The ( and ) in your regex have been replaced with \( and \), which is to tell the regex that this should match a literal ( and not be considered to start a capture.
The .* is changed to (.*) which means "whatever text appears here, capture it for later use.
The replacement text refers to the text captured as \1 and puts it back in the replacement.