♻️ Update regex to be lazy and handle ! in text

[jeremyfelder]

Description

This PR reverts the regex to use a match all (.+) token with the addition of being lazy (? - match as little as possible, consuming as needed) instead of greedy (match as much as possible, giving back as needed) which allows for using an ! in the text.

Regex Explanation

[!]{2} - match two leading exclamation points
(?P<class>.+?)[|^] - match anything once or more (lazily) until a | or ^ is found
(?P<text>.+?(?=[!]{2})) - match anything once or more (lazily) until two consecutive !s are found
(?=(?P<lookahead>[!]{3}))? - check if there are three consecutive !s
(?(lookahead)(?P<exclaim>!)) - if three consecutive !s were found match one !
[!]{2} - match two trailing exclamation points

Tested Cases

• multiple spans in a single line
• using an ! anywhere in the text
• using an ! anywhere in the class

NOTE: The edge case of an exclamation point at the end of the text (i.e. having 3 ! in a row) required appending an ! to the text

See here for tested cases: https://regex101.com/r/BZnHIO/2

NOTE:
Markdown is using the python standard library re as its regex library. re doesn't allow for conditional expressions unless the condition is a previous capture group - https://docs.python.org/3/library/re.html?highlight=(?(id/name). This forced a conditional matching capture group named "lookahead" which is then used in the conditional expression instead of having the lookahead inline in the conditional.

[wilhelmer]

Looks good to me, thanks for contributing!