Dart regex for capturing groups but ignoring certain similar patterns - regex

Question

I'm trying to capture a group from a string with ~, ~~ and ~~~ symbols. I was successful with extracting single symbols but it doesn't ignore the other occurrences in the string.
This is my code I tried experimenting with:
String f = '~the calculator is on and working~I entered 50 into the calculator'+
'~~I press add button~~holding equal button ~~~The result should be 50';
List<String>givens = f.split(RegExp(r'~+'));
List<String>whens = f.split(RegExp(r'~~+'));
List<String>thens = f.split(RegExp(r'~~~+'));
for(String ss in givens){
print(ss);
}
print('xxxxxxxxxxxx');
for(String ss in whens){
print(ss);
}
print('xxxxxxxxxxxx');
for(String ss in thens){
print(ss);
}
Which will result with:
The givens capture group also captured the ones with ~~ and ~~~ which is not intended.
The whens capture group also captured the ones single ~ which made it very confusing.
Lastly, the thens capture group also captured the others which is also not intended.
I only need to capture the strings starting with the specific pattern but will stop when they see a different one.
Example: givens should only capture 'the calculator is on and working' and 'I entered 50 into the calculator' only.
Any hints or help is greatly appreciated!

Answer 1

I think the problem is that you started off by splitting the string into pieces. But it might be easier to search for the elements with a pattern that will look for some text preceeded with either one, two or three ~ chars.
This can be done with regex positive lookbehind patterns.
Typically, if you want to find a string preceeded by one tild then you have to avoid that it matches if we have other tilds before it.
Find givens
(?<=(?:[^~]|^)~)[^~]+ would be the pattern to find only givens.
Test it here: https://regex101.com/r/9WLbM3/2
Explanation
[^~] means search for any character which is not a ~. This is because [abc] means any char which is in the list, so a, b or c. If you add the ^ char at the beginning of the list then it means "not these chars".
[^~]+ means search for one or multiple times a character which is not ~. This will capture phrases between the tilds.
A positive lookbehind is done with (?<=something present). We want to search for a tild so we would put (?<=~) as positive lookbehind. But the problem is that it will also match the ones with several tilds in front. To avoid that we can say that the tild should either be prefixed by ^ (meaning the beginning of a string) or by [^~] (meaning not a tild). To say "either this or that", we use the syntax (this|that|or even that). But using parenthesis will capture the content and we don't need that. To disable group capturing we can add ?: at the beginning of the group, leading finally to (?:[^~]|^) meaning either a non-tild char or the beginning of the string, without capturing it.
Find whens and thens
The regular expression is almost the same. It's just that we replace ~ by ~{2} or ~{3}.
Pattern for whens: (?<=(?:[^~]|^)~{2})[^~]+
Pattern for thens: (?<=(?:[^~]|^)~{3})[^~]+