Allow to specify characters that have to be converted to character class

[NightWatch0]

First of all thank you for this great tool.
When using, I often need to convert text into more detailed character classes, not just non-digits or non-blank characters.
Is it possible to customize the range of characters to be converted into character classes, like [a-e\d], [①-⑨⒈-⒙] or specific languages such as Chinese and Japanese.
For example, if the source text is 我的名字是Tom, I hope to get the regular expression [\u{4e00}-\u{9fa5}]{5}\w{3} instead of \w{8}, by specifying character class [\u{4e00}-\u{9fa5}].
And I want to specify the maximum and minimum length of repeated substrings. Sometimes I get results like (\w{5}|\w{7,8}|\w{10,17}), but the regular expression I expected is (\w{3,20}). So I hope to be able to specify the minimum and maximum repetition times of the substring, or combine the repetition times into an interval instead of multiple branches.
I think these two points can be specified together, using multiple formats similar to \w{3,20} to specify characters that must be converted into character classes.

[pemistahl]

Thanks for your ideas @NightWatch0. I will think about how to improve grex even further and I will keep your suggestions in mind.

[suliveevil]

Awesome idea.