Using regex to fix broken paragraph in Chinese

[icearch]

Originally Posted by ElMiko

See the attached txt file. As I said before, I wasn't able to discern from your screenshots what text editor you were using, but that may be contributing to the issue if they are using a different regex engine (or different version of the same regex engine).

My screenshot all comes from Sigil.

And your code does not work.



The right bottom says "didn't find any matching to replace"



You can read this. And this is that I fixed </i> with <\/i> so no error.

[ElMiko]

Originally Posted by icearch

My screenshot all comes from Sigil.

And your code does not work.

Attachment 223602

The right bottom says "didn't find any matching to replace"

Attachment 223603

You can read this. And this is that I fixed </i> with <\/i> so no error.

icearch... you're killing me, bud.

The reason the search didn't match anything is because you changed the text from plain text to html! You stated your original example was plain text so I modified my regex to accommodate that instead of an HTML doc.

All the lines in the new text you showed in the last post are within <p> tags. So of course the search that I modified to match your plain text example won't match anything.

Plain Text:


HTML:


As you can see, depending on which version of the regex search i use, i get the same number of matches in both the plain text and html samples. And it doesn't require (nor should it include) a backslash before the "/i" because a forward slash isn't a special character in regex.

---

P.S. What is the regex validation checker you're using?

[icearch]

My bad, I see you tend to use html tags, so I thought you give me regex works with tags. I'm not a competent regex user so I did not spot that.

And your code works great, thank you.

I use

https://regex101.com/

and

https://regexr.com/

to test regex.

[ElMiko]

You mentioned you're still polishing your regex skills (as we all are), so I thought I'd simplify the search and try to break down what all the elements of the plain text version of the search are doing.

(\p{L}(|[,;-])|,”|[MD][rs]\.|Mrs\.|(”|—)(?=\n\p{Ll}))\n

It will match any thing that:

(
\p{L}(|[,;-])| — is a letter followed by nothing, a comma, a semi-colon, or a hyphen (FYI, I don't include colons in this search because I find that more often than not, a colon used in a fiction is supposed to be followed by a paragraph break, but if you want to add it in, you can, of course), or
,”| — is a comma followed by a curly closing quote, or
[MD][rs]\.| — is any version of "Mr., Dr., or Ms.", or
Mrs\.| — is "Mrs.", or
(”|—)(?=\n\p{Ll}) — is a closing curly quote or an em dash [provided it is followed by a line break AND a lowercase letter],
)

and

\n — is followed by a line break.

For the first instance of "\p{L}", you should be able to replace it with "\p{Han}" and Chinese equivalents of punctuation marks that shouldn't denote a paragraph break (e.g. the Chinese comma). But it won't work for the second-to-last bit of regex—i.e. (”|—)(?=\n\p{Ll})— because there's no such thing as a lowercase character, and that's a critical limiter in the functioning of that search element.

[icearch]

Thank you for the information, it helps alot.