MobileRead Forums - Regex examples

Page 45 of 46

« First

Show 40 post(s) from this thread on one page

MobileRead Forums (https://www.mobileread.com/forums/index.php)

- Sigil (https://www.mobileread.com/forums/forumdisplay.php?f=203)

- - Regex examples (https://www.mobileread.com/forums/showthread.php?t=167971)

leschek

09-16-2020 07:09 AM

I hope this is correct topic to post to.

In my language we use one letter prepositions and conjunctions (a, i, o, u, k, s, v, z) which shouldn't be on the end of lines. Here is example from book I try to "epubize":
"spatřil člun a v tom člunu". (translation: "he saw a boat and in that boat")
What I want is to find letters "a" and "v" and replace them with no-break space to connect them to following word. I have this regex (I found somewhere)

Code:

\s([aiouksvz])\s

for searching, but it finds only the first letter and then skip the second one. I tried to change the searching direction in Sigil to "up", but it doesn't help. I guess there must be some problem with regex I'm using.

I also tried this example and again it finds only every second letter:

Code:

some words a s i k v some words

It seems there is some problem with space between letters. When I double it to:

Code:

some words a s i k v some words

the searching works.

davidfor

09-16-2020 08:00 AM

I think you want:

Code:

\b([aiouksvz])\s

That will pick a single letter followed by whitespace.

leschek

09-16-2020 08:15 AM

Thank you, it works partialy, but it does find also parts of html code as

Code:

<a href...

and words ending with searched characters with previous character from non English alphabet as nás, při etc.

DiapDealer

09-16-2020 12:36 PM

Quote:

Originally Posted by leschek (Post 4035291)

Thank you, it works partialy, but it does find also parts of html code as

Code:

<a href...

and words ending with searched characters with previous character from non English alphabet as nás, při etc.

I'm tackling your exceptions in reverse order. :)

To make \b honor unicode codepoints, turn on the Unicode Character Properties flag with (*UCP)

So the above"

Code:

\b([aiouksvz])\s

becomes:

Code:

(*UCP)\b([aiouksvz])\s

This should exclude the 'i' and the 'a' characters in your 'nás' and 'při' examples

To make the expression ignore the character class matches that immediately follow an angled (x)html bracket (<) you can use a negative lookbehind. Something like:

Code:

(*UCP)(?<!\<)\b([aiouksvz])\s

should ignore the 'a' and 'i' characters used in (x)html's anchor and italic tags.

The (*UCP) flag and the (?<!\<) lookbehind are not captured groups despite the appearance. So the replacement you're looking for will still be something like:

Code:

\1 

leschek

09-16-2020 05:53 PM

Quote:

Originally Posted by DiapDealer (Post 4035380)

Code:

(*UCP)(?<!\<)\b[^<]([aiouksvz])\s

So the replacement you're looking for will still be something like:

Code:

\1 

Thank you for your time and explanation, but unfortunately it's working partially again. It ignores the html code (a href, i), which is great, but it doesn't find all letters I need to find. For example in sentence "spatřil člun a v tom člunu", it should find letters "a" and "v", but it only finds "a" and ignores "i". It also find some two-letters words as "na", "do" or in English "as" and "is".

DiapDealer

09-16-2020 08:13 PM

Apologies... I pasted the wrong full expression. It had an extraneous (and incorrect) negative character class that I was testing out.

This is the one that works for me for all of your examples so far:

Code:

(*UCP)(?<!\<)\b([aiouksvz])\s

leschek

09-17-2020 06:05 AM

Quote:

Originally Posted by DiapDealer (Post 4035562)

This is the one that works for me for all of your examples so far:

Code:

(*UCP)(?<!\<)\b([aiouksvz])\s

Thank you very much. I tried it on a few pages and it seems it's working as expected. Awesome.

ShdwMnrch

09-18-2020 09:54 AM

Hello,
I need help on regex, i have lines like these

Code:

<p>– Wahahahaha!</p>



<p>Grasha got drunk, raged and got on the table.</p>



<p>– Wahahahaha! This is a celebration party! Drink and sing guys!</p>

The dialogues are preceded with "– " but I wanna wrap the dialogues with 「」these characters. I had no problem replacing the
"–" to " 「" but I having problems replacing "" when "– " is present in the beginning of the lines.

I have tried the regex search of:

Code:

(?<=– .*)<\/p>

It should find only the "" of the 1st and 3rd line, ignoring the 2nd line. But it seems like sigil regex doesn't support positive lookbehind and it returns nothing. Please help for any workaround. Thanks!

BeckyEbook

09-18-2020 12:09 PM

Try (as long as I understand your problem correctly):

Code:

(?<=– )(.+)

Replace:

Code:

\1」

DiapDealer

09-18-2020 12:18 PM

Sigil's PCRE regex engine certainly supports positive lookbehinds. It just doesn't support variable-length lookbehinds--positive or negative. It's a known limitation of the PCRE engine.

Use \K to simulate a variable-length lookbehind:

Code:

–( .*?)\K<\/p>

Make sure "Minimal Match" is unchecked when using the above expression.

More on the use of \K here: https://www.regular-expressions.info/keep.html

ShdwMnrch

09-18-2020 12:22 PM

Quote:

Originally Posted by BeckyEbook (Post 4036157)

Try (as long as I understand your problem correctly):

Code:

(?<=– )(.+)

Replace:

Code:

\1」

It worked, didn't thought of tinkering with the replace value. thank you!

DiapDealer

09-18-2020 12:30 PM

Not sure why it looks like there's an extra space in my above expression. It seems to copy and work fine, though. *shrug*

ShdwMnrch

09-18-2020 12:31 PM

Quote:

Originally Posted by DiapDealer (Post 4036162)

Code:

–( .*?)\K<\/p>

Make sure "Minimal Match" is unchecked when using the above expression.

More on the use of \K here: https://www.regular-expressions.info/keep.html

I see. I just started to learn how to use regex so that helps a lot, thank you

BillPearl

09-18-2020 05:06 PM

Suggestion

\[\s][a,i,o,u,k,s,v,zç]\[\s]

will handle '<a ' case finds space before and after letter. You may want to run this with just one letter at a time using Replace All

hobnail

10-12-2020 03:34 PM

I don't understand why this isn't working; my search string is:

<a id="Page_([xvi]+)|([\d]+)" class="x-ebookmaker-pageno" title="\[([xvi]+)|([\d]+)\]"></a>

When the file contains

<a id="Page_iv" class="x-ebookmaker-pageno" title="[iv]"></a>

and I click on the Find button, it highlights only

<a id="Page_i

What's wrong with my regexp?

All times are GMT -4. The time now is 07:52 PM.

Page 45 of 46

« First

Show 40 post(s) from this thread on one page