Regex examples

[1v4n0]

More idle questions: some way to match everything except a specific string, or a set of specific strings?

[eschwartz]

Yep, with a negative lookaround (see HERE for the first explanation I saw, way back, of this trick):

((?!string).)*

Matches the dot-match-all, if NOT followed by a "string", and then repeats the search ad nauseam.


regex is not meant to do this sort of stuff, but it can be shoehorned into it.

There will be a performance hit, but how often does that become a practical matter?

[1v4n0]

Weird. This finds the first string in the xhtml file, but when I re-click "find" it skips to the next file, instead of finding the second string of the file, i.e. the second interval between the "excluded" strings. If I position the pointer after the excluded string it finds the right stuff, but I have to do it manually each time.

Anyway ty. What I'm thinking about is even more complicated: I got a text with lots of line-breaks, <br/>, and I want to wrap the first of each three lines into a span.

I could achieve the same goal with some <p> and nth-child css, but they're not as widely supported.

Hope this all makes sense

[eschwartz]

I suspect you could also do the same with manually declared classes what difference does it make whether they are wrapped in spans or ps?

[1v4n0]

Well I found out (at my expense ) that I can't nest p's, though calibre does not make it clear immediately, and what I'm working on is already in a p. But anyway, this is details. Even if I were to use <p>'s, what difference would it make? I'd still need to select the first of every three.

(This is for some poetry, in which the first verse of evry tercet is more indented than the other two.)

[1v4n0]

I got some footnotes. Some of them are one paragraph only, others are longer. They only have "backlinks" to the main text at the beginning of the note

  <p><a href="#13" id="13-fn">13.</a> bla bla bla</p>

<p>alb alb alb</p>

I'd like to have a backlink at the end too (makes more sense, especially if the note is very long).

  <p><a href="#13" id="13-fn">13.</a> bla bla bla</p>

<p>alb alb alb<a href="#13">back to text</a></p>

So I need a regex that searches for all the instances of a paragraph starting with a backlink, followed by the closing </p> and zero or more paragraphs until it finds another one that starts with another backlink.

I tried to fiddle with sigil a bit, but to no avail.

Ty

[eschwartz]

Try:

Find:

(<p(?: [^>]+)?>)(<a [^>]+>[^<>]+</a>)((?:(?!</?a(?: [^>]+)?>).)*)(?=</p>)

Replace:

\1\2\3\2

Match 1 is the opening paragraph tag. We could use a positive lookbehind and skip replacing that, but then we can't match optional classes and stuff.

Match 2 (in blue above) captures the link and link text:

<a [^>]+>[^<>]+</a>

Match 3 (in red above) captures the note content, by matching a string that doesn't contain an "a" tag. I lurves my negative lookarounds :

(?:(?!</?a(?: [^>]+)?>).)*

Finally, we look ahead to spot the closing paragraph tag.

[1v4n0]

Doesn't work. Just finds all the single paragraphs between

<p><a href=

and

</p>

Which is the same thing I get with

<p><a href=(.*?)</p>

I need something that selects everything from one occurrence of

<p><a href=

to the following one. There are "line breaks" in between (I don't know if that's the right word. Line breaks, something like this

</p>

  <p>

)

THIS WORKS, but it's not very clean work:

First, find all the occurrences of \s+ and replace with space.

Then look for

<p><a href=(.*?)<p><a href=(.*?)

The code is barely readable this way though, as there are no line breaks anymore.

Ty

[eschwartz]

What about:

Find:

(<p(?: [^>]+)?>)(<a [^>]+>[^<>]+</a>)((?:(?!</?[ap](?: [^>]+)?>).)*)(?=</p>)

Replace:

\1\2\3\2

I should've made it find text without "a" or "p" tags.

EDIT: No, not sure what I was thinking of at all. That makes no sense...

[1v4n0]

Nope, still the same

[eschwartz]

Tested in calibre editor on the content:

    <p><a href="#13" id="13-fn">13.</a> bla bla bla</p>

<p>alb alb alb</p>

  <p><a href="#14" id="14-fn">14.</a> bla bla bla</p>

<p>alb alb alb</p>

  <p><a href="#15" id="15-fn">15.</a> bla bla bla</p>

<p>alb alb alb</p>

  <p><a href="#16" id="16-fn">16.</a> bla bla bla</p>

<p>alb alb alb</p>

And my original code worked just fine, becoming:

    <p><a href="#13" id="13-fn">13.</a> bla bla bla</p>

<p>alb alb alb<a href="#13" id="13-fn">13.</a></p>

  <p><a href="#14" id="14-fn">14.</a> bla bla bla</p>

<p>alb alb alb<a href="#14" id="14-fn">14.</a></p>

  <p><a href="#15" id="15-fn">15.</a> bla bla bla</p>

<p>alb alb alb<a href="#15" id="15-fn">15.</a></p>

  <p><a href="#16" id="16-fn">16.</a> bla bla bla</p>

<p>alb alb alb<a href="#16" id="16-fn">16.</a></p>

So I'm not sure what problem you were having.
As for post #474 I am not sure what, if any, thought processes were running through my mind!

[leftright]

Can someone help me with the syntax for the following please,
I want to remove the numbers between these sentences and then append the sentences as in below,

"The cat jumped over the moon many times before astronauts' or argonauts'

37

ever manged to do so, even before any cow attempted this amazing feat, and when it succeeded, milk shake for everyone."

It appears as <p class="calibre1">37</p>

It should be;
"The cat jumped over the moon many times before astronauts' or argonauts' ever manged to do so, even before any cow attempted this amazing feat, and when it succeeded, milk shake for everyone."

Many thanks.

[Doitsu]

There are probably more elegant solutions, but the following expression should do the trick assuming that the text before and after the page number is also wrapped in paragraph tags:

Find:<p([^>]*)>(.*?)</p>\s*<p[^>]*>\d+</p>\s*<p[^>]*>(.*?)</p>

Replace:<p\1>\2 \3</p>

[doubleshuffle]

Originally Posted by leftright

"The cat jumped over the moon many times before astronauts' or argonauts' ever manged to do so, even before any cow attempted this amazing feat, and when it succeeded, milk shake for everyone."

Totally unrelated to the page number: You really want to get rid of the apostrophes after astronauts and argonauts.

[gipsy]

Hi,
i user the following regex to check for hypen errors

FIND:(?<![\x{0370}-\x{03FF}\x{1F00}-\x{1FFF}])([\x{0370}-\x{03FF}\x{1F00}-\x{1FFF}]+)-(?!\ |\,|\1)
REPLACE:\1

In greek language we have sometimes a - before a capital letter, where i can find the unicode rage to exclude them from the finding? I don't want to use something like this

(?<![\x{0370}-\x{03FF}\x{1F00}-\x{1FFF}])([\x{0370}-\x{03FF}\x{1F00}-\x{1FFF}]+)-(?!\ |\,|\1|\Δ|\Π|\Α|\Ρ|\Χ|\Ν|\Ζ|\Φ)

Thanks