Replace <br/> in <h1>...<h1/>

[foosion]

I'd like to replace <br/>s with a space when they're between opening and closing h1 tags. However, something simple such as

<h1>(.*?)<br/>(.*?)<h1/>
<h1>\1 \2<h1/>

doesn't work well since there are some opening and closing h1s that don't have a br between them, so this matches everything until the next br and then closing h1, which could be a lot. I just want to replace a br when it's between an opening and closing h1. In other words find an opening and closing h1 tag, then if there's a br in that range replace it with a space, otherwise do nothing.

How to?

[enuddleyarbl]

Odd. In my Calibre editor that works just fine.

A couple of thoughts, though (could just be typos). First, you should close your <br> tags. IOW, <br/>, not just <br>. Second, are you sure all your <h1></h1> tags will be on the same line? Because, as written, your search string won't find multiple line <h1>s. And, third (possibly related to the problem you're seeing), do you have the "Dot All" box checked at the bottom of the editor screen? With it unchecked, it works as you want for single line <h1>s. With it checked, it'll pick up multiple line <h1>s, but also those <h1>s without a <br/> and stretch the selection to the next </h1>.

[foosion]

<br> was a typo - I meant <br/>

The code I posted finds all of my <h1>s, the problem is it also finds large areas of text not within <h1>...<h1/> if there isn't a <br/> inside. I do have Dot All checked, otherwise it doesn't find the <h1>...<h1/>s; none are single line.

So the question is how to just find <br/> within <h1>...<h1/> with Dot All checked? IOW, when there are <h1>s without a <br/>, how to avoid stretching the selection to the next </h1>?

[theducks]

<h1/> is also a typo H1 is not a self closing tag.

<h1>(.*?)\s*<br />\s*(.*?)</h1>

I included 0 or more spaces as part of my search as you do put 1 back in your replace.

FWIW I usually do the reverse. I insert a trailing space after \1.

\1 <br />\2

for use on my small (not a desktop) screen when chapter titles are more the just Chapter #. The advantage is, the TOC builder works.
You can even resize the second line with a <span class=> and not mess up the TOC inclusion

[foosion]

Originally Posted by theducks

<h1>(.*?)\s*<br />\s*(.*?)</h1>

That seems to have the same issue if there's a <h1> </h1> pair without a <br/> inside - it stretches until the next h1.

[theducks]

did you have Dot all checked? That makes it Greedier than you want.

[Quoth]

Sometimes a heading tag is for a multi-line heading where a a separate heading tag for each line would result in erroneous auto-generated TOC.

Also a <br /> is often for a heading that is logically two lines or two long for a single line would look strange with an arbitrary position for the word wrap with big font or smaller screen.


So while it's perfectly possible to find only the newlines embedded in a heading and put a space instead it may not be what is really needed.

Though, in an ebook, I'd only ever have <br /> in a heading, if anywhere, as extra space elsewhere is better done with paragraph CSS.

[theducks]

I also do it for an attempt at (anti-ugly) control:

Chapter One Hundred Twenty-Seven

I would rathe have, than a arbitrary break because it will not fit on the line.

Chapter
One Hundred Twenty-Seven

[enuddleyarbl]

I, too, am always (we'll, often) adding <br/>s to my chapter headings just to make sure they fit decently on the page. And, sorry I forgot to mention the ending </h1> tag. I saw it, but forgot.

[foosion]

Originally Posted by theducks

did you have Dot all checked? That makes it Greedier than you want.

If Dot all is unchecked I don't get any hits. With it checked it's greedier than I want.

I'm starting to think I need a function that first finds <h1>(.*?)</h1>, then searches \1 for <br/> and does the replace. I'm just not sure how to do this.

[foosion]

Originally Posted by Quoth

Though, in an ebook, I'd only ever have <br /> in a heading, if anywhere, as extra space elsewhere is better done with paragraph CSS.

Agreed, but this is a commercial book and it's using <br/> as a line break in many places.

[theducks]

I do this kind of search all the time. (eg add a <span around \2 to style it a bit different)
There is something hidden in the text that is throwing the find off.
One thing I found can help: Beautify all HTML. That does get rid of space spaces (or use \s* )

[foosion]

This seems to work as a regex function:

def replace(match, number, file_name, metadata, dictionaries, data, functions, *args, **kwargs):
    return '<h1>'+match.group(1).replace('<br/>',' ').replace('<br/>',' ').replace('<br/>',' ')+'</h1>'

I did multiple replaces because there are up to three <br/>s in headings.

I used create TOC from headings and it's added ids throughout, including blanks ones, e.g.:

<h1 id="toc_21">Security measures adopted by Atlantis/Shanghai.</h1><h1 id="toc_22"></h1>

Is this normal?

[Quoth]

Originally Posted by foosion

Agreed, but this is a commercial book and it's using <br/> as a line break in many places.

Some of or all those need to be replaced
/p> <br /> <p class="existing"
with
/p> <p class="revised"