Remove Italic Tags from Selection?

[enuddleyarbl]

Is there some way to remove italic tags (<i> and </i>, but this could apply to other tags as well) from a selected group of paragraphs in the Editor?

I ask because almost all the time, the author/publisher italicizes things with those tags in the HTML instead of with a css style. So far, the only things I've come up with are manually selecting each tag pair and deleting them, or marking the selection and running a Regex search/replace on the marked text:

find: (<p[^>]*>)<i>(.+?)</i>([\.\?]+)?</p>
replace: \1\2\3</p>

[theducks]

Diaps Toolbag (plugin)
Outright delete or Modify (assign different)

[Karellen]

Originally Posted by enuddleyarbl

I ask because almost all the time, the author/publisher italicizes things with those tags in the HTML instead of with a css style.

What is wrong with that?

[enuddleyarbl]

@theducks: Unless I'm missing something with Diap's Toolbag, it can only change all the stuff on a given page (EDIT: sorry, the current HTML file) or all the stuff in the whole book (all the files). I just need to remove those tags from specific clumps of text on a page.

@Karellen: Usually, I've no problem with things like <i>, <em>, <b> and <strong> in the HTML. But, where there are clumps of text that should be treated as a syntactic whole (i.e., a quote at the top of a chapter, a poem, a letter, etc.), I'd like to apply a CSS class to handle the formatting. It's possible the styling embedded in the content will conflict with that present in the class.

If the <i> tags were only around small bits of text like ship names or foreign words, that's probably not a problem. I think I could use another class to override their italics to something like bold:

i i {
  font-style: normal;
  font-weight: bold;
}

But, if I included something like that where large swaths of text are italicized, all I'd end up with is large swaths of text bolded.

[DNSB]

That really sounds like something where it would be easier just to make the changes manually. You could search for the <i>blah blah blah</i> text and either replace or find next. Unless there is something very specific that you can use to select the chunks of text you want to remove the italics from, there are no good ways to automatically make the changes.

[enuddleyarbl]

Yep. That's what I was afraid of. As it is, unless the number of paragraphs in a the block of text is big enough (say, for instance, 4 lines of a poem or more), I just manually delete the tags. If there are more, I'll go through the select, mark, type regex, replace, unmark dance.

[JSWolf]

Originally Posted by enuddleyarbl

Yep. That's what I was afraid of. As it is, unless the number of paragraphs in a the block of text is big enough (say, for instance, 4 lines of a poem or more), I just manually delete the tags. If there are more, I'll go through the select, mark, type regex, replace, unmark dance.

There's also no need to make these changes. So the simplest solution is to let them be.

[DNSB]

Jon, in your opinion, there is no reason to remove those italics. Sadly, it would appear that enuddleyarbl has their own opinion which disagrees with yours. His ebook, his formatting.

[JSWolf]

Originally Posted by DNSB

Jon, in your opinion, there is no reason to remove those italics. Sadly, it would appear that enuddleyarbl has their own opinion which disagrees with yours. His ebook, his formatting.

There is no need. That's a fact. There's a want to do it. That's also a fact.

I never said there's no reason. I said there's no need. There isn't a need. There's a want. A need would be if it wasn't working or causing some sort of display problem.

Some of the things I do when editing an eBook are not needed to be done but are done because I want to do them.

[enuddleyarbl]

I guess it's not a needful thing, but in case some other person wants to do something similar, I'm starting to look at this from a different direction. Instead of working with the <i>/</i> pair and the text in between, I've decided to look for the whitespace between adjacent italicized paragraphs:

find: </i>([\.\?]?</p>\s+?<p[^>]*>)<i>
replace: \1

I'm assuming if there are consecutive entire paragraphs italicized, there's probably some kind of semantic block involved. Once I find the start of such a block, I can remove the intervening </i> and <i> pairs with "Replace and Find" and leave only the single starting and ending pair. I can probably work with that more easily.

[Tex2002ans]

Originally Posted by enuddleyarbl

If the <i> tags were only around small bits of text like ship names or foreign words, that's probably not a problem. I think I could use another class to override their italics to something like bold:

i i {
  font-style: normal;
  font-weight: bold;
}

One trick that I do when proofing something like this...

I add a temporary background-color:

i i {
	background-color: red; <---- Add this line.
}

(This would highlight all <i> nested within other <i>s!)

As I skim through the preview of the ebook, the different colors stand out like a sore thumb. This lets me take a much closer look at the (problematic) areas.

When I'm done fixing up all the code, I remove the background-color.

- - -

Side Note: I sometimes use multiple colors for different things:

• yellow
• blue
• orange
• [...]

Just choose any of the HTML Color Names if you want it easy.

You could then use something like:

background-color: green;

to let you know you already fixed it!

- - -

Originally Posted by enuddleyarbl

[...] where there are clumps of text that should be treated as a syntactic whole (i.e., a quote at the top of a chapter, a poem, a letter, etc.), I'd like to apply a CSS class to handle the formatting. It's possible the styling embedded in the content will conflict with that present in the class.

Pretty good idea.

In that case, I mark with a <div> or <blockquote> plus a class="".

For example:

<blockquote class="openingquote">
	<p>Some amazing quote.</p>
	<p class="right">—Tex</p>
</blockquote>

<div class="poem">
	<div class="stanza">
		<p class="line">The cow jumped over the moon.</p>
		<p class="line">And Tex jumped over the <em>house</em>!</p>
	</div>
</div>

Then you can easily toggle the italics on/off if needed:

Spoiler:

blockquote.openingquote {
	font-style: italic; <--- Optional
	margin-top: 1em;
	margin-bottom: 1em;
	margin-left: 10%;
	margin-right: 10%;
}

div.poem {
	font-style: italic; <--- Optional
	margin-top: 1em;
	margin-bottom: 1em;
	margin-left: 3em;
}

What I tend to do is something along these lines:

Search: <i>(.+?)</i>
Replace: <span class="temppoem">\1</span>

and just go through the book on a case-by-case basis.

Combined with that background-color trick above, it makes it much easier to find these spots + fix them up.

- - -

After I'm done, I can then do a:

Search: <span class="temppoem">

and go mass replacing those new <span>s with whatever code I need.

(Usually use Diap's amazing Toolbag to shift to something clean!)

- - -

Originally Posted by enuddleyarbl

But, if I included something like that where large swaths of text are italicized, all I'd end up with is large swaths of text bolded.

What does the original book look like?

Italics/Emphasis within italics = Roman (Straight up-and-down, normal text).

So let's say you were reading a Fiction book where all the character's inner thoughts are in italics:

<p><i class="thoughts">The ghost is going to <em>kill</em> my cat!</i>
Tex doused the door handle in holy water.
<i class="thoughts">Remind me to never read the <i class="booktitle">Necronomicon</i> or go aboard the <i class="shipname">HMS Haunty</i> ever again!</i></p>

These 4 words would all be Roman (= normal text):

• kill
• Necronomicon
• HMS Haunty

Similar to quotes within quotes, where you alternate:

• “Outer ‘Inner “Double-Inner” Inner’ Outer” (American)
• ‘Outer “Inner ‘Double-Inner’ Inner” Outer’ (British)

You do the same thing with italics!

Think of it like an ON/OFF switch:

• 1st layer = ON
• 2nd layer = OFF
• 3rd layer = ON

Then if you made all your thoughts be bold instead, you'd toggle the italics back on for those "inner" emphasis + book titles + ship names.

Originally Posted by enuddleyarbl

Once I find the start of such a block, I can remove the intervening </i> and <i> pairs with "Replace and Find" and leave only the single starting and ending pair. I can probably work with that more easily.

One more trick I sometimes use:

Search: <i>([^<]{100,})</i>
Replace: <span class="replace">\1</span>

Here's a breakdown of that Regex:

• <i> = "Look for an italics."
• (...) = "Stick all this stuff into Group 1."
  
  • [^<] = "Look for any character that's NOT a less than sign."
    
  • {100,} = "Look for 100 OR MORE of the previous thing."
    
    • Adjust # as needed.
• </i> = "Look for close of italics."

This will find really long italics, while ignoring any italics <100 characters long.

(Sometimes I start with high numbers, like 200, then progressively work my way down.)

Originally Posted by enuddleyarbl

I'm assuming if there are consecutive entire paragraphs italicized, there's probably some kind of semantic block involved.

Yeah... every single ebook is going to be completely different.

We can give general principles/guidelines, but there definitely isn't a "one-button press" solution for complicated situations like this.

Like you might have another book with hundreds of different "calibre123" classes that all apply italics, and you'd have to figure out what each one does.

99.99% chance the ebooks aren't super clean with their <i> + <em> markup... especially when dealing with nested cases like this!

- - -

Complete Side Note: If you want the ultimate info on Italics <i> and Emphasis <em>, see my posts in:

• 2021: "Italics and Bold"
• 2020: "<i>, <em> or <span> for italics ?"
  
  • My stuff begins at Post #67.

If you want even more examples of weird edge-cases like "quotes-within-quotes" or "should the punctuation be in italics?" see:

• 2020: "Punctuation rules of thumb?"
  
  • My stuff begins at Post #16.

[enuddleyarbl]

Tex2002ans: Thanks. I especially like the idea of temporarily changing colors of the types of things you're looking for. That should make seeing them much easier.

[enuddleyarbl]

And, of course, it helps a lot now that I've found the Calibre Editor's Saved Search function: Search > Saved Searches (is my face red for not having found that up until now).

[Tex2002ans]

Originally Posted by enuddleyarbl

And, of course, it helps a lot now that I've found the Calibre Editor's Saved Search function: Search > Saved Searches (is my face red for not having found that up until now).

Yes, Saved Searches are fantastic.

Here's an example where I explained them in detail:

• 2022: "Convert text formating from CSS to HTML"

(I use them in Sigil, not so much in Calibre.)

For example, I have a 12-step Saved Search I use to instantly clean up Finereader HTML/EPUB cruft.

And here's an example EPUB that I used that method on:

• 2022: "Optimize PDFs from archive.org for E-Ink devices"

Within seconds, I was able to generate a (relatively clean) EPUB from the PDF. Compared to the Archive.org "EPUB" version... mine is wayyyyyy better.

[capink]

You can use the Editor Chains plugin to do this. But it is going to be two part process

1. Mark the tag(s) you want to remove. (Using the first chain attached below).
   This chain will mark tags by adding an attribute to them. First bind the chain to whatever keyboard shortcut you want. After that you can start marking the tags you want to act on. All you need is to have the cursor somewhere inside the tag before clicking the keyboard shortcut.
2. Run the second chain to make the necessary changes to the tags you previously marked.

Also you can merge adjacent italic tags into one. You can find an example for that in the plugin's thread. It will need modifying to change the name form <span> to <i>.

Notes:

• To import the chains. go to Editor Chains > Add/modify chains > right click on table > import chain.
• To bind a chain to a keyboard shortcut. go to Editor Chains > Add/modify chains > Keyboard shortcuts.
• The first action of the second chain will add a class name "italic" to the parent tag. If you don't want that, edit the chain to remove it.
• In your case we are already unwrapping the <i> tag, so the marks will go away as will. In other scenarios, we need to add an action to the second chain to remove the marking attribute; I added it anyway as it does not hurt.
• The "Mark Tag" of the plugin is configurable. So you can make as many mark chains as you want to use for different purposes