Q: Regex Find and Replace delete surrounding tags

hidden.platypus · 05-30-2015, 12:44 PM

Hi all,

I have no clue what I'm doing, and now I'm stuck.

I apparently understand the find part of the Find and Replace function. It finds what it I want it to find, but the Replace section won't cooperate. So there's something I don't understand. Any help would be appreciated.

I converted a docx to epub using Calibre.

I am now using the Edit Book app in Calibre to do final polishing.

Calibre created a ton of tags. I'm learning to use css to format ebooks (and learning css). What I'm attempting to do is delete the opening span tag and the closing span tag and leave what's inside the tag alone. I am using css to do what the tag is already doing.

So I want to remove the span tags surrounding a string and leave the string untouched.

In the code I have this:

Code:

<span class="calibre5">The law is inevitable.</span>

I want to do a search and replace that leaves only this

Code:

The law is inevitable.

My Search string is:

Code:

<span class="calibre5">[^<>]+</span>

But anything I put in the Replace box gets literally replaced.

So if I put:

Code:

(?<=<span class="calibre5">)(.*?)(?=<\/span>)

Then instead of getting

Code:

The law is inevitable.

when I push Replace

I get:

Code:

(?<=<span class="calibre5">)(.*?)(?=<\/span>)

And if i put

Code:

/1

I get

Code:

/1

So you get the picture.

How can I learn what I'm doing wrong?

DiapDealer · 05-30-2015, 01:24 PM

Your slash is backwards in your replace expression.

Should be \1 and not /1

I liked your first expression best:

Code:

<span class="calibre5">[^<>]+</span>

You just forgot to capture the span contents...

Code:

<span class="calibre5">([^<>]+)</span>

... so you could stick them back in without the span tags using just \1 for the replace.

**Note that your expression will ignore any nested span situations (or any situations with or or the like included in the span). In other words it won't match anything in:

Code:

<span class="calibre5">The law is <span>oops</span> inevitable.

Nested tags can cause problems when trying to use regex alone to alter them if you're not careful.

My suggestion (and I know this going to sound like rampant self-promotion, but this is exactly why I created it the first place)?? I'd use my editor plugin that allows you to remove an opening span and and it's matching closing tag without having to worry about nesting and the like.

hidden.platypus · 05-30-2015, 01:47 PM

Thanks for responding.

I was just about to do a

Eureka! I got it kind of post but I really appreciate you responding.

Taught me some things I wasn't seeing properly for some reason.

Quote:

Originally Posted by DiapDealer

Your slash is backwards in your replace expression.

Should be \1 and not /1

So in my frustration I wouldn't have noticed that. Thank you for that. I'll burn it into my memory which slash is proper.

Quote:

Originally Posted by DiapDealer

You just forgot to capture the span contents...

Code:

<span class="calibre5">([^<>]+)</span>

I don't understand what that means. Where do I learn that?

Here's what I ended up doing:

Find:

Code:

<span class="calibre5">(?<mygroup>([^<>]+))</span>

Replace:

Code:

\g<mygroup>

Now I have to sit and think about:

Quote:

Originally Posted by DiapDealer

**Note that your expression will ignore any nested span situations (or any situations with or or the like included in the span). In other words it won't match anything in:

Code:

<span class="calibre5">The law is <span>oops</span> inevitable.

Why would you do that to someone?

Why?

JK

Thank you much.

What's your tool btw? I'm going to click your profile and find out but I felt it would probably be polite to ask you first.

hidden.platypus · 05-30-2015, 01:49 PM

Didn't notice that you'd linked it.

Nevermind. Frustration and euphoria have combined to make me drunk.
Such it always is.

DiapDealer · 05-30-2015, 04:01 PM

Quote:

Find:

Code:

<span class="calibre5">(?<mygroup>([^<>]+))</span>

Replace:

Code:

\g<mygroup>

Not really worth going to the trouble of creating a named group in this case.

Anything (OK not anything--but for the purpose of this exercise) in parentheses is a capture group.

So in my expression:

Code:

<span class="calibre5">([^<>]+)</span>

Everything matched within the parentheses (a capture group) can be substituted in the Find expression using \1. If you had more than one capture group, they would be \2 \3 ... etc.

Your way (named groups) will work, but there's no need to go to all that trouble, IMO.

I learned nearly everything I know about regex from http://www.regular-expressions.info/

gbm · 05-31-2015, 12:14 AM

Quote:

Originally Posted by hidden.platypus

Hi all,

I have no clue what I'm doing, and now I'm stuck.

I apparently understand the find part of the Find and Replace function. It finds what it I want it to find, but the Replace section won't cooperate. So there's something I don't understand. Any help would be appreciated.

I converted a docx to epub using Calibre.

I am now using the Edit Book app in Calibre to do final polishing.

Calibre created a ton of tags. I'm learning to use css to format ebooks (and learning css). What I'm attempting to do is delete the opening span tag and the closing span tag and leave what's inside the tag alone. I am using css to do what the tag is already doing.

So I want to remove the span tags surrounding a string and leave the string untouched.

Spoiler:

I just do a search and replace on the opening span tag, then under tools use Fix Html- all files, to remove the closing span tag.

bernie

Thom* · 06-03-2015, 10:52 AM

You have probably already resolved this, but I simply use:

Search: (.*?)
Replace: \1

Make sure you have selected "regex" for searsh method and that is it.

Now, nested tags could cause a problem, but that is for another discussion.

Best of Luck

DiapDealer · 06-03-2015, 11:22 AM

Quote:

Originally Posted by Thom*

You have probably already resolved this, but I simply use:

Search: (.*?)
Replace: \1

Make sure you have selected "regex" for searsh method and that is it.

Now, nested tags could cause a problem, but that is for another discussion.

Best of Luck

Code:

<span class="calibre5">([^<>]+)</span>

Is infinitely less potentially destructive than

Code:

<span class="calibre5">(.*?)</span>

precisely because of nested spans.

Thom* · 06-03-2015, 12:12 PM

Interesting, I am new to this and was not familiar with ^. The concern I have is that your search would not find:

Now is the time for all good men to enjoy a coke. or:
Now is the time for all good men to enjoy a coke.

This search would:

(((?!<span).)*?)

Code:

Unrelated, how do you get those "code boxes" to show up in your comment?

Ah, I figured it out.

In case I sound stupid, remember that I am just learning.

phossler · 06-03-2015, 02:56 PM

I really like DiapDealer's plug in

https://www.mobileread.com/forums/sho...d.php?t=251365

You might look into it since I think it might handle even complex cases like that

DiapDealer · 06-03-2015, 05:19 PM

Quote:

Originally Posted by Thom*

Interesting, I am new to this and was not familiar with ^. The concern I have is that your search would not find:

Now is the time for all good men to enjoy a coke. or:
Now is the time for all good men to enjoy a coke.

Correct.
But it will also not match:

Code:

<span class="calibre5">Now is the time for all <span class="italic">good men</span> to enjoy a coke.</span>

Whereas the first (.*?) might match:

Now is the time for all good men to enjoy a coke.

and create malformed html when replaced.

I'd rather have to make a second pass for the calibre5 spans that didn't get picked up with the first search than have the first search botch my html.

I usually try to avoid using regex for matching ending tags with opening tags entirely (especially in potentially nested situations).

Thom* · 06-03-2015, 08:44 PM

So you are suggesting multiple passes to clear tags. I like the idea.

In your example:

Quote:

Now is the time for all good men to enjoy a coke.

I could first run

Code:

<span class="italic">(((?!<span).)*?)</span>

to get the inner tags, then run

Code:

<span class="calibre5">(((?!<span).)*?)</span>

to clear the outer tags.

Great idea, I will give it a try.
I think it might take more that two passes in some cases.

DiapDealer · 06-03-2015, 08:52 PM

Quote:

Originally Posted by Thom*

Great idea, I will give it a try.
I think it might take more that two passes in some cases.

Do it for the practice/experience ... then try my editor plugin (which would get all the calibre5 spans in one go--regardless of any nesting).

eschwartz · 06-16-2015, 09:59 PM

@Thom*-- See: s&r for paired tags

I recommended something like:

Code:

<span(?: class="calibre5")?>((?:(?!<span).)*?)</span>

to match all content inside those span tags, unless there is another span tag in the content.

But really, the OP should avoid the problem to begin with, by opening his DOCX in the editor directly. It will be converted into an EPUB for editing -- without having the CSS flattened!
Or use Toxaris' Word addin.

calibre's conversion is meant for finished products, to move to another device. And usually by end-users, not content creators. That is why the editor was designed to avoid the creation of tag soup.

Thom* · 06-16-2015, 11:16 PM

I thought I was smart but you had this figured out a year ago.

Foiled again.

05-30-2015, 12:44 PM	#1
hidden.platypus Connoisseur Posts: 89 Karma: 190508 Join Date: May 2014 Device: Android	Q: Regex Find and Replace delete surrounding tags Hi all, I have no clue what I'm doing, and now I'm stuck. I apparently understand the find part of the Find and Replace function. It finds what it I want it to find, but the Replace section won't cooperate. So there's something I don't understand. Any help would be appreciated. I converted a docx to epub using Calibre. I am now using the Edit Book app in Calibre to do final polishing. Calibre created a ton of <span class="calibre5"> tags. I'm learning to use css to format ebooks (and learning css). What I'm attempting to do is delete the opening span tag and the closing span tag and leave what's inside the tag alone. I am using css to do what the <span class="calibre5"> tag is already doing. So I want to remove the span tags surrounding a string and leave the string untouched. In the code I have this: Code: <span class="calibre5">The law is inevitable.</span> I want to do a search and replace that leaves only this Code: The law is inevitable. My Search string is: Code: <span class="calibre5">[^<>]+</span> But anything I put in the Replace box gets literally replaced. So if I put: Code: (?<=<span class="calibre5">)(.?)(?=<\/span>) Then instead of getting Code: The law is inevitable. when I push Replace I get: Code: (?<=<span class="calibre5">)(.?)(?=<\/span>) And if i put Code: /1 I get Code: /1 So you get the picture. How can I learn what I'm doing wrong?

05-30-2015, 01:24 PM	#2
DiapDealer Grand Sorcerer Posts: 27,595 Karma: 193191846 Join Date: Jan 2010 Device: Nexus 7, Kindle Fire HD	Your slash is backwards in your replace expression. Should be \1 and not /1 I liked your first expression best: Code: <span class="calibre5">[^<>]+</span> You just forgot to capture the span contents... Code: <span class="calibre5">([^<>]+)</span> ... so you could stick them back in without the span tags using just \1 for the replace. *Note that your expression will ignore any nested span situations (or any situations with <i> or <b> or the like included in the span). In other words it won't match anything in: Code: <span class="calibre5">The law is <span>oops</span> inevitable. Nested tags can cause problems when trying to use regex alone to alter them if you're not careful. My suggestion (and I know this going to sound like rampant self-promotion, but this is exactly* why I created it the first place)?? I'd use my editor plugin that allows you to remove an opening span and and it's matching closing tag without having to worry about nesting and the like. Last edited by DiapDealer; 05-30-2015 at 01:36 PM.

06-03-2015, 10:52 AM	#7
Thom* The Fumbler Posts: 66 Karma: 10 Join Date: Jun 2015 Device: android 4.2/fbreader	You have probably already resolved this, but I simply use: Search: <span class="calibre5">(.*?)</span> Replace: \1 Make sure you have selected "regex" for searsh method and that is it. Now, nested <span> tags could cause a problem, but that is for another discussion. Best of Luck

06-03-2015, 12:12 PM	#9
Thom* The Fumbler Posts: 66 Karma: 10 Join Date: Jun 2015 Device: android 4.2/fbreader	Fumbling Interesting, I am new to this and was not familiar with ^. The concern I have is that your search would not find: <span class="calibre5">Now is the time for all <i>good men</i> to enjoy a coke.</span> or: <span class="calibre5">Now is the time for all <em>good men</em> to enjoy a coke.</span> This search would: <span class="calibre5">(((?!<span).)*?)</span> Code: Unrelated, how do you get those "code boxes" to show up in your comment? Ah, I figured it out. In case I sound stupid, remember that I am just learning.

06-03-2015, 02:56 PM	#10
phossler Wizard Posts: 1,079 Karma: 412718 Join Date: Jan 2009 Location: Valley Forge, PA, USA Device: Kindle Paperwhite	I really like DiapDealer's plug in https://www.mobileread.com/forums/sho...d.php?t=251365 You might look into it since I think it might handle even complex cases like that Attached Thumbnails

05-30-2015, 01:49 PM	#4
hidden.platypus Connoisseur Posts: 89 Karma: 190508 Join Date: May 2014 Device: Android	Didn't notice that you'd linked it. Nevermind. Frustration and euphoria have combined to make me drunk. Such it always is.

06-16-2015, 09:59 PM	#14
eschwartz Ex-Helpdesk Junkie Posts: 19,421 Karma: 85397180 Join Date: Nov 2012 Location: The Beaten Path, USA, Roundworld, This Side of Infinity Device: Kindle Touch fw5.3.7 (Wifi only)	@Thom-- See: s&r for paired tags I recommended something like: Code: <span(?: class="calibre5")?>((?:(?!<span).)?)</span> to match all content inside those span tags, unless there is another span tag in the content. But really, the OP should avoid the problem to begin with, by opening his DOCX in the editor directly. It will be converted into an EPUB for editing -- without having the CSS flattened! Or use Toxaris' Word addin. calibre's conversion is meant for finished products, to move to another device. And usually by end-users, not content creators. That is why the editor was designed to avoid the creation of tag soup. Last edited by eschwartz; 06-16-2015 at 10:01 PM.

06-16-2015, 11:16 PM	#15
Thom* The Fumbler Posts: 66 Karma: 10 Join Date: Jun 2015 Device: android 4.2/fbreader	I thought I was smart but you had this figured out a year ago. Foiled again.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
RegEx to replace only inside certain tags?	phossler	Editor	6	03-03-2015 08:24 PM
Regex find and replace	SanatyrZeo	Sigil	5	10-29-2012 07:03 AM
Regex problem: Trying to replace surrounding text without effecting the middle	ghostyjack	Workshop	3	10-09-2012 04:26 PM
RegEx find and replace	iblesq	Sigil	1	01-10-2011 09:26 PM
REGEX find and replace help please	potestus	Sigil	13	09-18-2010 04:14 PM

Advert

Advert