Regex - replace only part of a string - how?

flameproof · 02-20-2012, 08:34 AM

I have often problems with books that were auto-converted by Calibre. Here is one issue:

Text has often wrong line breaks.

Example:

Code:

  <p class="calibre2">This is just a sample</p>

  <p class="calibre2">text with no meaning.</p>

I can find it with the string:

Code:

 [a-z]</p>

  <p class="calibre2">

But when I replace it then (of course) the last letter is missing. Without the [a-z] I would catch normal end of sentence line breaks.

Is there a way?

DOH! I found it!

Search string: (\w+)</p>

<p class="calibre2">

Replace with: \1

flameproof · 02-20-2012, 09:21 AM

Please let me add one more common problem: wrong hyphen (probably from the PDF)

Search: (\w)-(\w)
replace with: \1\2

How can I make it case sensitive? I like to correct 'ele-phant' but not 'John-Bob' ?

PeterT · 02-20-2012, 09:29 AM

Quote:

Originally Posted by flameproof

Please let me add one more common problem: wrong hyphen (probably from the PDF)

Search: (\w)-(\w)
replace with: \1\2

How can I make it case sensitive? I like to correct 'ele-phant' but not 'John-Bob' ?

I think a brute force approach would be
([a-z]\w)-(\w)

flameproof · 02-20-2012, 09:37 AM

Quote:

Originally Posted by PeterT

I think a brute force approach would be
([a-z]\w)-(\w)

Thanks.

Seems '([a-z])-([a-z])' with a clicked 'Match Cases' is OK too.

Serpentine · 02-21-2012, 07:24 PM

Use the posix character classes rather, for your example : ([[:lower:]])-([[:lower:]])

List of posix char classes

WS64 · 02-22-2012, 11:11 AM

[[:lower:]]?

I guess you mean [:lower:]
But honestly, I find [a-z] way easier to write. Especially since I have to add some German letters often, like [a-zäöüß]

DiapDealer · 02-22-2012, 11:37 AM

Quote:

I guess you mean [:lower:]

Nope. [[:lower:]] is the correct usage.

Timur · 02-22-2012, 11:50 AM

I was used to write [a-z] for lowercase letters too, but since I discovered that unicode properties flag is working in Sigil 0.5 (*UCP), I simply use character classes with it to cover non-ASCII letters.
\w, \W, [:lower:], [:upper:], [:alpha:], [:alnum:] are all affected by (*UCP).

WS64 · 02-22-2012, 01:36 PM

Quote:

Originally Posted by DiapDealer

Nope. [[:lower:]] is the correct usage.

I really had to check these...
You are (of course) right, I was wrong.
I never tried those since I never saw a reason to use them...

Serpentine · 02-22-2012, 07:50 PM

Quote:

Originally Posted by WS64

I never tried those since I never saw a reason to use them...

They're great, since they cover your unicode characters too, for example you dont have to add ä,ß, etc - they are already understood to be lowercase.

[a-zäöüß] will all be captured by [[:lower:]], as well as a load more of edge cases which you might not have thought of. Saving you time, making edits more complete. The punct class is also especially useful - and very often overlooked.

The more you know!

WS64 · 02-23-2012, 02:51 AM

Quote:

Originally Posted by Serpentine

They're great, since they cover your unicode characters too, for example you dont have to add ä,ß, etc - they are already understood to be lowercase.

I just checked. [[:lower:]] does NOT find äöüß.

Timur · 02-23-2012, 04:43 AM

@WS64: add (*UCP) in front of your pattern like:

Code:

(*UCP)\b[[:lower:]]+

02-20-2012, 08:34 AM	#1
flameproof Member Posts: 17 Karma: 10 Join Date: Dec 2011 Device: Sony PRS-T1	Regex - replace only part of a string - how? I have often problems with books that were auto-converted by Calibre. Here is one issue: Text has often wrong line breaks. Example: Code: <p class="calibre2">This is just a sample</p> <p class="calibre2">text with no meaning.</p> I can find it with the string: Code: [a-z]</p> <p class="calibre2"> But when I replace it then (of course) the last letter is missing. Without the [a-z] I would catch normal end of sentence line breaks. Is there a way? DOH! I found it! Search string: (\w+)</p> <p class="calibre2"> Replace with: \1 Last edited by flameproof; 02-20-2012 at 08:45 AM.

02-23-2012, 04:43 AM	#12
Timur Connoisseur Posts: 54 Karma: 37363 Join Date: Aug 2011 Location: Istanbul Device: EBW1150, Nook STR	@WS64: add *(UCP)** in front of your pattern like: Code: (UCP)\b[[:lower:]]+ Last edited by Timur; 02-23-2012 at 04:43 AM. Reason: typo*

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
regex search/replace	Sharlene	Sigil	10	01-28-2012 04:14 AM
regex replace???	schuster	Conversion	14	01-29-2011 09:02 AM
need regex help search and replace	schuster	Calibre	4	01-10-2011 09:00 AM
My RegEx isn't doing what I hoped to remove page numbers and a fixed string	winterminute	Calibre	6	12-19-2010 10:55 PM
Find and replace string with wildcard	jhempel24	Sigil	15	11-12-2010 01:50 PM

02-20-2012, 09:21 AM	#2
flameproof Member Posts: 17 Karma: 10 Join Date: Dec 2011 Device: Sony PRS-T1	Please let me add one more common problem: wrong hyphen (probably from the PDF) Search: (\w)-(\w) replace with: \1\2 How can I make it case sensitive? I like to correct 'ele-phant' but not 'John-Bob' ?

02-21-2012, 07:24 PM	#5
Serpentine Evangelist Posts: 416 Karma: 1045911 Join Date: Sep 2011 Location: Cape Town, South Africa Device: Kindle 3	Use the posix character classes rather, for your example : ([[:lower:]])-([[:lower:]]) List of posix char classes

02-22-2012, 11:11 AM	#6
WS64 ♫ Posts: 660 Karma: 506380 Join Date: Aug 2010 Location: Germany Device: Kobo Aura / PB Lux 2 / Bookeen Frontlight / Kobo Mini / Nook Color	[[:lower:]]? I guess you mean [:lower:] But honestly, I find [a-z] way easier to write. Especially since I have to add some German letters often, like [a-zäöüß]

02-22-2012, 11:50 AM	#8
Timur Connoisseur Posts: 54 Karma: 37363 Join Date: Aug 2011 Location: Istanbul Device: EBW1150, Nook STR	I was used to write [a-z] for lowercase letters too, but since I discovered that unicode properties flag is working in Sigil 0.5 *(UCP)*, I simply use character classes with it to cover non-ASCII letters. \w, \W, [:lower:], [:upper:], [:alpha:], [:alnum:] are all affected by (UCP).

Advert

Advert