regex help

najgori · 01-21-2014, 07:42 AM

1.
p(?!([^<]+)?>) "find character 'p' everywhere except within html tags"

2.
p(?!([^&]+)?;) "find character 'p' everywhere except within named entities"

i would like to merge these 2 searches to 1, so my new search would skip both tags and named entities.

i could replace all named entities with decimal, but this is indirect solution.

mzmm · 01-21-2014, 10:08 AM

--edit, misread the question, can't figure out how to delete this.

najgori · 01-21-2014, 02:06 PM

it looks like a common task: "mark every letter 'p' but skip tags and named entities".

you can "cheat" by converting named entities to numbers which is recommended for ebooks (if i am correct?), but i am looking for "elegant" solution for this problem.

eschwartz · 01-21-2014, 02:57 PM

use the following format to check for "either _ or _": (either|or)

najgori · 01-21-2014, 04:01 PM

p((?!([^<]+)?>)|(?!([^&]+)?;)) does not work.

eschwartz · 01-21-2014, 08:04 PM

instead of saying "(either not this | or not that)", you need to say "not (either this | or that)". Currently, if either condition is true, a match is made. A "p" in an html tag isn't in an entity, and vice versa, so of course it matches.

...

Here you go:

Code:

p(?!([^<&]+)?(>|;))

First we search for the "p" then we check ahead to see if the following is not there: any character string that does NOT contain either a "<" or "&" , followed by either a ">" or ";"

EDIT: If your sentence ends in a ";" this won't match, (you can try adding a negative lookbehind for a "&" and seeing where that leads you) so you're probably better off using calibre to convert all entities to unicode. And tags won't have that problem.

EDIT #2: added regex pipe and parentheses to beginning of answer, to clarify how it works.

mrmikel · 01-21-2014, 08:42 PM

Quote:

Originally Posted by najgori

you can "cheat" by converting named entities to numbers which is recommended for ebooks (if i am correct?), but i am looking for "elegant" solution for this problem.

Unless this is for a school project in epubs, the reader will not care or even know what kind of entities are there, so where does elegance enter the picture?

najgori · 01-22-2014, 02:49 AM

@mrmikel. elegant meaning single regular expression instead of many - entiites are not important

@eschwartz. also does not work. it misses some characters.

i am in a process of learning regular expressions and this was just a question "could this be done?" - not very important to me. big thanx to everybody who tried to find solutions.

eschwartz · 01-22-2014, 11:06 AM

Quote:

Originally Posted by najgori

@mrmikel. elegant meaning single regular expression instead of many - entiites are not important

@eschwartz. also does not work. it misses some characters.

i am in a process of learning regular expressions and this was just a question "could this be done?" - not very important to me. big thanx to everybody who tried to find solutions.

Like I said. What you really want to do is convert entities to something else. calibre will change them to unicode characters, which I think is a very good way to deal with them.

01-21-2014, 07:42 AM	#1
najgori Klak Posts: 174 Karma: 150374 Join Date: Sep 2011 Location: Belgrade, Serbia Device: many	regex help 1. p(?!([^<]+)?>) "find character 'p' everywhere except within html tags" 2. p(?!([^&]+)?;) "find character 'p' everywhere except within named entities" i would like to merge these 2 searches to 1, so my new search would skip both tags and named entities. i could replace all named entities with decimal, but this is indirect solution.

01-21-2014, 08:04 PM	#6
eschwartz Ex-Helpdesk Junkie Posts: 19,422 Karma: 85397180 Join Date: Nov 2012 Location: The Beaten Path, USA, Roundworld, This Side of Infinity Device: Kindle Touch fw5.3.7 (Wifi only)	instead of saying "(either not this \| or not that)", you need to say "not (either this \| or that)". Currently, if either condition is true, a match is made. A "p" in an html tag isn't in an entity, and vice versa, so of course it matches. ... Here you go: Code: p(?!([^<&]+)?(>\|;)) First we search for the "p" then we check ahead to see if the following is not there: any character string that does NOT contain either a "<" or "&" , followed by either a ">" or ";" EDIT: If your sentence ends in a ";" this won't match, (you can try adding a negative lookbehind for a "&" and seeing where that leads you) so you're probably better off using calibre to convert all entities to unicode. And tags won't have that problem. EDIT #2: added regex pipe and parentheses to beginning of answer, to clarify how it works. Last edited by eschwartz; 01-22-2014 at 01:51 AM.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Need help for a regex	wobohohoho	Sigil	4	01-02-2013 04:42 AM
Help me with regex please.	eVrajka	Library Management	5	08-15-2011 12:17 PM
regex help please	thevoiceofcheese	Calibre	2	08-01-2011 11:27 PM
Regex	Faster	Sigil	2	04-24-2011 09:08 PM
Help with a regex	A.T.E.	Calibre	1	04-05-2010 07:50 AM

01-21-2014, 10:08 AM	#2
mzmm Groupie Posts: 171 Karma: 86271 Join Date: Feb 2012 Device: iPad, Kindle Touch, Sony PRS-T1	--edit, misread the question, can't figure out how to delete this. Last edited by mzmm; 01-21-2014 at 11:02 AM.

01-21-2014, 02:06 PM	#3
najgori Klak Posts: 174 Karma: 150374 Join Date: Sep 2011 Location: Belgrade, Serbia Device: many	it looks like a common task: "mark every letter 'p' but skip tags and named entities". you can "cheat" by converting named entities to numbers which is recommended for ebooks (if i am correct?), but i am looking for "elegant" solution for this problem.

01-21-2014, 02:57 PM	#4
eschwartz Ex-Helpdesk Junkie Posts: 19,422 Karma: 85397180 Join Date: Nov 2012 Location: The Beaten Path, USA, Roundworld, This Side of Infinity Device: Kindle Touch fw5.3.7 (Wifi only)	use the following format to check for "either _ or _": (either\|or)

01-21-2014, 04:01 PM	#5
najgori Klak Posts: 174 Karma: 150374 Join Date: Sep 2011 Location: Belgrade, Serbia Device: many	p((?!([^<]+)?>)\|(?!([^&]+)?;)) does not work.

01-22-2014, 02:49 AM	#8
najgori Klak Posts: 174 Karma: 150374 Join Date: Sep 2011 Location: Belgrade, Serbia Device: many	@mrmikel. elegant meaning single regular expression instead of many - entiites are not important @eschwartz. also does not work. it misses some characters. i am in a process of learning regular expressions and this was just a question "could this be done?" - not very important to me. big thanx to everybody who tried to find solutions.

Advert

Advert