Searching NOT

[tarisea]

For the life of me I can't figure out how to do these 2 searches.

1. How would I search for </p> where the is NOT a . before it?
The the first line would come up but the second one wouldn't.

<p class="calibre11">But that didn’t mean he wanted to turn out</p>
<p class="calibre11"> the lights and go to sleep either.</p>


2. Same problem different version. How would I search for </p> where it is followed by a character [a-z] instead of a <


<p class="calibre11">But that didn’t mean he wanted to turn out</p>
the lights and go to sleep either.</p>


Any help would be greatly appreciated.

TTRS

[theducks]

Originally Posted by tarisea

For the life of me I can't figure out how to do these 2 searches.

1. How would I search for </p> where the is NOT a . before it?
The the first line would come up but the second one wouldn't.

<p class="calibre11">But that didn’t mean he wanted to turn out</p>
<p class="calibre11"> the lights and go to sleep either.</p>


2. Same problem different version. How would I search for </p> where it is followed by a character [a-z] instead of a <


<p class="calibre11">But that didn’t mean he wanted to turn out</p>
the lights and go to sleep either.</p>


Any help would be greatly appreciated.

TTRS

([a-z,"])</p>\s+<p class=calibre\d+">([a-z])
searches for lowercase or comma or straight quote

\1 \2

This will miss some words that start with a capital: eg Martha or a quote.

[DiapDealer]

1) Regex mode.

[^.]</p>

Although you're probably going to want something like:

[^.?!"']</p>

if you're looking for broken paragraphs.

Number two seems like a really odd search. Your code's in a lot of trouble if you have naked text immediately following a closing paragraph tag. But:

</p>[a-z]

should find it, if for any reason there is.

Or if it should literally be anything OTHER than the beginning of a new tag... then:

</p>[^<]

[Doitsu]

theducks's suggestion is, of course, perfectly fine, however, if you want to merge sentences, you might find this extended expression helpful that I use to merge sentences:

Find: ([[:lower:]],*;*:*)</p>\s+<p[^>]*>\s*([[:lower:]])
Replace: \1 \2

This expression will search for:
- a single lower case character followed by zero or more commas, semi-colons or colons
- followed by </p>,
- followed by one or more white-spaces (including line-breaks),
- followed by <p> (with optional attributes),
- followed by zero or more spaces and a single lower-case character.

[Tex2002ans]

These are the three regex I use.

Regex #1: This catches all hyphenation at the end of the paragraphs. I replace this in a case by case basis just to make sure the combined word does not actually require the hyphen.

-</p>\s+<p>

Replace:

(NOTHING, just a complete blank)

Sample:

<p>This is a paragraph which has a hyphen-</p>
<p>ated paragraph.</p>

Regex #2: This catches every paragraph that ends in a character that is NOT a '>', '”' (right double quote), '?', '!', '.':

([^>”\?\!\.])</p>\s+<p>

Replace (make sure there is a SPACE afterwards):

\1

Sample:

<p>This is a sample,</p>
<p>of a paragraph that will</p>
<p>be caught by the regex above.</p>

Regex #3: This usually catches all paragraphs which were not combined by the above two, but begin with a lowercase letter (usually these should be combined, or there was a blockquote beforehand, or something odd in the text).

<p>[a-z]

Sample:

<p>In 2014, Tex gave an informative sample:</p>
<blockquote><p>Here is a quote.</p></blockquote>
<p>here is more information.</p>

These three tackle nearly all broken paragraphs in my experience.

Other oddities such as semi-colons or colons will be pointed out by Regex #2, and those can be fixed on a case-by-case basis. (Sometimes they should be combined, sometimes they should not).

I keep these three in my Sigil Saved Searches (Tools - Saved Searches). More info on how to use Saved Searches can be found here:

http://web.sigil.googlecode.com/git/..._searches.html