new excerpt function suggestion

[Denis de Bernardy]

/*

• sem_fancy_excerpt()
• -------------------
• Creates a fancy default excerpt

/

function sem_fancy_excerpt( $excerpt )
{
global $post;

$excerpt_length = 30;

if ( $excerpt == )
{

$excerpt = $post->post_content;
$excerpt = apply_filters( 'the_content', $excerpt );
$excerpt = strip_tags( $excerpt );
$excerpt = preg_replace( "/^{(\w+(\W+\w+){".($excerpt_length-1)."}[}\.]*\.)[^{$]+/", "$1 (...)", $excerpt );}

}
return $excerpt;
} end sem_fancy_excerpt()

remove_filter('get_the_excerpt', 'wp_trim_excerpt');
add_filter('get_the_excerpt', 'sem_fancy_excerpt');

[Denis de Bernardy]

the above works a bit like wp_trim_excerpt, but returns full sentences.

[Denis de Bernardy]

preg_replace( "/^{\W*(\w+(\W+\w+){".($excerpt_length-1)."}[}\.]*\.)[^{$]+/", "$1 (...)", $excerpt );}

(just in case)

[Denis de Bernardy]

I added the file to the alpha version of my plugin. I left the test code at the end of it, should you want to play with it.

The backlink_excerpt method is extremely forgiving to html errors and will eat all but the worst coded html pages.

I think there is room for a few more enhancements. In particular, if you have:

• upper text
• <a ...>blah blah</a>
• lower text

The excerpt will be 'blah blah', whereas it would make more sense to have the upper and lower text. But presumably, it remains good enough for now.

[Denis de Bernardy]

a few dozen test cases later, it is:

$excerpt = preg_replace( "/^{\W*(\w+(\W+\w+){".($excerpt_length-1)."}((?!(\.|\n|\r)).)*(\.|\n|\r))(\w|\W)+/", "$1 (...)", $excerpt );}

rather than /^{(\w+(\W+\w+){".($excerpt_length-1)."}[}\.]*\.)[^$]+/

[Denis de Bernardy]

/^{\W*(\w+(\W+\w+){".($excerpt_length-1)."}((?!(!|\?|\.|\n|\r)).)*(!|\?|\.|\n|\r))(\w|\W)+/}

even. ;)

edited on: 04-20-05 21:11

[MC_incubus]

Probably not something to be considered for WP 1.5.1 (as it's a feature, and a regex feature, at that!)

Isn't this essentially in plugin form anyway?

[Denis de Bernardy]

the fancy excerpt is a plugin, yes. however, the attached file contains the following function, that Matt requested the other day in the hackers list:

function sem_backlink_excerpt( $text, $link )
{

clean up the text

$text = preg_replace( "/<!DOC/", "<DOC", $text ); strip_tags bug
$text = preg_replace( "/[\s\r\n\t]+/", " ", $text ); normalize spaces
$text = preg_replace( "/ <(h1|h2|h3|h4|h5|h6|p|th|td|li|dt|dd|pre|caption|input|textarea|button|body)[^{>]*>/", "\n\n", $text );
$text = strip_tags( $text, "<title><a>" ); just keep the tags we need}

echo Markdown( $text );

split into paragraphs

$p = explode( "\n\n", $text );

fetch the title

$title = preg_replace( "/^[<]*<[^>]*title[>]*>([^<]*)<\/[>]*title[^{>]*>(\w|\W)*/", "$1", $p[0] );}

fetch the first paragraph with the link

$sem_regexp_pb = "/(
/|
\|\*|\?|\+|\.|\^{|
$|\(|\)|\[|\]|\}

\{|\})/";

$sem_regexp_fix = "

$1";
$link = preg_replace( $sem_regexp_pb, $sem_regexp_fix, $link );

for ( $i = 0; $p[$i] && !$excerpt; $i++ )
{

if ( preg_match( "/<a[^{>]+".$link."[}>]*>/", $p[$i] ) )
{

$context = preg_replace( "/.*<a[^{>]+".$link."[}>]*>([^{>]+)<\/a>.*/", "$1", $p[$i] );
$excerpt = trim( strip_tags( $p[$i] ) );}

}

}

return the result
return array( $title, $context, $excerpt );
} end sem_backlink_excerpt ()

edited on: 04-21-05 09:58

[Denis de Bernardy]

sem_backlink_excerpt() dines on an html page and an url. it returns an array with:

• the html page's title
• the excerpt, = the first paragraph that contains the link (rather than a more or less random string containing it)
• the context = the link text

at the moment the context is the link's text. in the future, i'll likely change this to be the full sentence. likewise, i might change the function to return the paragraphs before and after, when the paragraph is a mere sentence.

from visiting wp source, i think using the function requires rewriting some areas of code related to fetching pingback excerpts. and that you do not necessarily require the $context variable.

i do, however, for two plugins i've in my pipe. thus, it would be nice if there were some hook i can catch to bypass it in the future.

[matt]

Rolled in some of this code and it seems to be working well. Thanks! :)

[(none)]

Milestone 2.0 deleted