[Resolved] news excerpt...

jassh

okie guys..
here's a quick one for ya..

i've flip thru the manual and books..
but i yet (or probably missed out) to come accross
on how to make an excerpt out of a paragraph...

okay.. for instance, i just want the first 20 words of a paragraph
which of course available in the form of variables,
taken from a database..

so, a sentence like this:

Within the current incarnation of (X)HTML tables are the building blocks to well organized and attractive designs. They allow the developer an incredible amount of control over the visual representation of information and graphics. However they come with a catch, they are an absolute headache to write into your PHP code. Thankfully there is a solution: the PEAR HTML_Table class.

would appear as something like this:

Within the current incarnation of (X)HTML tables are the building blocks to well organized and attractive designs. They allow the...

any idea how i can achieve this?
i tried to manipulate (str_word_count),
problem is, it doesnt includes things like period, open bracket.. etc.
my other concern is that, imagine making an array out of a full article..
hundreds of words.. maybe thousand...
now, isnt that ridiculous??

so guys..
any better idea??

thanks in advanced..
-jassh

drawmack

the average english word is 5 characters long. Let's use that.

<?php
$synop_words = 20 //the number of words in the synopsis
$padding = 10 //extra characters padded to the end
$word_length = 5 //the length of an arverage word
$synop = substr($article,0,(($word_length * $synop_words)+$padding);

//okay so $synop is close to want we want but we want 
//to ensure that it does not break in the middle of a 
//word or contain more then 20 real words.

$ary_synop = explode(' ',$synop);
$i=0;
$synop = '';
while($i<count($ary_synop) && $i<20) {
  $synop .= ' ' . $ary_synop[$i];
}
$synop = substr($synop,1) . '...';

echo $synop;
?>

Weedpacket

I'd just use preg_split() (using /\s+/ as the regexp), limit it to 21, and then join the first twenty elements of that back together with spaces.

Or use strtok() for a while.

drawmack

but my way is better because it's more lines of code and IBM pays per line.😃

Weedpacket

Originally posted by drawmack
but my way is better because it's more lines of code and IBM pays per line.😃

Plus yours would be more easily adapted to reading the input incrementally from a file, I guess.

majik-sheff

preg_match('/^{\s*(\S+\s+){'} . $words . '}/', $in, $out);

Put your word count into $words, and your paragraph into $in.

Your result will show up in $out[0].

jassh

whoaa..
way cool guys...
guess somebody has been dippin in the news excerpt businees eh

okie dokie...
i've been thru all the code..
and you guys rawk... totally..

so here to sums up the winner,
plus with some personal comment...

the second runner up..

$news = "Within the current incarnation of (X)HTML tables are the building blocks to well organized and attractive designs. <br>";
$news .= "They allow the developer an incredible amount of control over the visual representation of information and graphics. <br>";
$news .= "However they come with a catch, they are an absolute headache to write into your PHP code. <br>";
$news .= "Thankfully there is a solution: the PEAR HTML_Table class.";


$synop = substr($news,0,125); 
$news_array = explode(' ',$synop); 
$words = count($news_array) - 1;

for ($i=0; $i<$words; $i++) {
$excerpt .= " " . $news_array[$i];
}
echo strip_tags($excerpt) . "....";

Comment: the best code if u are paid per line..
and probably the most simplest logic ever..
i love this one.. its a classic...
although i did a lil' modifications on it
but as weedpacket said, it can be adapted to read lines from files
and by far the best in performance speed...
coz i think it read only the first 125 character..
this is truely amazing..

and the winner is

$news = "Within the current incarnation of (X)HTML tables are the building blocks to well organized and attractive designs. <br>";
$news .= "They allow the developer an incredible amount of control over the visual representation of information and graphics. <br>";
$news .= "However they come with a catch, they are an absolute headache to write into your PHP code. <br>";
$news .= "Thankfully there is a solution: the PEAR HTML_Table class.";

$wordcount = 20;
preg_match('/^\s*(\S+\s+){' . $wordcount . '}/', $news, $out); 
echo strip_tags($out[0]) . "...";

Comment: By far, the most simplest, yet technically great..
well, i'd say this would be the quickest way to do it,
i dont know if this method superseeds the above for speed
but i truely love this one too...

and the rest of the code..
well, using strtok i believe it tokenize the string
into its individual words by using the space character as the token
so by theory, i guess if there is 1000 words,
it will tokenize every single words, hence performance degration
but please correct me if i'm wrong eh weedpacket

preg_split().. hmm.. i know it can somehow split text,
but ya need some sort of pattern... and i guess this would do too
lemme see...



$wordsmax = 20;
$keywords = preg_split ("/[\s,]+/", $news);
for ($i=0; $i<$wordsmax; $i++) {
$excerpt .= " " . $keywords[$i];
}

echo $excerpt . "...";

hey... what did cha know..
it works...
nice one 😃

anyway...
just wondering..
with so much method for one objectives..
how do i know which code take the least time, the less resource,
and the most simplest way to execute???

to sum it all up...
u guys are cool... totally... 😃 😃
thanks a lot guys!!!!

and now.. its time to click the resolved buttons 😃

cheers;
-jassh

drawmack

Originally posted by jassh
anyway...
just wondering..
with so much method for one objectives..
how do i know which code take the least time, the less resource,
and the most simplest way to execute???
[/B]

you go into the code library and download my benchmarking class. Then you test eash method for efficiency on differing sized inputs. I have a feeling what you'll find is that speed is so close as to not even be a concern.

majik-sheff

I benchmark the winner at about 3x faster than the runner up and about 7x faster than the also-ran. The numbers for the first two do not vary much by the input size, the last one really suffers on large inputs.

pats self on back😉

jassh

Originally posted by majik-sheff
pats self on back😉

aww.. lemme pats u on your back too :p

anyway, yeah.. i aggree on majik's results...
i guess he really deserve a pat on his back..

but kudos goes to drawmack..
ur benchmark's classes is a darn good tool for application efficiency..
thanks pal... 😃

hope you guys have a nice day ahead of ya..

cheers;
-jassh

OMNIpotus

Props to your mom, Sheff. Big ups indeed.