3

I'm trying to make excerpts from long texts while deleting everything that comes after the last space if the string length is more than 110 characters.

        $string = 'Стихи похожи на людей: помнят прошлое и ничего не знают о будущем, хотят жить вечно, a страница уже перелистывается.';

        if (mb_strlen($string ) > 110) {
            $pos = mb_strpos($string , ' ', 110);
            $excerpt = rtrim(mb_substr($string, 0, $pos), '.,—-_!@\'"()*#~').'...';
        }

If I print with print_r(mb_strlen($pos)); the result of $pos is 0, and its working correctly if I change $pos to $pos = mb_strpos($quote_content, ' ', 99);.

The last word in this case is 16 characters long and the whole string is 116 characters long so it makes perfect sense as to why a 99 offset works while anything above will result a $pos value of 0 thus instead of making an excerpt it just returns ... (based on the current example).

I have quite a few strings here with different string length and words length so I need a dynamic solution that will work in all cases. Any ideas?

Ricardo
  • 1,653
  • 8
  • 26
  • 51
  • 1
    Sorry, It's early. Have you tried to loop counting backwards to see if it will shorten the string to the desired length? So since 100 returns 0 then loop back to 99. etc etc – thecoolestguyever123 Aug 31 '18 at 11:50
  • 1
    There is a pretty good example in [the PHP manual](http://php.net/manual/en/function.wordwrap.php) See the sample code from `ju1ius` from 6 years ago – RiggsFolly Aug 31 '18 at 12:00

3 Answers3

1

A simple (and fast) way to cut a string to a fixed number of characters is with preg_replace:

$string = 'Стихи похожи на людей: помнят прошлое и ничего не знают о будущем, хотят жить вечно, a страница уже перелистывается.';
$excerpt = preg_replace('/^(.{1,110})\s.*$/u', '$1...', $string);
echo $excerpt;

Output:

Стихи похожи на людей: помнят прошлое и ничего не знают о будущем, хотят жить вечно, a страница уже...

The regex works by looking for some number of characters ^(.{1,110})\s (from 1 to 110) from the start of a string and up to a space character. Since the quantifier is greedy it takes as many characters as it can. Those characters are captured in a group. The rest of the string is then matched by .*$, and the whole string is replaced by the first capture group and three .'s ($1...), giving just the first portion as desired. The u flag on the regex means it will count unicode characters correctly. To adjust the length of the excerpt, simply change the 110 to whatever length you need.

Regex101 demo

Edit

The regex can also be modified to strip off any non-word characters (so you don't end up with the quick brown fox,...) by modifying it to insist that the last character of the capture group is a word character and then allowing the following character to be a non-word character:

$string = 'Стихи похожи на людей: помнят прошлое и ничего не знают о будущем, хотят жить вечно, a страница уже перелистывается.';
$excerpt = preg_replace('/^(.{1,23}\w)\W.*$/u', '$1...', $string);
echo $excerpt;

Output:

Стихи похожи на людей...

Updated demo

Nick
  • 138,499
  • 22
  • 57
  • 95
  • This is amazing. I didn't know you can do that with regex. Is there a way to remove any special characters and/or commas/periods from the end before we add three `.'s`? Otherwise, if there's a comma at the end it will become `a страница уже,...`. – Ricardo Aug 31 '18 at 14:41
  • @Ricardo FYI I benchtested 100k iterations of the accepted answer and my edited (which is more flexible) and the times were 0.315 for the accepted answer and 0.192 for this answer. – Nick Sep 01 '18 at 01:12
  • A lot of new information in this answer, no other choice than marking it as the (better) answer :) – Ricardo Sep 01 '18 at 08:56
1

This will cut the string at the last space without cutting words:

$excerpt = mb_substr($string, 0, mb_strrpos($string, ' ', -(mb_strlen($string) - 110)));

strrpos and mb_strrpos go backwards so you can search the last occurrence starting from the given position

Francesco G.
  • 164
  • 1
  • 6
  • This works, thanks :) Now I'm trying to decide whether `mb_substr` is better than regex (Nick's answer) in terms of performance, any insight on that? – Ricardo Aug 31 '18 at 14:56
  • I don't know about performance... as for my personal taste I usually prefer using regular functions if they exist, especially if you have to deal with other programmers not so fond on regular expressions. Of course this applies to yourself too: if you're not that fond wit regex you may have problems maintaining that code you didn't write ;-) This article may help about performances: [link](https://stackoverflow.com/questions/14342427/are-preg-match-and-preg-replace-slow) – Francesco G. Aug 31 '18 at 15:10
0

A lazy fix by checking all chars from 110 until space was found

    // lazy fix by checking all chars from 110 until space was found

    if (mb_strlen($string) > 110) {
        $p = 110;
        while(!($pos = mb_strpos($string , ' ', $p--))){};
        $excerpt = rtrim(mb_substr($string, 0, $pos), '.,—-_!@\'"()*#~') . ' ... ';
    }
Ulrich Dohou
  • 1,509
  • 14
  • 25
TsV
  • 629
  • 4
  • 7