[RESOLVED] Regex to match sentences [a-z] ?

lar5

Hi, i am searching for a way to find full sentences in a html page.

maybe a regex will help that matches the following:
Min. 5 words followed by an "." // containing only [a-z 0-9 , .]

I tried for 2 hours now but my regex knowledge is just too limited.

Best Regards,
Lar

bpat1434

Hmm... something like:

([a-zA-Z0-9\,\"\'-\s] [a-zA-Z0-9\,\"\'-\s] [a-zA-Z0-9\,\"\'-\s] [a-zA-Z0-9\,\"\'-\s] [a-zA-Z0-9\,\"\'-\s]) ([a-zA-Z0-9\,\"\'-\s]).

it would match at least 5 words (with numbers)...

$pattern = "~([a-zA-Z0-9\,\"\'\-\s]* [a-zA-Z0-9\,\"\'\-\s]* [a-zA-Z0-9\,\"\'\-\s]* [a-zA-Z0-9\,\"\'\-\s]* [a-zA-Z0-9\,\"\'\-\s]*) ([a-zA-Z0-9\,\"\'\-\s]*)\.~";

$sent = 'This is a sentence of 7 words.  I am a man of few words.  This sucks.  How about 5 words only.';

preg_match_all($pattern, $sent, $matches);

echo '<pre>';
var_dump($matches);

Outputs:

array(3) {
  [0]=>
  array(3) {
    [0]=>
    string(30) "This is a sentence of 7 words."
    [1]=>
    string(26) "  I am a man of few words."
    [2]=>
    string(25) "  How about 5 words only."
  }
  [1]=>
  array(3) {
    [0]=>
    string(23) "This is a sentence of 7"
    [1]=>
    string(19) "  I am a man of few"
    [2]=>
    string(19) "  How about 5 words"
  }
  [2]=>
  array(3) {
    [0]=>
    string(5) "words"
    [1]=>
    string(5) "words"
    [2]=>
    string(4) "only"
  }
}

You can modify it as you want, but this also takes into account hyphenated words (non-payer) and quotes (Sally said "yes", I'm okay with that).

Hope that helps....

lar5

You are a genius! Thanks! Works almost perfect.
Do you know a way to remove all items from the array that include more then 2 commas?

bpat1434

Loop through the array with a for()...

IF the sentence has 2+ commas, don't add to new array
ELSE add it to array

Something like:

for($i=0; $i<count($matches[0]); $i++)
{
  $string = $matches[0][$i].$matches[1][$i];
  if(substr_count($string, ',') <= 2)
  {
    $matching[] = $string;
  }
}

Then, you shoudl just be able to use $matching to have an array of all sentences that have less than or equal to 2 commas....

Merve

[man]preg_grep/man

I think that the expression would be /,{3,}/ but I'm not sure; test it out.

lar5

Thanks, the code works great!