Regular Expression Help Needed

voidstate

I'm rather new to regular expressions and am having trouble working out how to write one to solve the following problem:

I want to split a page of text into paragraphs anywhere there is two or more new lines. SO the following:

Lorem ipsum dolor sit amet, consectetuer adipiscing elit. Nunc eleifend , , fringilla nec, iaculis in, condimentum id, felis. Nullam at turpis. Proin ligula. Praesent fringilla mi et justo.

Duis turpis. Quisque egestas. Cras fringilla sapien non dui. Cras sit amet justo in justo sagittis adipiscing. Aliquam interdum. Aliquam ut condimentum eleifend. In id mauris vitae felis luctus lacinia. Quisque adipiscing. Sed lobortis sem vitae velit.

Vestibulum posuere odio a arcu. Pellentesque egestas. Fusce consequat mollis mi. Duis at urna nec nulla sollicitudin facilisis. Sed tristique sagittis velit, id accumsan sem metus ac augue. Vivamus blandit metus. Nullam consequat. Ut leo quam, dignissim et, auctor ut, placerat vel, velit. Maecenas molestie laoreet wisi.

would be split into an array with three members, which I would then place paragraph tags around an implode to create this:

Lorem ipsum dolor sit amet, consectetuer adipiscing elit. Nunc eleifend , , fringilla nec, iaculis in, condimentum id, felis. Nullam at turpis. Proin ligula. Praesent fringilla mi et justo. 
Duis turpis. Quisque egestas. Cras fringilla sapien non dui. Cras sit amet justo in justo sagittis adipiscing. Aliquam interdum. Aliquam ut condimentum eleifend. In id mauris vitae felis luctus lacinia. Quisque adipiscing. Sed lobortis sem vitae velit. 
Vestibulum posuere odio a arcu. Pellentesque egestas. Fusce consequat mollis mi. Duis at urna nec nulla sollicitudin facilisis. Sed tristique sagittis velit, id accumsan sem metus ac augue. Vivamus blandit metus. Nullam consequat. Ut leo quam, dignissim et, auctor ut, placerat vel, velit. Maecenas molestie laoreet wisi.

Obviously the expression would have to ignore white space and recognise more than two newlines as a single gap.

Anyone help?

voidstate

PS I'm guessing this board is doing something very similar when I post this question. How ironic... 😃

ClevaTreva

Assuming the text is where? A Text file?

If so, can you read it into an array?

e.g.

$textbit=file('filename');

This will split the text file up into individual rows.

then

$newpara=1;
for ($i=0;$i<count($textbit);$i++)
{
if ($newpara==1)
{
echo "<p>";
$newpara=0;
}
if (rtrim($textbit[$i])=="")
{
echo "</p>";
$newpara=1;
} else {
echo rtrim($textbit[$i]);
}
}

That's assuming the text file is in the format of:

text
text
text
blank line
text
text
etc

Otherwise play around with it a bit. I didn't check the code, so make sure it is syntax correct!

Trevor

devinemke

assuming the text in is "page.txt"

$page = file_get_contents ('page.txt');
$page_array = explode ("\r\n\r\n", $page);
foreach ($page_array as $value) {echo '<p>' . $value . '</p>';}

The tricky thing is the newlines. The above code would work fine on Windows where the newlines are\r\n. For UNIX use \n

devinemke

or simply convert all newlines to the UNIX style before exploding:

$page = file_get_contents ('page.txt');
$page = str_replace ("\r\n", "\n", $page); // win to unix
$page = str_replace ("\r", "\n", $page); // mac to unix
$page_array = explode ("\n\n", $page);
foreach ($page_array as $value) {echo '<p>' . $value . '</p>';}

ClevaTreva

Using file_get_contents assumes he has php 4.3.0 or higher. Quite a lot of hosts are stuck on 4.2.2 or 4.2.3

Otherwise, I like that script. 😃

Trevor

swr

The advice of exploding on the double newline is good for what you've asked for, but I wonder if you wouldn't be content to use [man]nl2br/man to convert your ASCII line breaks to HTML line breaks.

goldbug

Dunno if it works (haven't tested), but something like this?


$pattern = "/(.*?)(\n|\r\n|\r){2,}/i";
$replacement = "<p>\$1</p>";

echo preg_replace($pattern,$replacement,$content);

voidstate

Thanks for all the feedback. I am getting the text from a html textarea.

After loads of experimentation with what goldbug posted, this seems to work:

$textElements = split("((\s|\t)*(\n|\r\n|\r){1}(\s|\t)*){3,}", $text);

Seems kinda cumbersome... but it seems to work.

Thanks everyone.

voidstate

edit: the "s" characters have backslashes before them but it' not showing up for some reason