Converting unless HTML

[Hikari]

I need some quick help. I need to run some str_replace() functions on a variable (what will be a webpage) UNLESS the code is HTML.

Does anyone know how I would go about doing this?

[pnoeric]

Please be more specific: do you have one string that has the entire contents of the web page? Are you trying to str_replace the HTML tags, or everything that is NOT part of an HTML tag, or ... ?

[Hikari]

Ok, heres an example.

I'm grabbing a webpage using this:

$OpenFile = fopen("$url", "r");
$RetrieveFile = fread($OpenFile, 200000);
$GrabData = eregi("$GrabStart(.*)$GrabEnd", $RetrieveFile, $DataPrint);

So.. what I want to do is several str_replaces to it
(ie:

$output = str_replace('a', 'e', $DataPrint[1]);
$output = str_replace('n', 'x', $DataPrint[1]);

)

• Just as an example. So I want to do a whole bunch of those, UNLESS its HTML code. I want to leave all the HTML code alone. I assume some type of regex would be used in it.

[pnoeric]

Yeah, you're probably right, a regex would do the trick. It'd be a hell of a regex though! :-) I think I had to do something like this once, and ended up writing a mini-parser.

[Hikari]

Umm.. well, it wouldn't be that bad. I mean.. all it would have to be is:

<[^>]*[^>]>

That would be the string to see the HTML. But how do I take that, and make it so that in my replacing it ignores all of the HTML (using that regex)

[Weedpacket]

I'm guessing it could be done with a preg_replace, involving suitable assertions about what characters may surround which characters. I may come back with one, but for now I've got another suggestion.

I'm assuming that the HTML is to spec - the only < and > characters in the file are those that begin and end tags. Things get really messy otherwise, and you end up like pnoeric writing a mini-parser.

//Using fread() with a ridiculously big size
// is not a good idea - PHP has to find and
// reserve that much memory before it can
// begin reading.
$string = join('',file($url)); // Either do this
                               // use a proper
                               // fread() loop.
$stringbits = preg_split('/[<>]/',$string);

What does that achieve? As it happens, all the even-numbered array entries contain the text outside tags, and all the odd-numbered entries contain text that is inside tags.

So do the conversion on all the even-numbered elements:

$len=count($stringbits);
for($i=0;$i<$len; $i+=2)
  $stringbits[$i]=convert($stringbits[$i]);

Put <> back around the odd-numbered ones:

$len=count($stringbits);
for($i=1;$i<$len; $i+=2)
  $stringbits[$i]='<'.$stringbits[$i].'>';

And reassemble the pieces:

$convertedstring = join('',$stringbits);

I think....

[pnoeric]

WP, as usual, an elegant solution ;-) as you hinted at, I wrote my parser to deal with poorly formed HTML and also with "<" and ">" characters inside of html tags-- I don't think those can be put into the unicode "&lt;" tags, can they? anyhow, your regexp code is great!

[Weedpacket]

Thank you; however, I've got a better solution. Still assuming a separate convert() function to do the actual str_replace work, (and assuming that $string contains the text in question) replace the above code with the following:

$string = preg_replace("/>([^<]+)</es","'>'. convert('\\1').'<'", $string);

I like the /e modifier 🙂

< and > are explicitly not allowed anywhere except when starting and ending tags, but I understand Everyone's Favourite Browser (for one) does let them in without complaining (even within attribute values). The result of course is a lot of dodgy HTML out there...

[Hikari]

Thanks for all your help WP. I'm trying to put this into application, but I'm having a bit of trouble

keeps trying hehe