Looking for the best way to parse a specific HTML tag

s_shrum

I have created a script that I am adding some features too.

One of these features will be to allow the script to return back META tag information from a directory of locally stored hypertext documents. Specifically, this data will (for me) be formatted as such:

....
<head>
<meta http-equiv="author" content="Sean Shrum">
<meta http-equiv="email" content="sean@shrum.net">
<meta http-equiv="created" content="020101">
<meta http-equiv="modified" content="030407">
....
</head>

I figure I can use the '</head>' as a trigger to end looping so I don't parse the entire HTML files.

I've got the code to read in the data line by line but I am having issues about how to parse the meta lines for the data. Issues like:

How do I search for substrings with double quotes in them?
What if the user switches the order of the keys?
etc.

BTW: I would rather code the process myself as opposed to using some sort of lib that would do this for me...seems like it should be a fairly easy thing to accomplish, knowing the particulars of PHP.

Any help would be apprechiated.

TIA

trexx

first i would parse each line between the starting and closing head tag into an array.

then i'd loop through the array and parse out key and value of the meta tags with a regular expression. (take a look at http://www.php.net/manual/de/function.preg-match.php)

and finally you store key and value in an asssociative array.

s_shrum

But how do I handle the double quotes in search strings??? As escaped entities?

ala: "\""

Weedpacket

Yes. Assuming your strings are double-quoted themselves...

trexx

yes. it's long ago since i used regex the last time but i think it should look like (untested)...

preg_match( "/http-equiv=\"([^\"]*)\" content=\"([^\"]*)\"/i", $line,  $matches );

then you SHOULD have the results of the first and second brace-pair in $matches ( try print_r($matches) );

s_shrum

Ah! Got it working. Here is the code (with a bit extra) if anyone is interested:

// get meta data out of html documents
if ( $ext == "shtml" ) {
	$fp = fopen(".." . $HTTP_GET_VARS['dir'] . "/" . $basename . "." . $ext, 'r');
	while(!feof($fp)) {
		$line = trim(fgets($fp, 1024));
		if( strpos( $line, "<meta") !== false ) {
			$key_spos = strpos($line, "http-equiv=\"") + 12;
			$key_epos = strpos($line, "\"", $key_spos);
			$key = substr($line, $key_spos, $key_epos - $key_spos);
			$val_spos = strpos($line, "content=\"") + 9;
			$val_epos = strpos($line, "\"", $val_spos);
			$val = substr($line, $val_spos, $val_epos - $val_spos);
			print $basename." - ".$key_spos."/".$key_epos."-".$val_spos."/".$val_epos." = " . $key."+++".$val."<br>";
		} elseif ( strpos($line, "</head>" ) === true ) {
			wend;
 		}
	}
}

Thanx again

trexx

Ok... guess your solution looks more impressive 😃

superwormy

As a note, often times meta tags appear over several lines, so I would be careful about making it parse every line into a seperate elelment on the array... specifically meta keywords and meta descriptions often run more than one line...

A better method might be to write a function that first finds <head>, read until you find </head>

Now, search within that for <meta and read until you find >

Then, use regex's ( or the general expression you just wrote that you used to get between the <head> and <meta tags ) on each of what you foudn before to get teh content and type...