[RESOLVED] Extracting description and keywords metatag

[teguh123]

Is there a special PhP command for this or are we left with good old regex?

[bpat1434]

What request are you examining? If it's one you create with [man]fopen/man or [man]fsockopen/man you can use regex to do it. Alternatively if you know the URL, you can use [man]get_headers[/man]($url, 1); to give you an associative array of headers from that URL.

And of course if you're using cURL you can use [man]http_parse_headers/man on the curl headers response to get an associative array of headers.

So it depends on how you're calling the request and what you're trying to achieve.

[teguh123]

Good pointer. Hmm... I don't think headers meta descriptions are in headers. It's in the head section. Not header section. http://id2.php.net/manual/en/function.http-parse-message.php

I'll play around.

[bradgrafelman]

Note another option might be using [man]DOMDocument[/man].

[teguh123]

Does this domdocument and http_parse_headers

Is there a special library that need to be there like curl?

[bpat1434]

DOMDocument: There is no installation needed to use these functions; they are part of the PHP core.

http_parse_headers: This » PECL extension is not bundled with PHP.

So, to recap:

DOMDocument: Included in php5 by default, not not included in php4. Available for php4 via the PECL extension.
http_parse_headers: Not included in any php installation, a part of the PECL extension.

[teguh123]

I think I'll just use regex

<.?meta.?description.*?=\"\'[\"\']

Is the syntax correct?

It should match <meta anything description anything= 'or " then the next ' or ""

Am I correct?

<meta name="description" content="This is a discussion forum for discussing all aspects of the serverside scripting language PHP" />

Should produce This is a discussion forum for discussing all aspects of the serverside scripting language PhP

[leatherback]

something like this should work:

preg_match_all('/<[\s]*meta[\s]*name="?' . '([^>"]*)"?[\s]*' . 'content="?([^>"]*)"?[\s]*[\/]?[\s]*>/si', $contents, $match);

[teguh123]

Thanks.

I just need description and keywords meta. So it should be

preg_match_all('/<[\s]*meta[\s]*name="description?' . '([^>"]*)"?[\s]*' . 'content="?([^>"]*)"?[\s]*[\/]?[\s]*>/si', $contents, $match); 

and

preg_match_all('/<[\s]*meta[\s]*name="keywords?' . '([^>"]*)"?[\s]*' . 'content="?([^>"]*)"?[\s]*[\/]?[\s]*>/si', $contents, $match); 

right?

[bpat1434]

No, you'd want something like

preg_match_all('~~<.*meta.*name=["|\'](description|keywords)["|\'].*content=["|\'](.*)["|\'].*/>~iU');

You can view a demo of that exact pattern for this page here: click here. You can also view the source I used. Hope it helps.

[Weedpacket]

If I may tweak that a notch... 🙂

 preg_match_all('~<.*meta.*name=(["|\'])(description|keywords)\1.*content=(["|\'])(.*)\3.*/>~iU'); 

Not that I've tested it, but it should ensure that the quotes around the attribute values match.

[bpat1434]

Okay, tested & updated. You can see it at my above links. Forgot about that Weedpacket. Thanks for the tweak 😉

[teguh123]

Damned I humbly bow to the wisdom of superior men.

[bpat1434]

If this is resolved, please mark it as such.