search script help...

[mitov]

I have a php indexing and search script. Problem is that it indexes all of the information on the pages.

Is it possible to get it to index only what i specify? For example, in the normal page, have it layed out like this:

...menu....
<spider>
Content i want indexed
</spider>
Copyright etc. etc.

Is it possible to get the script to only index what is between the <spider> tags?

Any help will be appreciated! I am struggeling and have no knowledge php... thats why im using a ready made script.

Please...

[kakki]

preg_match_all( '#<spider>.*</spider>#msiU', $content, $filteredContent );
// then all is stored array $filteredContent

[mitov]

Ok. My script runnes of a file called class.search, which can be found at www.hairandbeautypages.com/find/classes (right click and save target as).

Where do i have to place that in the script?

[mitov]

Actualy, i think that the script which indexes is class.index which can be found at www.hairandbeautypages.com/find/admin/classes

[kakki]

// ... grab content of page and store that in $content
preg_match_all( '#<spider>.*</spider>#msiU', $content, $filteredContent );
// save $filteredContent[ 0 ] in DB

[mitov]

Hey,

I found this in my script:

// Grab the files content
$pageData = "";

				while(!feof($fp))
					$pageData .= fgets($fp, 4096);

				fclose($fp);

				// Try and get the pages title
				if(eregi("<title>([^<]*)</title>", $pageData, $titleArray))
				{
					// The page as a title tag, extract it
					$pageTitle = $titleArray[1];
				}
				else
				{
					// Page contains no title, give it the default title
					$pageTitle = $GLOBALS["FF_LANG"]["defaultPageTitle"];
				}

				unset($titleArray);

				// Has this page got any ignore tags?
				while(preg_match('%<!--FF_IGNORE-->(.*?)<!--FF_END_IGNORE-->%is', $pageData, $ignore_matches)) {
					$pageData = str_replace($ignore_matches[0], '', $pageData);
				}
				unset($ignore_matches);

				// Has this page got any script tags?
				while(preg_match('%<script[^>]*>.*?</script>%is', $pageData, $ignore_matches)) {
					$pageData = str_replace($ignore_matches[0], '', $pageData);
				}
				unset($ignore_matches);

				// Only grab the contents of the page body and then
				// strip out any script tags
				eregi("(<body[^>]*>.*</body>)", $pageData, $bodyArray);

				if (sizeof($bodyArray) > 0) {
					$strippedPageData = strip_tags($bodyArray[1]);
				} else {
					$strippedPageData = '';
				}

				unset($bodyArray);[/QUOTE]

Where you see the body tags, i changed them to spider but it still dosent work. It keeps indexing the entire page...

Thanks for all your help by the way.

[kakki]

i had the problem, that eregi cant do a regexp ungreedy( or i am too dumb for this )
maybe you can change all eregi's to:

preg_match_all( '#<spider>.*</spider>#msiU', $pageData, $filteredContent ); 
$pageData=$filteredContent[ 0 ]; // but you will loose all other occurences of <spider>...</spider>

[mitov]

so.. remove the entire line

eregi("(<body[^{>]>.</body>)",} $pageData, $bodyArray);

and put this in its place:

preg_match_all( '#<spider>.*</spider>#msiU', $pageData, $filteredContent );
$pageData=$filteredContent[ 0 ]; // but you will loose all other occurences of <spider>...</spider>

Is that what i should do? Sorry for making you break it down for me... im not a php guy 🙁

Thanks for all your help!

[kakki]

this line:

eregi("(<body[^>]*>.*</body>)", $pageData, $bodyArray);

into this:

preg_match_all( '#<body.*</body>#msiU', $pageData, $bodyArray ); 
$bodyArray[ 1 ]=preg_replace( '#^.*>#msiU','',$bodyArray[ 1 ] );

i think, that should do the trick

[thepeccavi]

maybe you could try changing the line:

eregi("(<body[^>]*>.*</body>)", $pageData, $bodyArray);

to

eregi("(<spider>.*</spider>)", $pageData, $bodyArray);

This changes the code less than above...

PS... this may not work... I have rarly used eregi() functions

Hope this helps!

[mitov]

thanks to both of you. thepeccavi, i had already tried that... thank though. I will try it out now and let you know what happens...

Miro

[mitov]

It still dosent work... Have a look at the last search result:

http://hairandbeautypages.com/find/index.php?query=hebe&type=advanced&searchType=2

you can still see the menu... and the class.index is with kakki's modification...

[thepeccavi]

try deleting all the indexed pages and start indexing them again... this may help...

[mitov]

Have just tried it and still no luck...

Any other ideas? This search has been driving me mad and i feel like ripping it appart if it was paper! I am realy despirate now as the entire site has to be up and runing by wednesday...

Any other suggestions would be VERY HIGHLY appreciated!

Thanks for all your help guys...

[thepeccavi]

This would possibly mean MAJOR re-designing, but you could move the entire site onto PHP and MySQL generation... the menus, and footers could all be in your PHP script, but the data that you want to index could be in a MySQL database...

This would mean that searching would be just a simple MySQL query, and you could use mod_rewrite to make nice looking URLs!

Another benefit of this would be that you could create a cool admin centre for your client where they could edit the content and even the menus!

What do you think of this idea? Is it too late to try it now?

[mitov]

Well, the thing is that 1 its a bit too late, 2 i have no knowledge of php what so ever, and 3 that to get it made professionaly, i am looking at roughtly £3000, which i have not got at the moment.

Is there ANYTHING which i may try?

Miro

[thepeccavi]

£3000????????? WOW!!! maybe web design IS well paid! I'd do it for less then £100... oh well.....

Sorry, I can't think of anything else, as I have never (I think) used eregi() or preg() and the related functions... Sorry!

[mitov]

Its Ok. Thanks for your help...

Less than £100 and make it look like i have it at the moment? WOW! You are cheap!

Do you know anyone who may be able to help me?

[thepeccavi]

I just noticed that one image on your searhc page links to "file:///C:/Documents and Settings/Owner/My Documents/hb PAGES/img/salons.png"! no wonder it appears as broken!

[mitov]

Which image is that? Where is it positioned?