Help Isolating Words

Sharif

Okay so I am trying to creating a google-like search script. Of course it can't be that powerful but I can try!

Here is what I am trying to do, say the text being searched is:

"The text here being search is so boring but it looks so cool!";

And the search term was "boring"... the script would then display the result as:

...ch is so boring but...

Get it so far?

Ok I got all that to work, even with things that could screw it up like weird characters, lowercase and all that. Now the thing is, how can I get it to isolate for multiple keywords?

Is it possible? I attached the code I have so far, you can test it and try and help me if you can. All help is appreciated. Thanks.

rehfeld

how do you want it to "isolate" multiple keywords?

the most simplisitc would be to repeat your procedure for each search term, for example using a loop. if you want to group them and stuff, you are going to need to decide upon some strict rules of how it should work.

some functions you would will likely find usefull:

http://php.net/metaphone
http://php.net/soundex
http://php.net/levenshtein
http://php.net/similar_text

make sure you read the user comments, theres some pretty powerful code snippits in there.

and when your done reading all that, you will have even more respect for google than ever 🙂

Sharif

I don't understand how I'd use a loop you know, since the result would highlight all the matching words in ONE block...

rehfeld


$search = 'foo bar baz';

$text = 'i am some foo text';


$words = explode(' ', $search);


foreach ($words as $word) {
    $text = highlight_word($word, $text);
}

echo $text;

Sharif

LOL it doesn't work !!!

rehfeld

and "doesnt work" means what?

rehfeld

oh, i just looked at your code, i didnt know you were cutting the string, i thought just highlighting.

in that case

$result = '';

foreach ($words as $word) {
    $result .= highlight_word($word, $text);
}

Sharif

I don't think you understood what the script is supposed to do. Did you test the code you posted?

Lars

I assume it supposed to be

highlight_string

(IMHO, don't LOL me if I am wrong please).

http://my.php.net/manual/en/function.highlight-string.php

Sharif

Yes I've tried with highlight_string as well but it will not work. Basically if you were searching a big blob of text with two search terms, i'd like to highlight those search terms in the text but not show teh entire blob... just cutoffs.

Sharif

I figured I post a good example:

In google I typed in "microsoft midtown"... this is one of the results:

... Note: Microsoft DirectX® 7.0 or later is required to play Midtown Madness on Zone.com . Midtown Madness Trial Version. Download the Trial Version now! ...

rehfeld

Originally posted by Sharif
I don't think you understood what the script is supposed to do. Did you test the code you posted?

i realize you dont use a function called highlight_word

im implying for you to use your code there. im not getting specific, because you arent either.

again, if you want your results grouped together, and you just showed that you do want this by posting that example, you will need to decide on a strict set of rules of how you desire your code to behave.

what if they searched for 2 words, and 1 word was at the beggining of a long paragraph, yet the other was farther down the page?
should you highlight just 1 word? should you highlight both, showing ALL text in between(possibly pages of irrelevant text)?
should you break the text up and just show the relevant chunks around each word?
what if 2 word matches are next to each other? should you chunk this or not?

maybe this should depend on how much other text is between the 2 words?
if so, how should it depend?

i could think of a ton of these questions.....but YOU need to decide them, not us.

and then you said this

"i'd like to highlight those search terms in the text but not show teh entire blob... just cutoffs."

i dont see any "cutoffs" in your example, i see a single cutoff. there is a very large difference. again, decide upon a behavior you desire.

if this behavior is acceptable

... Note: Microsoft DirectX® 7.0 ........to play Midtown Madness......

then take a closer look at the last way i looped, as thats essentially what it will produce.

eriksmoen

$string = eregi_replace($query, "<B>$query</B>",$string);

if multiple words then:

$words = explode(" ",$query);
for($a = 0; $a < sizeof($words); $a++){
$string = eregi_replace($words[$a],"<B>".$words[$a]."</B>,$string);
}

hope that helps you a bit.