regular expression help newbie

beckstei

I'm trying to match words in a sentence to entries in a dictionary database, so that words in the sentence that match a word in the dictionary are turned into links. This means matching words in the sentence that are plural, or past tense to words in the dictionary that are not plural or past tense as well as dealing with words that are immediately followed by a comma or period.

Suppose I have the following sentence:

The planets revolve around the sun primarily in the plane of the solar system, as seen in the planetarium demonstration.

where I have planet, plane and planetarium in the database.

entry1=planet;
entry2=plane;
entry3=planetarium;

I want a script that does the following;

for($i=0;$i=$end_of_sentence;$i++){
for($j=0;$j=$num_words_in_dictionary,$j++){
if ($word_in_sentence[$i] == $entry[$j]){
$word_in_sentence[$i] =

'<a href="link_to_definition">'.$word_in_sentence.'</a>';
}
}

I've been splitting the sentence on spaces and looping through the sentence one word at a time. I've tried various forms of if(eregi( )), but I always end up matching plane, i.e., <a href="..">plane</a>t, or <a href="..">plane</a>tarium, but i can't seem to match planet or planetarium. If I go with
if($word_in_sentence == $entry) I don't know how to deal with plurals, past tense, commas or periods or whatever on the end of the word. Can someone give me a push in the right direction?
Thanks
B

Anon

Use perl regular expressions using the function preg_replace(). You can look it up at http://www.php.net/preg_replace
and for a good explanation on how to use them try the perlre man page at http://www.aisd.com/technology/perl/man/perlre.shtml
You basically want to use word boundaries \b to match the word exactly. For instance,
in the string:
$string = "Cats taste like chicken when cooked properly";
$regexp = "#\b(Cats)\b#si";
$replace = "<a href="$pathtolink/\1.html">\1</a>";
$string = preg_replace ( $regexp, $replace, $string);
Will match every instance of that word and replace it with a link defined by $pathtolink and given that the definition will be (the matching word).html it will link there. However, the above does NOT match the following string:
$string = "Catfood is something my cat does not like."
It is because it is matching the word "cats" in that regexp and nothing else. The word boundary in there forbids anything other than cats and will not match scaredycat, catfood, or even catsup. It will match the word "cats" and only the word cats.
I think...

Anon

One other thing, I'm sure that if you tweak the regexp I showed you, then you can get to where you should be going. The problem lies in pluralizations and whether or not to even try to link them. I mean, when you get to certain words like platypus to platypi, things can get hairy and I'm sure as @!#$ that you don't want to hand edit each and every word. Just make it a point to only link the words that you need. Dealing with commas, periods, ect... should be dealt with by using the word boundary (\b) like I had mentioned earlier.
However, you can be sure of it by making that part of your regexp.
Say you have a list of all the things you want replaced in an array.
$wordarray = ("dog", "cat", "groundhog", "frog");
$wordlist = implode("|",$wordarray); //probably not the best way to do this...
$string = "Sometimes I have my pet groundhog fight to the death with my neighbors litter of baby cats".
string = "Cats taste like chicken when cooked properly";
$regexp = "#\b($wordlist)\b#si";
$replace = "<a href="$pathtolink/\1.html">\1</a>";
$string = preg_replace($regexp, $replace, $string);
There, that is more help than you deserve. @!#$! No one's ever helped me out on my damn questions.

Anon

Thanks big bunches. hugs and kisses to your cats.
B