Searching a text file

chuckury

Lets say I have a text file ("search.txt") and in it I have something like this:

<a href="index.php?page=Pseudo">Creating Pseudo Frames</a><br>Tutorial on creating "pseudo" frames in php.

All on one line of course. But alot of lines like this a new line everytime with a different link and description. How do I search the text file then print all the closest results on the webpage. so basically a search engine using a .txt file (cant use mysql server doesnt have it)

dalecosp

Well, [man]file/man would read the file into an array. Then, [man]in_array/man might get you an exact match.

For getting "close" to an item, you might look at [man]soundex/man, [man]levenshtein/man, [man]metaphone/man and such-like.

Another approach might be [man]ereg/man or perhaps [man]preg_match[/man], if you'd rather use regular expressions of one type or another.

Lastly, I have seen systems that break apart each line in pieces (like [man]explode/man ), and then simply use logical operations to attempt a match.

Then, there's the problem of storage of the result, but that should be a relatively small issue in comparison.

HTH,

chuckury

Umm :o Could please explain how to use those Im a newb.

Merve

I'll explain what levenshtein() is.

The Leventshtein distance between two strings is the minimum number of characters you'd have to change, add, delete, or move to change one string into the other. levenshtein() calculates this distance.

So for example, you could use a loop to compare the search to each link, and say if the levenshtein distance is bigger than for example, 5 (I'm just giving a pretty random number here...you'll want to test your script to figure out which number works best), then it is not a match. If it is smaller than 5, it is a match.

Also, you take the Levenshtein distance of the metaphone key of each word, but this doesn't work well with phrases.

For example, the metaphone key of hello is HL. This is also the metaphone key of hell and hill. What you could do is calculate the length of the user input string, or metaphone key - which you can put in a string using strlen(), then if the length of the string is smaller, then it must be an identical match, but if the length of the metaphone key is larger, then use the Levenshtein distance trick.

Also, if a user inputs something with spaces, explode the string with explode(), get the metaphone keys of each element of the array with metaphone() and put the metaphone keys together with join() or implode(). And then use the Levensthein trick.

The possiblities are endless with PHP!

chuckury

Thank you for you're reply I will look into the leveshtein() later first I just want to get it to work using the in_array() then get more complicted why doesnt this work I dont get an error but whenever I type exact words that are contained in the text file it still gives me no matches found.

<?
$searchdb = file ('search.txt');
   $_REQUEST['search'];
   import_request_variables('p', 'p_');
if (in_array($p_search, $searchdb, TRUE)) {
    print "Got Irix";
} else {
    print "No Matches Found!";
}
?>

Check it out here http://www.peerideas.com/htmlfunction/search.php
("go" without the quotes should work, but it doesnt)

dalecosp

I imagine some error checking would help you figure out why it's not working. My code might be something like:

<?
$file="somefile.txt";
if (file_exists($file)) {
   $fp=fopen($file, 'r');
   if ($fp) {
      $data=file($file);
   }
   else {
      echo "Unable to open file for reading";
   }
} else {
   echo "File does not exist!";
}
$x=0;
$search_term=$_POST['search'];
if ($search_term) {
   while ($x<count($data)) {
      $line=$data($x);
      if ($search_term==$line) {
         echo "Match found!";
         $match=$line;
         exit;
      }
      else {
         $x++;
      }
   }
} else {
   echo "No search term!";
}