Neverside

Outputting flat file search results

Ok, I am pretty sure this is possible, but I am very confused of how to go about doing this. I want to create a script that searches a flat file for a keyword, and if that keyword is within a line, have that line diplayed as a search result. It would have to work for multiple lines.

PHP:

<?php

//songs.txt

The Who - My Wife
Radiohead - Like Spinning Plates
Radiohead - Just
The Beatles - I Want You (Shes so Heavy)

?>

So if the user searched for 'radiohead' it would output:

Radiohead - Like spinning plates
Radiohead - Just

The search would have to be case insensitive as well.

All I have so far is this:

PHP:

<?

$fp = fopen('songs.txt','r');
$ul = fread($fp,filesize('songs.txt'));
fclose($fp);

// break up the file into lines
//
$bits = explode("\n",$ul);

// look for our arttist / song
if (in_array($song,$bits)) {
   echo $song . ' exists!';  //this is where it would output the lines where the song is found
} else {
   echo $song . ' not found.';
}

?>

Thanks in advance guys!

I assume the file is not so large so you could theoretically do something like:

PHP:

<?
$search_soundex = soundex( $search );
$fp = file( 'songs.txt' );

foreach( $fp as $lineNumber => $text ){
  $words = explode( " ", $text );
  foreach( $words as $iWord ) {
    if( trim( strtolower( $search ) ) == strtolower( $iWord ) || soundex( $iWord ) == $search_soundex )
      $found[] = $lineNumber;
 }
}
?>

Not the best and really untested but something like that should work. Note: I have never used the function soundex() before in an actual script but I have played around with it and it's really cool. You can also substitue similar_text() or whatever for soundex().

___________________

wow, thanks for the quick response. Can you sort of explain what each section does a little becasue I am totally lost. I can figure some stuff out but a few general explanation would really help...thanks again.

I think I have figured a few things out

PHP:

<?
$search_soundex = soundex( $search ); //$search is the keyword being searched for
$fp = file( 'songs.txt' );  //songs file

foreach( $fp as $lineNumber => $text ){ //explode the file into line numbers and store in $text variable
  $words = explode( " ", $text ); //seperate into each word
  foreach( $words as $iWord ) {//this is where i get lost
    if( trim( strtolower( $search ) ) == strtolower( $iWord ) || soundex( $iWord ) == $search_soundex )
      $found[] = $lineNumber;
 }
}
?>

$search is the search phrase.

file() reads an entire file into an array.

foreach() goes through an array element by element and $lineNumber is the current array element it's reading. $text is that array's value.

The next foreach goes through each word and if the word.

If a word is found that matches the search phrase or the word's soundex matches the search phrase soundex then it puts the line number into a "found" array.

You should then run an array_unique on the results to make sure there aren't duplicate line numbers.

You can view the results by doing something like:

PHP:

<?php

foreach( $found as $line ) {
  echo $fp[$line]."<br/>";
}

?>

___________________

Thanks, it's working jsut great.

It's working!! LMAO

I really didn't expect that. Try typing in some things that are _almost_ the same as your search phrase and see if it works. Like "radohead" instead of "radiohead". Tell me if it actually works.

___________________

ya, that sounex() function is the ****. This is working great! Thanks for the help.

I was messing with this... it's reasonably accurate, but mess with the percents up top to change boundries. It's a bit faster then the other one...

PHP:

<?php

define('FILE', 'file.txt');
define('SEARCH_TERM', 'ludo');
define('EXACT', false);
define('MATCH_PERCENT_SHORT', 15);
define('MATCH_PERCENT_LONG', 35);

$lines = file(FILE);

$short_search = explode('-', SEARCH_TERM);
$search['short'] = $short_search[0];
$search['long'] = str_replace('-', '', SEARCH_TERM);

foreach($lines as $number=>$content)
{
    similar_text($content,$search['short'], $short_percent);
    similar_text($content,$search['long'], $long_percent);
    $match = EXACT ? stristr($content, SEARCH_TERM) stristr($content, SEARCH_TERM) ? true : (($short_percent > MATCH_PERCENT_SHORT) || (($long_percent > MATCH_PERCENT_LONG) && ($short_percent > MATCH_PERCENT_SHORT-10))));
    if($match)
    {
        $matches[] = $content;
    }
}

if(isset($matches))
{
    print count($matches) . ' matches.<br />';
    foreach($matches as $match_number=>$item)
    {
        print ($match_number+1) . ': ' . $item . '<br />';
    }
}
else
{
    print 'no matches';
}
?>

___________________

http://www.philbrodeur.com - Expert PHP Development and Tutorials

Nice script, Phil. But perhaps you should use levenshtein() which is faster and also supports a greater complexity. soundex() is the fastest of the bunch but it performs a different task than similar_text() or levenshtein().

___________________