// simplepie makes the rss parsing very easy.
include_once( "libs/simplepie.inc" );
$keywords=
    array(
                    ‘fsbo’=>"Hey come check out great prices on real estate auctions: http://www.fsboauction.info"
);

$username = ‘fsbohomes’;
$password = ‘somepassword’;

// read the public timeline. Note that this is cached for 60 seconds, so asking for it more often
// than that is silly. Also notice that it won’t catch _everything_ but you should get at least one a day.

$read=‘http://twitter.com/statuses/public_timeline.rss’;
$feed=new SimplePie( );
$feed->set_feed_url( $read );
$feed->set_cache_duration( 60 );
$feed->init();
$feed->handle_content_type();

$count=0;
foreach( $feed->get_items() as $item ){
// read the tweet.
                $tweet=$item->get_title();
// split it into username and the contents.
// we’ll be comparing our keywords against the contents and
// sending the username an @ message.
                list( $a, $t)=split(‘:’, $tweet, 2);
// I had a lot of multibyte-encoded lines but didn’t feel like parsing them.
// since I don’t speak the languages anyway.
                $e= mb_detect_encoding( $t );
                if ( $e == ‘ASCII’ ) {
// check each keyword
                    foreach( $keywords as $seek=>$message ){
                        $count++;
                        if ( strpos( $t, $seek ) !== false ) {
// if the keyword is found, send a message.
                                    $msg = ‘@’.$a.‘ ‘.$message;
                                    $msg = substr( $msg, 0, 140 );
                                    sendTweet( $msg );
                        }
                    }
                }
}
// http://twitter.com/statuses/public_timeline.format
// formats: xml, json, rss, atom

// this is from the twitter bot post referenced above
// it’s just a function for posting a twitter.
function sendTweet($msg){
                global $username;
                global $password;

    $url = ‘http://twitter.com/statuses/update.xml’;

    $curl_handle = curl_init();

    curl_setopt($curl_handle, CURLOPT_URL, $url);
    curl_setopt($curl_handle, CURLOPT_CONNECTTIMEOUT, 2);
    curl_setopt($curl_handle, CURLOPT_RETURNTRANSFER, 1);
    curl_setopt($curl_handle, CURLOPT_POST, 1);
    curl_setopt($curl_handle, CURLOPT_POSTFIELDS, "status=$msg");
    curl_setopt($curl_handle, CURLOPT_USERPWD, "$username:$password");
    $buffer = curl_exec($curl_handle);
    curl_close($curl_handle);
    if (empty($buffer)) {
        echo "failed\n";
    }
    else {
        echo "succeeded\n";
    }
}
 

// how many words is our goal?
$wordcount=1000;

// read the input files. Really this should be a glob or a readdir() loop.
$i[]=file(‘source/after-you-pay-off-credit-card-debt.txt’);
$i[]=file(‘source/agency-card-credit-debt-settlement.txt’);
$i[]=file(‘source/a-problem-called-credit-card-debt.txt’);
/*
$i[]=file(‘source/bad-debt-credit-card-what-is-that.txt’);
$i[]=file(‘source/before-you-go-for-credit-card-debt-help.txt’);
$i[]=file(‘source/blogging-consolidation-debt-and-new-information-technology-239.txt’);
$i[]=file(‘source/card-com-credit-debt-en-language-site.txt’);
$i[]=file(‘source/college-student-credit-card-debt.txt’);
$i[]=file(‘source/consolidate-credit-card-debt.txt’);
$i[]=file(‘source/consolidate-your-credit-card-debt.txt’);
$i[]=file(‘source/credit-card-debt-consolidation-loan.txt’);
$i[]=file(‘source/credit-card-debt-consolidation.txt’);
$i[]=file(‘source/credit-card-debt-counseling.txt’);
$i[]=file(‘source/credit-card-debt-management.txt’);
$i[]=file(‘source/credit-card-debt-negotiation.txt’);
$i[]=file(‘source/credit-card-debt-reduction.txt’);
$i[]=file(‘source/credit-card-debt-relief.txt’);
$i[]=file(‘source/credit-card-debt-settlement.txt’);
$i[]=file(‘source/credit-card-debt.txt’);
$i[]=file(‘source/creditcarddebt.txt’);
$i[]=file(‘source/eliminate-credit-card-debt.txt’);
$i[]=file(‘source/excessive-credit-card-debt.txt’);
$i[]=file(‘source/get-out-of-credit-card-debt.txt’);
$i[]=file(‘source/is-consolidating-credit-card-debt-a-good-option.txt’);
$i[]=file(‘source/reduce-credit-card-debt.txt’);
$i[]=file(‘source/re-financing-to-consolidate-debt.txt’);
$i[]=file(‘source/taking-a-step-towards-credit-card-debt-elimination.txt’);
$i[]=file(‘source/teen-credit-card-debt-statistics.txt’);
$i[]=file(‘source/the-benefits-from-credit-card-debt-consolodation.txt’);
*/

// some print output to mark my spot while I troubleshoot.
print count( $i )." files loaded\n";
// put them all together.
for ($j=0;$j<count($i);$j++ ){
    $input=array_merge($input, $i[$j] );
}
print "merged\n";

// clean up the input. We’re winding up with just a list of words separated by ONE space.
$input=implode(" ", $input );
$input=preg_replace(‘|[^a-zA-Z0-9 ]|’, ‘ ‘, $input );
// this next line cleans out all double spaces.
while( strpos( $input, ‘  ’) !== false ) { $input = str_replace( ‘  ’, ‘ ‘, $input ); }

$inputstring=$input;
$holder=explode(‘ ‘, $input );
unset( $input );
// not sure why I had empty elements in the $input array but this is to remove all of them.
$input=array();
foreach( $holder as $k=>$v) {
    if ( strlen( trim( $v )) > 0 ) {
        $input[]=trim( $v );
    }
}
print "cleaned\n";

// this array is for the word pairs.
$segments=array();
$holder=array_shift( $input );
while( count( $input ) ) {
        // take the current word and the next word from the input array.
    $matchme=$holder.‘ ‘.$input[0];
    $pattern="|$matchme (.*?) |";
        // get all the matches from input- the long string of words not the array
    preg_match_all( $pattern, $inputstring, $matches );
        // add the list of matched words to the segments array
    $segments[ $matchme ] = $matches[1];
        // cut the first item from the input array and move it over one.
        // note that the array_shift compares against the while strlen above, so the shift is important
    $holder=array_shift( $input );
}
// just some output
print count($segments)." segments calculated\n";
$keys=array_keys( $segments );

// starting to assemble the output. Using $keys[0] to be the first bit of the markoved content.
$output[]=$keys[0];
// $bracket is reused each loop, but $output holds the … wait for it … output.
$bracket=$output[0];
print "assembling\n";
// yeah yeah compare the length of output against the desired length.
while( count( $output )  < $wordcount ){
        // $bracket is the key for $segments.
    $index=$bracket;
    $possiblenext=$segments[$index];
        // $possiblenext is the array of words-that-follow ‘bracket’.
    if ( count( $possiblenext ) == 0 ) {
                // this is where it’s dying oddly
        print "no possible next for $bracket\n";
        die("\n\naiee!!\n");
    }
    $rand=rand(0, count( $possiblenext )-1 );
        // pick a random number out of the array. I had this all wrapped together but broke it apart when it wasn’t working
    $nextword=$possiblenext[$rand];
    $output[] = ‘ ‘.$nextword; // prepend a space so the words don’t mash together
        // the next bracket is the second word of this group plus the nextword variable.
    $holder=split( " ", $bracket );
        // makes me sad that I can’t use split(‘ ‘, $bracket)[1];
    $nextbracket=$holder[1].‘ ‘.$nextword;
    $bracket = $nextbracket;
}
$output = implode( ‘ ‘, $output );
print $output;
print "\n";