/*
 * Fun filters by Dougal <URL:http://dougal.gunters.org/>
 *
 * This started when I added a pirate filter to my blog for Talk Like
 * a Pirate Day (talklikeapirate.com). My first version was easier
 * than I expected it to be (though it had flaws), which inspired me
 * to locate and convert some of Kalsey's MovableJive filters.
 * 
 * The earliest version had a flaw, in that it would filter text inside of
 * HTML tags, causing it to mangle links and such. I fixed this by borrowing
 * an idea from Simon Willison. Simon's use of a callback function to only
 * match text that was not part of a tag was good, but it included the '>'
 * and '<' brackets from surrounding tags, requiring you to hack them back
 * in at the end of your content filter.  
 * 
 * After an afternoon studying the pcre pattern syntax and wrestling regexps
 * with the help of the Regex Coach (http://weitz.de/regex-coach/) I came up
 * with an improved pattern, which doesn't require us to tack the '>' and
 * '<' back on manually. Cool, huh?
 */

// An idea from Simon Willison (http://simon.incutio.com/)
// This tries to make sure that we only filter text *between*
// any HTML tags, and not *within* them.
function filter_cdata_content($content, $filter='none') {
	if (function_exists($filter)) {
		$content = preg_replace_callback('/(?(?<=>)|\A)([^<>]+)(?(?=<)|\Z)/s', $filter, $content);
	}

	return $content;
}

// This function takes an array of ('/pattern/' => 'replacement') pairs
// and applies them all to $content.
function array_apply_regexp($patterns,$content) {
	// Extract the values:
	$keys = array_keys($patterns);
	$values = array_values($patterns);
	
	// Replace the words:
	$content = preg_replace($keys,$values,$content);

	return $content;
}

// For Talk Like a Pirate Day
function pirate($content) {
	return filter_cdata_content($content,'pirate_filter');
}

function pirate_filter($content) {
	$content = $content[1];

	$patterns = array( 
			'%\bmy\b%' => 'me',
			'%\bboss\b%' => 'admiral',
			'%\bmanager\b%' => 'admiral',
			'%\b[Cc]aptain\b%' => "Cap'n",
			'%\bmyself\b%' => 'meself',
			'%\byour\b%' => 'yer',
			'%\byou\b%' => 'ye',
			'%\bfriend\b%' => 'matey',
			'%\bfriends\b%' => 'maties',
			'%\bco[-]?worker\b%' => 'shipmate',
			'%\bco[-]?workers\b%' => 'shipmates',
			'%\bearlier\b%' => 'afore',
			'%\bold\b%' => 'auld',
			'%\bthe\b%' => "th'",
			'%\bof\b%' =>  "o'",
			"%\bdon't\b%" => "dern't",
			'%\bdo not\b%' => "dern't",
			'%\bnever\b%' => "ne'er",
			'%\bever\b%' => "e'er",
			'%\bover\b%' => "o'er",
			'%\bYes\b%' => 'Aye',
			'%\bNo\b%' => 'Nay',
			"%\bdon't know\b%" => "dinna",
			"%\bhadn't\b%" => "ha'nae",
			"%\bdidn't\b%"=>  "di'nae",
			"%\bwasn't\b%" => "weren't",
			"%\bhaven't\b%" => "ha'nae",
			'%\bfor\b%' => 'fer',
			'%\bbetween\b%' => 'betwixt',
			'%\baround\b%' => "aroun'",
			'%\bto\b%' => "t'",
			"%\bit's\b%" => "'tis",
			'%\bwoman\b%' => 'wench',
			'%\blady\b%' => 'wench',
			'%\bwife\b%' => 'lady',
			'%\bgirl\b%' => 'lass',
			'%\bgirls\b%' => 'lassies',
			'%\bguy\b%' => 'lubber',
			'%\bman\b%' => 'lubber',
			'%\bfellow\b%' => 'lubber',
			'%\bdude\b%' => 'lubber',
			'%\bboy\b%' => 'lad',
			'%\bboys\b%' => 'laddies',
			'%\bchildren\b%' => 'little sandcrabs',
			'%\bkids\b%' => 'minnows',
			'%\bhim\b%' => 'that scurvey dog',
			'%\bher\b%' => 'that comely wench',
			'%\bhim\.\b%' => 'that drunken sailor',
			'%\bHe\b%' => 'The ornery cuss',
			'%\bShe\b%' => 'The winsome lass',
			"%\bhe's\b%" => 'he be',
			"%\bshe's\b%" => 'she be',
			'%\bwas\b%' => "were bein'",
			'%\bHey\b%' => 'Avast',
			'%\bher\.\b%' => 'that lovely lass',
			'%\bfood\b%' => 'chow',
			'%\broad\b%' => 'sea',
			'%\broads\b%' => 'seas',
			'%\bstreet\b%' => 'river',
			'%\bstreets\b%' => 'rivers',
			'%\bhighway\b%' => 'ocean',
			'%\bhighways\b%' => 'oceans',
			'%\bcar\b%' => 'boat',
			'%\bcars\b%' => 'boats',
			'%\btruck\b%' => 'schooner',
			'%\btrucks\b%' => 'schooners',
			'%\bSUV\b%' => 'ship',
			'%\bairplane\b%' => 'flying machine',
			'%\bjet\b%' => 'flying machine',
			'%\bmachine\b%' => 'contraption',
			'%\bdriving\b%' => 'sailing',
			'%\bdrive\b%' => 'sail',
			'/ing\b/' => "in'",
			'/ings\b/' => "in's",
			// These next two do cool random substitutions
			'/(\.\s)/e' => 'avast("$0",3)',
			'/([!\?]\s)/e' => 'avast("$0",2)', // Greater chance after exclamation
			);

	// Replace the words:
	$content = array_apply_regexp($patterns,$content);
	
	return $content;
}

// support function for pirate()
// this could probably be refactored to make it more generic, allowing
// different filters to pass their own patterns in.
function avast($stub = '',$chance = 5) {
	$shouts = array(
				", avast$stub",
				"$stub Ahoy!",
				", and a bottle of rum!",
				", by Blackbeard's sword$stub",
				", by Davy Jones' locker$stub",
				"$stub Walk the plank!",
				"$stub Aarrr!",
				"$stub Yaaarrrrr!",
				", pass the grog!",
				", and dinna spare the whip!",
				", with a chest full of booty$stub",
				", and a bucket o' chum$stub",
				", we'll keel-haul ye!",
				"$stub Shiver me timbers!",
				"$stub And hoist the mainsail!",
				"$stub And swab the deck!",
				", ye scurvey dog$stub",
				"$stub Fire the cannons!",
				", to be sure$stub",
				", I'll warrant ye$stub",
				);
				
	shuffle($shouts);
	
	return (((1 == rand(1,$chance))?array_shift($shouts):$stub) . ' ');
}

// Borrowed fudd, jive, kraut, and chef from Kalsey's MovableJive filter
//   http://kalsey.com/blog/2003/02/movablejive/

// These filters (particularly 'kraut' and 'jive' might contain something
// that could offend you or your readers. Edit and/or use at your own risk.

function fudd($content) {
	return filter_cdata_content($content,'fudd_filter');
}

function fudd_filter($content) {
	$content = $content[1];

	$patterns = array(
			'%(r|l)%' => 'w',
			'%qu%' => 'qw',
			'%th(\s)%' => 'f$1',
			'%th%' => 'd',
			'%n\.%' => 'n, uh-hah-ha-ha.',
			'%(R|L)%' => 'W',
			'%(Qu|QW)%' => 'QW',
			'%TH(\s)%' => 'F$1',
			'%Th%' => 'D',
			'%N\.%' => 'N, uh-hah-hah-hah.'
			);

	$content = array_apply_regexp($patterns,$content);

	return $content;
}

function kraut($content) {
	return filter_cdata_content($content,'kraut_filter');
}

function kraut_filter($content) {
	$content = $content[1];

	$patterns = array(
	'%ing%' => 'ingkt',
	'% the %' => ' ze ',
	'%The %' => 'Ze ',
	'% with %' => ' mitt ',
	'%With %' => 'Mitt ',
	'% is%' => ' iss',
	'% Is%' => ' Iss',
	'%wr%' => 'w-r-r',
	'%Wr%' => 'W-r-r',
	'%R%' => 'R-r-r',
	'% r%' => ' r-r-r',
	'%Yes( |\.|!)%' => 'Jawohl$1',
	'%YES!%' => 'JAWOHL!',
	'% yes( |.|!)%' => ' ja$1',
	'%No( |!|\?) %' => 'Nein$1',
	'% no( |\.|!|\?)%' => ' nein$1',
	'% not%' => ' nicht',
	'%Not%' => 'Nicht',
	'%[Mm]r.%' => 'Herr',
	'%[Mm]rs.%' => 'Frau',
	'%Miss%' => 'Fraulein',
	'% of %' => ' uff ',
	'%Of %' => 'Uff ',
	'%my%' => 'mein',
	'%My%' => 'Mein',
	'% and %' => ' undt ',
	'%And %' => 'Undt ',
	'%and%' => 'ent',
	'%One %' => 'Ein ',
	'% one%' => ' ein',
	'%Is %' => 'Ist ',
	'% is %' => ' ist ',
	'%ow %' => 'ow ',
	'%w %' => 'w ',
	'% sc%' => ' shc',
	'%Sc%' => 'Shc',
	'% st%' => ' sht',
	'%St%' => 'Sht',
	'%sh%' => 'sch',
	'%Sh%' => 'Sch',
	'%ch%' => 'ch',
	'%Ch%' => 'Ch',
	'% c%' => ' k',
	'% C%' => ' K',
	'% for %' => ' fur ',
	'%Have%' => 'Haf',
	'%have%' => 'haf',
	'%j%' => 'ch',
	'%J%' => 'Ch',
	'%Qu%' => 'Qv',
	'%qu%' => 'qv',
	'%rd%' => 'rt',
	'%v%' => 'f',
	'%V%' => 'F',
	'% w%' => ' v',
	'%W%' => 'V',
	'%ward%' => 'verrt',
	'%wh%' => 'v',
	'%Wh%' => 'V',
	'% th%' => ' z',
	'%Th%' => 'Z',
	'%th%' => 'zz',
	'%[Cc]offee%' => 'Kafe',
	'%[Tt]hank[s]*%' => 'Danke',
	'%[Jj]ohn%' => 'Johann',
	'%[Ww]illiam%' => 'Wilhelm',
	'%[Bb]rad%' => 'Wilhelm',
	'%[Gg]ary%' => 'Gerhardt',
	'%[Jj]on%' => 'Hansel',
	'%[a-f]!%' => '$1 Naturlich!',
	'%[p-z]!%' => '$1 Seig Heil!'
	);

	$content = array_apply_regexp($patterns,$content);

	return $content;
}

function jive($content) {
	return filter_cdata_content($content,'jive_filter');
}

function jive_filter($content) {
	$content = $content[1];

	$patterns = array(
	'%file%' => 'stash',
	'%send%' => "t'row",
	'%program%' => 'honky code',
	'%atlas%' => 'Isaac',
	'%unix%' => 'slow mo-fo',
	'%UNIX%' => 'that slow mo-fo',
	'%takes%' => "snatch'd",
	'%Mexican%' => 'wet-back',
	'%Italian%' => 'greaser',
	'%take%' => 'snatch',
	"%don't%" => "duzn't",
	'%jive%' => 'JIBE',
	'%fool%' => 'honkyfool',
	'%modem%' => 'doodad',
	'%e the %' => 'e da damn ',
	'%a the %' => 'a da damn ',
	'%t the %' => 't da damn ',
	'%d the %' => 'd da damn ',
	'% man %' => ' dude ',
	'%woman%' => 'mama',
	'%girl%' => 'goat',
	'%somethin%' => "sump'n",
	'% lie %' => ' honky jibe ',
	'%([a-b]\.)%' => '$1 Sheeeiit.',
	'%([e-f]\.)%' => '$1 What it is, Mama!',
	'%([i-j]\.)%' => "$1 Ya' know?",
	'%([m-n]\.)%' => "$1 'S coo', bro.",
	'%([q-r]\.)%' => '$1 Ah be baaad...',
	'%([u-v]\.)%' => '$1 Man!',
	'%([y-z]\.)%' => '$1 Slap mah fro!',
	'%Sure%' => "Sho' nuff",
	'%sure%' => "sho' nuff",
	'% get	%' => ' git',
	'%will have%' => "gots'ta",
	'%will%' => "gots'ta ",
	'%got to%' => "gots'ta",
	'%I am%' => "I's gots'ta be",
	'%am not %' => 'aint ',
	'%is not %' => 'aint ',
	'%are not %' => 'aint',
	'% are your	%' => " is yo' ",
	'% are you	%' => ' you is ',
	'% hat %' => ' fedora ',
	'% shoe %' => ' kicker ',
	"%haven't%" => 'aint',
	'%have to%' => "gots'ta",
	'%have%' => "gots'",
	'% has%' => " gots'ta",
	'%come over%' => 'mosey on down',
	'% come %' => ' mosey on down ',
	'%!%' => '.  Right On!',
	'%buy	%' => 'steal',
	'% car %' => ' wheels ',
	'%drive%' => 'roll',
	'% eat %' => ' feed da bud ',
	'% black %' => ' brother ',
	'% negro %' => ' brother',
	'%white%' => 'honky',
	'%nigger%' => ' gentleman ',
	'%nice%' => "supa' fine",
	'%person%' => "sucka'",
	'% thing%' => ' wahtahmellun',
	'%home%' => 'plantation',
	'%name%' => 'dojigger',
	'% path%' => ' alley',
	'%computer%' => 'clunker',
	'%or%' => "o'",
	'%killed%' => 'wasted',
	'%president%' => 'super-dude',
	'%prime minister%' => 'super honcho',
	'%injured%' => 'hosed',
	'%government%' => "guv'ment",
	'%knew%' => 'knowed',
	'%because%' => "a'cuz",
	'%Because%' => "A'cuz",
	'%your%' => "yo'",
	'%Your%' => "Yo'",
	'%four%' => 'foe',
	'%got%' => 'gots',
	"%aren't%" => "ain't",
	'%young%' => 'yung',
	'%you%' => "ya'",
	'%You%' => "You's",
	'%first%' => 'fust',
	'%police%' => 'honky pigs',
	'% string%' => " chittlin'",
	'% read%' => ' eyeball',
	'%write%' => 'scribble',
	'%th%' => 'd',
	'%Th%' => 'D',
	'%ing%' => "in'",
	'% a %' => ' some ',
	'% to\s+%' => " t'",
	'%tion%' => 'shun',
	'% almost %' => " mos' ",
	'% from%' => ' fum',
	'% because %' => " cuz' ",
	"%you're%" => 'youse',
	"%You're%" => 'Youse',
	'%alright%' => "coo'",
	'%okay%' => "coo'",
	'%er %' => "a' ",
	'%known%' => 'knode',
	'%want%' => "wants'",
	'%beat%' => "whup'",
	'%exp%' => "'sp",
	'%exs%' => "'s",
	'% exc%' => " 's",
	'% ex%' => " 'es",
	'%like%' => 'likes',
	'%did	%' => 'dun did',
	'%kind of%' => "kind'a",
	'%women%' => 'honky chicks',
	'% men %' => ' dudes ',
	'% mens %' => ' dudes ',
	'% man %' => ' dude ',
	'%woman%' => 'honky chick',
	'%dead%' => 'wasted',
	'%good%' => 'baaaad',
	'%open %' => 'jimmey ',
	'%opened %' => "jimmey'd ",
	'% very%' => ' real',
	'%per%' => "puh'",
	'%pera%' => "puh'",
	'%oar%' => "o'",
	'% can%' => ' kin',
	'%just %' => 'plum ',
	'%detroit%' => 'Mo-town',
	'%western electric%' => "da' cave",
	'% believe%' => " recon'",
	'%[Ii]ndianapolis%' => 'Nap-town',
	'% [Jj]ack%' => ' Buckwheat',
	'% [Bb]ob %' => " Liva' Lips ",
	'% [Pp]hil %' => ' dat fine soul ',
	'% [Mm]ark %' => ' Amos ',
	'%[Rr]ober%' => 'Leroy',
	'%[Ss]andy%' => 'dat fine femahnaine ladee',
	'%[Jj]ohn %' => "Raz'tus ",
	'% [Pp]aul%' => " Fuh'rina",
	'%[Rr]eagan%' => 'Kingfish',
	'%[Dd]avid%' => 'Issac',
	'%[Rr]onald%' => 'Rolo',
	'% [Jj]im %' => ' Bo-Jangles ',
	'% [Mm]ary%' => ' Snow Flake',
	'%[Ll]arry%' => 'Remus',
	'%[Jj]oe%' => "Massa' ",
	'%[Jj]oseph%' => "Massa' ",
	'%mohammed%' => "liva' lips",
	'%pontiff%' => "wiz'",
	'%pope%' => "wiz'",
	'%pravda%' => 'dat commie rag',
	'%broken%' => "bugger'd",
	'%strange %' => 'funky ',
	'%dance %' => 'boogy ',
	'% house%' => ' crib',
	'%ask	%' => "ax'",
	'% so %' => " so's ",
	'%head%' => "'haid",
	'%boss%' => 'main man',
	'%wife%' => 'mama',
	'%people%' => "sucka's",
	'%money%' => "bre'd",
	'%([a-z]):%' => '$1, dig dis:',
	'%amateur%' => "begina'",
	'%radio	%' => "transista'",
	'% of %' => ' uh ',
	'%what%' => 'whut',
	'%does%' => 'duz',
	'%was%' => 'wuz',
	'% were%' => ' wuz',
	'%understand it%' => 'dig it',
	'%understand%' => 'dig it',
	'% my%' => " mah'",
	'% [Ii] %' => " ah' ",
	'%meta%' => "meta-fuckin'",
	'%hair%' => 'fro',
	'%talk%' => 'rap',
	'%music%' => 'beat',
	'%basket%' => 'hoop',
	'%football%' => 'ball',
	'%friend%' => 'homey',
	'%school%' => 'farm',
	'%boss%' => 'Man',
	'%want to%' => 'wanna',
	'%wants to%' => "be hankerin' aftah",
	'%well%' => 'sheeit',
	'%Well%' => 'Sheeit',
	'%big%' => 'big-ass',
	'%bad%' => 'bad-ass',
	'%small%' => 'little-ass',
	'%sort of%' => 'radical',
	'% is %' => ' be ',
	'%water%' => 'booze',
	'%book%' => "scribblin'",
	'%magazine%' => 'issue of GQ',
	'%paper%' => 'sheet',
	'%\bup\b%' => 'down',
	'%\bdown\b%' => 'waaay down',
	'%break%' => 'boogie',
	'%Hi%' => "'Sup, dude",
	'%VAX%' => 'pink Cadillac',
	);

	$content = array_apply_regexp($patterns,$content);

	return $content;
}

function chef($content) {
	return filter_cdata_content($content,'chef_filter');
}

function chef_filter($content) {
	$content = $content[1];

	$patterns = array(

    '%an%' => 'un',
    '%An%' => 'Un',

    '%au%' => 'oo',
    '%Au%' => 'Oo',
    '%(\w)ew%' => '$1oo',
    '%(\w)ow%' => '$1oo',
    '%(\W)o%' => '$1oo',
    '%(\W)O%' => '$1Oo',
    '%(\w)u%' => '$1oo',
    '%(\w)U%' => '$1Oo',

    '%a(\w)%' => 'e$1',
    '%A(\w)%' => 'E$1',
    '%en(\W)%' => 'ee$1',

    '%(\w)e(\W)%' => '$1e-a$2',
    '%(\W)e%' => '$1i',
    '%(\W)E%' => '$1I',

    '%(\w)f%' => '$1ff',

    '%(\w)ir%' => '$1ur',

    '%([a-m])i%' => '$1ee',
    '%([A-M])i%' => '$1EE',

    '%(\w)o%' => '$1u',
    '%the%' => 'zee',
    '%The%' => 'Zee',
    '%th(\W)%' => 't$1',
    '%(\w)tion%' => '$1shun',
    '%v%' => 'f',
    '%V%' => 'F',
    '%w%' => 'v',
    '%W%' => 'V',


    '%f{2,}%' => 'ff',
    '%o{2,}%' => 'oo',
    '%e{2,}%' => 'ee',

    '%([\.!\?])\s*(</[^>]+>)?\s*$%' => '$1 Bork Bork Bork!$2',

	);

	$content = array_apply_regexp($patterns,$content);

	return $content;
}