location_parser.inc.php

<?php
/*
	Location Parser v1.01, 2005-04-17
	Copyright 2005 Michael Boone
	[email protected] (remove the X)
	http://boonedocks.net/

	Provided under the GNU General Public License
	Contact me for use outside the bounds of that license

	---------------------------------------------------------------	
	This program is free software; you can redistribute it and/or
	modify it under the terms of the GNU General Public License
	as published by the Free Software Foundation; either version 2
	of the License, or (at your option) any later version.
	
	This program is distributed in the hope that it will be useful,
	but WITHOUT ANY WARRANTY; without even the implied warranty of
	MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
	GNU General Public License for more details.

	The GNU General Public License can be found at:
	http://www.gnu.org/copyleft/gpl.html
	---------------------------------------------------------------	
	
*/
	
function findLocation($strLocation) {
	// given a free-form location string, determine two character state/province code and return it
	// currently United States and Canada only
	// returns an empty string if a location cannot be determined

	// define state/province array as static so we only define it once
	// generally I put these in alphabetical order, but I have moved certain states/provinces with
	// variations that may appear in other words to the end
	// final period is omitted from abbreviations since the period is a regex word boundary
	// periods in the middle like N.C. get kept, i.e. N.C. becomes N.C
	static $aryStates=array(
		array(abr=>'AL',poss=>'Alabama;AL;Ala'),
		array(abr=>'AK',poss=>'Alaska;AK'),
		array(abr=>'AZ',poss=>'Arizona;AZ;Ariz'),
		array(abr=>'AR',poss=>'Arkansas;AR;Ark'),
		array(abr=>'CA',poss=>'California;CA;Cal.;Calif'),
		array(abr=>'CT',poss=>'Connecticut;CT;Conn'),
		array(abr=>'DC',poss=>'District of Columbia;DC;D.C'),
		array(abr=>'FL',poss=>'Florida;FL;Fla'),
		array(abr=>'GA',poss=>'Georgia;GA'),
		array(abr=>'HI',poss=>'Hawaii;HI'),
		array(abr=>'ID',poss=>'Idaho;ID'),
		array(abr=>'IL',poss=>'Illinois;IL;Ill'),
		array(abr=>'IA',poss=>'Iowa;IA'),
		array(abr=>'KS',poss=>'Kansas;KS;Kan'),
		array(abr=>'KY',poss=>'Kentucky;KY'),
		array(abr=>'ME',poss=>'Maine;ME'),
		array(abr=>'MD',poss=>'Maryland;MD'),
		array(abr=>'MA',poss=>'Massachusetts;MA;Mass'),
		array(abr=>'MI',poss=>'Michigan;MI;Mich'),
		array(abr=>'MN',poss=>'Minnesota;MN;Minn'),
		array(abr=>'MS',poss=>'Mississippi;MS;Miss'),
		array(abr=>'MO',poss=>'Missouri;MO'),
		array(abr=>'MT',poss=>'Montana;MT;Mont'),
		array(abr=>'NE',poss=>'Nebraska;NE;Neb'),
		array(abr=>'NV',poss=>'Nevada;NV;Nev'),
		array(abr=>'NH',poss=>'New Hampshire;NH;N.H'),
		array(abr=>'NJ',poss=>'New Jersey;NJ;N.J'),
		array(abr=>'NM',poss=>'New Mexico;NM;N.M'),
		array(abr=>'NY',poss=>'New York;NY;N.Y'),
		array(abr=>'NC',poss=>'North Carolina;NC;N. Carolina;N.C'),
		array(abr=>'ND',poss=>'North Dakota;ND;N.D'),
		array(abr=>'OH',poss=>'Ohio;OH'),
		array(abr=>'OK',poss=>'Oklahoma;OK;Okla'),
		array(abr=>'OR',poss=>'Oregon;OR;Ore'),
		array(abr=>'PA',poss=>'Pennsylvania;PA;Penn'),
		array(abr=>'RI',poss=>'Rhode Island;RI;R.I'),
		array(abr=>'SC',poss=>'South Carolina;SC;S. Carolina;S.C'),
		array(abr=>'SD',poss=>'South Dakota;SD;S. Dakota;S.D'),
		array(abr=>'TN',poss=>'Tennessee;TN;Tenn'),
		array(abr=>'TX',poss=>'Texas;TX;Tex'),
		array(abr=>'UT',poss=>'Utah;UT'),
		array(abr=>'VT',poss=>'Vermont;VT'),
		array(abr=>'WA',poss=>'Washington;WA;Wash'),
		array(abr=>'WV',poss=>'West Virginia;WV;W. Virginia;W. Va'),
		array(abr=>'VA',poss=>'Virginia;VA'),
		array(abr=>'WI',poss=>'Wisconsin;WI;Wis.;Wisc'),
		array(abr=>'WY',poss=>'Wyoming;WY;Wyo'),
		array(abr=>'AB',poss=>'Alberta;AB;Alb'),
		array(abr=>'BC',poss=>'British Columbia;BC;B.C'),
		array(abr=>'MB',poss=>'Manitoba;MB'),
		array(abr=>'NB',poss=>'New Brunswick;NB'),
		array(abr=>'NL',poss=>'Newfoundland;Labrador;NL'),
		array(abr=>'NT',poss=>'Northwest Territories;NT'),
		array(abr=>'NS',poss=>'Nova Scotia;NS'),
		array(abr=>'PE',poss=>'Prince Edward Island;PE'),
		array(abr=>'QC',poss=>'Quebec;QC;PQ'),
		array(abr=>'SK',poss=>'Saskatchewan;SK'),
		array(abr=>'YT',poss=>'Yukon Territory;YT'),
		array(abr=>'NU',poss=>'Nunavut;NU'),
		array(abr=>'CO',poss=>'Colorado;CO;Colo'),
		array(abr=>'ON',poss=>'Ontario;ON;Ont'),
		array(abr=>'IN',poss=>'Indiana;IN;Ind'),
		array(abr=>'LA',poss=>'Louisiana;LA'),
		array(abr=>'DE',poss=>'Delaware;DE;Del')
		);
	
	$strReturn='';
	$strLocation=trim($strLocation);
	
	if ($strLocation!='') {
		foreach ($aryStates as $ary1State) {
			$aryTest=explode(';',$ary1State['poss']);
			
			foreach ($aryTest as $strTest) {
				// if a match, return the abbreviation and exit out of the regex testing
				$strTest=str_replace('.','\.',$strTest);
				if (preg_match('/^.*\b' . $strTest . '\b.*$/i', $strLocation)) {
					$strReturn=$ary1State['abr'];
					break 2;
				}
			}
		}
	}
	
	return $strReturn;
}

?>