include/geocoding.inc.php

   1 <?php
   2 /***************************************************************************
   3  *  Copyright (C) 2003-2009 Polytechnique.org                              *
   4  *  http://opensource.polytechnique.org/                                   *
   5  *                                                                         *
   6  *  This program is free software; you can redistribute it and/or modify   *
   7  *  it under the terms of the GNU General Public License as published by   *
   8  *  the Free Software Foundation; either version 2 of the License, or      *
   9  *  (at your option) any later version.                                    *
  10  *                                                                         *
  11  *  This program is distributed in the hope that it will be useful,        *
  12  *  but WITHOUT ANY WARRANTY; without even the implied warranty of         *
  13  *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the          *
  14  *  GNU General Public License for more details.                           *
  15  *                                                                         *
  16  *  You should have received a copy of the GNU General Public License      *
  17  *  along with this program; if not, write to the Free Software            *
  18  *  Foundation, Inc.,                                                      *
  19  *  59 Temple Place, Suite 330, Boston, MA  02111-1307  USA                *
  20  ***************************************************************************/
  21
  22 // Interface for an address geocoder. It provides support for transforming a free
  23 // form address into a fully structured one.
  24 // TODO: define and use an Address object instead of a key-value map.
  25 abstract class Geocoder {
  26     // Geocodes @p the address, and returns the corresponding updated address.
  27     // Unknown key-value pairs available in the input map are retained as-is.
  28     abstract public function getGeocodedAddress(array $address);
  29
  30     // Updates geoloc_administrativeareas, geoloc_subadministrativeareas and
  31     // geoloc_localities databases with new geocoded data and returns the
  32     // corresponding id.
  33     static public function getAreaId(array &$address, $area)
  34     {
  35         static $databases = array(
  36             'administrativeArea'    => 'geoloc_administrativeareas',
  37             'subAdministrativeArea' => 'geoloc_subadministrativeareas',
  38             'locality'              => 'geoloc_localities',
  39             );
  40
  41         if (isset($address[$area . 'Name']) && isset($databases[$area])) {
  42             $res = XDB::query("SELECT  id
  43                                  FROM  " . $databases[$area] . "
  44                                 WHERE  name = {?}",
  45                               $address[$area . 'Name']);
  46             if ($res->numRows() == 0) {
  47                 $address[$area . 'Id'] = XDB::execute("INSERT INTO  " . $databases[$area] . " (name, country)
  48                                                             VALUES  ({?}, {?})",
  49                                                       $address[$area . 'Name'], $address['countryId']);
  50             } else {
  51                 $address[$area . 'Id'] = $res->fetchOneCell();
  52             }
  53         }
  54     }
  55 }
  56
  57 // Implementation of a Geocoder using the Google Maps API. Please refer to
  58 // the following links for details:
  59 // http://code.google.com/apis/maps/documentation/services.html#Geocoding
  60 // http://code.google.com/intl/en/apis/maps/documentation/geocoding/
  61 // http://code.google.com/apis/maps/documentation/reference.html#GGeoAddressAccuracy
  62 //
  63 // It requires the properties gmaps_key and gmaps_url to be defined in section
  64 // Geocoder in plat/al's configuration (platal.ini & platal.conf).
  65 class GMapsGeocoder extends Geocoder {
  66
  67     // Maximum number of Geocoding calls to the Google Maps API.
  68     const MAX_GMAPS_RPC_CALLS = 5;
  69
  70     public function getGeocodedAddress(array $address) {
  71         $address = $this->prepareAddress($address);
  72         $textAddress = $address['text'];
  73
  74         // Try to geocode the full address.
  75         if (($geocodedData = $this->getPlacemarkForAddress($textAddress))) {
  76             return $this->getUpdatedAddress($address, $geocodedData, null);
  77         }
  78
  79         // If the full geocoding failed, try to geocode only the final part of the address.
  80         // We start by geocoding everything but the first line, and continue until we get
  81         // a result. To respect the limit of GMaps calls, we ignore the first few lines
  82         // if there are too many address lines.
  83         $addressLines = explode("\n", $textAddress);
  84         $linesCount   = count($addressLines);
  85         for ($i = max(1, $linesCount - self::MAX_GMAPS_RPC_CALLS + 1); $i < $linesCount; ++$i) {
  86             $extraLines = implode("\n", array_slice($addressLines, 0, $i));
  87             $toGeocode  = implode("\n", array_slice($addressLines, $i));
  88             if (($geocodedData = $this->getPlacemarkForAddress($toGeocode))) {
  89                 return $this->getUpdatedAddress($address, $geocodedData, $extraLines);
  90             }
  91         }
  92
  93         // No geocoding could be done, the initial address is returned as-is.
  94         return $address;
  95     }
  96
  97     // Updates the address with the geocoded information from Google Maps. Also
  98     // cleans up the final informations.
  99     private function getUpdatedAddress(array $address, array $geocodedData, $extraLines) {
 100         $this->fillAddressWithGeocoding(&$address, $geocodedData);
 101
 102         // If the accuracy is 6, it means only the street has been gecoded
 103         // but not the number, thus we need to fix it.
 104         if ($address['accuracy'] == 6) {
 105             $this->fixStreetNumber($address);
 106         }
 107
 108         // We can now format the address.
 109         $this->formatAddress($address, $extraLines);
 110
 111         // Some entities in ISO 3166 are not countries, thus they have to be replaced
 112         // by the country they belong to.
 113         // TODO: fixCountry($address);
 114
 115         return $address;
 116     }
 117
 118     // Retrieves the Placemark object (see #getPlacemarkFromJson()) for the @p
 119     // address, by querying the Google Maps API. Returns the array on success,
 120     // and null otherwise.
 121     private function getPlacemarkForAddress($address) {
 122         $url     = $this->getGeocodingUrl($address);
 123         $geoData = $this->getGeoJsonFromUrl($url);
 124
 125         return ($geoData ? $this->getPlacemarkFromJson($geoData) : null);
 126     }
 127
 128     // Prepares address to be geocoded
 129     private function prepareAddress($address) {
 130         $address['text'] = preg_replace('/\s*\n\s*/m', "\n", trim($address['text']));
 131         // TODO: $address['postalAddress'] = getPostalAddress($address['text']);
 132         $address['updateTime'] = time();
 133         unset($address['changed']);
 134         return $address;
 135     }
 136
 137     // Builds the Google Maps geocoder url to fetch information about @p address.
 138     // Returns the built url.
 139     private function getGeocodingUrl($address) {
 140         global $globals;
 141
 142         $parameters = array(
 143             'key'    => $globals->geocoder->gmaps_key,
 144             'sensor' => 'false',   // The queried address wasn't obtained from a GPS sensor.
 145             'hl'     => 'fr',      // Output langage.
 146             'oe'     => 'utf8',    // Output encoding.
 147             'output' => 'json',    // Output format.
 148             'gl'     => 'fr',      // Location preferences (addresses are in France by default).
 149             'q'      => $address,  // The queries address.
 150         );
 151
 152         return $globals->geocoder->gmaps_url . '?' . http_build_query($parameters);
 153     }
 154
 155     // Fetches JSON-encoded data from a Google Maps API url, and decode them.
 156     // Returns the json array on success, and null otherwise.
 157     private function getGeoJsonFromUrl($url) {
 158         global $globals;
 159
 160         // Prepare a backtrace object to log errors.
 161         $bt = null;
 162         if ($globals->debug & DEBUG_BT) {
 163             if (!isset(PlBacktrace::$bt['Geoloc'])) {
 164                 new PlBacktrace('Geoloc');
 165             }
 166             $bt = &PlBacktrace::$bt['Geoloc'];
 167             $bt->start($url);
 168         }
 169
 170         // Fetch the geocoding data.
 171         $rawData = file_get_contents($url);
 172         if (!$rawData) {
 173             if ($bt) {
 174                 $bt->stop(0, "Could not retrieve geocoded address from GoogleMaps.");
 175             }
 176             return null;
 177         }
 178
 179         // Decode the JSON-encoded data, and check for their validity.
 180         $data = json_decode($rawData, true);
 181         if ($bt) {
 182             $bt->stop(count($data), null, $data);
 183         }
 184
 185         return $data;
 186     }
 187
 188     // Extracts the most appropriate placemark from the JSON data fetched from
 189     // Google Maps. Returns a Placemark array on success, and null otherwise. See
 190     // http://code.google.com/apis/maps/documentation/services.html#Geocoding_Structured
 191     // for details on the Placemark structure.
 192     private function getPlacemarkFromJson(array $data) {
 193         // Check for geocoding failures.
 194         if (!isset($data['Status']['code']) || $data['Status']['code'] != 200) {
 195             // TODO: handle non-200 codes in a better way, since the code might
 196             // indicate a temporary error on Google's side.
 197             return null;
 198         }
 199
 200         // Check that at least one placemark was found.
 201         if (count($data['Placemark']) == 0) {
 202             return null;
 203         }
 204
 205         // Extract the placemark with the best accuracy. This is not always the
 206         // best result (since the same address may yield two different placemarks).
 207         $result = $data['Placemark'][0];
 208         foreach ($data['Placemark'] as $place) {
 209             if ($place['AddressDetails']['Accuracy'] > $result['AddressDetails']['Accuracy']) {
 210                 $result = $place;
 211             }
 212         }
 213
 214         return $result;
 215     }
 216
 217     // Fills the address with the geocoded data
 218     private function fillAddressWithGeocoding(&$address, $geocodedData) {
 219         // The geocoded address three is
 220         // Country -> AdministrativeArea -> SubAdministrativeArea -> Locality -> Thoroughfare
 221         // with all the possible shortcuts
 222         // The address is formatted as xAL, or eXtensible Address Language, an international
 223         // standard for address formatting.
 224         // xAL documentation: http://www.oasis-open.org/committees/ciq/ciq.html#6
 225         $address['geoloc'] = str_replace(", ", "\n", $geocodedData['address']);
 226         if (isset($geocodedData['AddressDetails']['Accuracy'])) {
 227             $address['accuracy'] = $geocodedData['AddressDetails']['Accuracy'];
 228         }
 229
 230         $currentPosition = $geocodedData['AddressDetails'];
 231         if (isset($currentPosition['Country'])) {
 232             $currentPosition      = $currentPosition['Country'];
 233             $address['countryId'] = $currentPosition['CountryNameCode'];
 234             $address['country']   = $currentPosition['CountryName'];
 235         }
 236         if (isset($currentPosition['AdministrativeArea'])) {
 237             $currentPosition                   = $currentPosition['AdministrativeArea'];
 238             $address['administrativeAreaName'] = $currentPosition['AdministrativeAreaName'];
 239         }
 240         if (isset($currentPosition['SubAdministrativeArea'])) {
 241             $currentPosition                      = $currentPosition['SubAdministrativeArea'];
 242             $address['subAdministrativeAreaName'] = $currentPosition['SubAdministrativeAreaName'];
 243         }
 244         if (isset($currentPosition['Locality'])) {
 245             $currentPosition          = $currentPosition['Locality'];
 246             $address['localityName']  = $currentPosition['LocalityName'];
 247         }
 248         if (isset($currentPosition['Thoroughfare'])) {
 249             $address['thoroughfareName'] = $currentPosition['Thoroughfare']['ThoroughfareName'];
 250         }
 251         if (isset($currentPosition['PostalCode'])) {
 252             $address['postalCode'] = $currentPosition['PostalCode']['PostalCodeNumber'];
 253         }
 254
 255         // Gets coordinates.
 256         if (isset($geocodedData['Point']['coordinates'][0])) {
 257             $address['latitude'] = $geocodedData['Point']['coordinates'][0];
 258         }
 259         if (isset($geocodedData['Point']['coordinates'][1])) {
 260             $address['longitude'] = $geocodedData['Point']['coordinates'][1];
 261         }
 262         if (isset($geocodedData['ExtendedData']['LatLonBox']['north'])) {
 263             $address['north'] = $geocodedData['ExtendedData']['LatLonBox']['north'];
 264         }
 265         if (isset($geocodedData['ExtendedData']['LatLonBox']['south'])) {
 266             $address['south'] = $geocodedData['ExtendedData']['LatLonBox']['south'];
 267         }
 268         if (isset($geocodedData['ExtendedData']['LatLonBox']['east'])) {
 269             $address['east'] = $geocodedData['ExtendedData']['LatLonBox']['east'];
 270         }
 271         if (isset($geocodedData['ExtendedData']['LatLonBox']['west'])) {
 272             $address['west'] = $geocodedData['ExtendedData']['LatLonBox']['west'];
 273         }
 274     }
 275
 276     // Formats the text of the geocoded address using the unused data and
 277     // compares it to the given address. If they are too different, the user
 278     // will be asked to choose between them.
 279     private function formatAddress(&$address, $extraLines) {
 280         $same = true;
 281         if ($extraLines) {
 282             $address['geoloc'] = $extraLines . "\n" . $address['geoloc'];
 283         }
 284         $geoloc = strtoupper(preg_replace(array("/[0-9,\"'#~:;_\- ]/", "/\r\n/"),
 285                                           array("", "\n"), $address['geoloc']));
 286         $text   = strtoupper(preg_replace(array("/[0-9,\"'#~:;_\- ]/", "/\r\n/"),
 287                                           array("", "\n"), $address['text']));
 288         $arrayGeoloc = explode("\n", $geoloc);
 289         $arrayText   = explode("\n", $text);
 290         $countGeoloc = count($arrayGeoloc);
 291         $countText   = count($arrayText);
 292
 293         if (($countText > $countGeoloc) || ($countText < $countGeoloc - 1)
 294             || (($countText == $countGeoloc - 1)
 295                 && ($arrayText[$countText - 1] == strtoupper($address['country'])))) {
 296             $same = false;
 297         } else {
 298             for ($i = 0; $i < $countGeoloc && $i < $countText; ++$i) {
 299                 if (levenshtein($arrayText[$i], trim($arrayGeoloc[$i])) > 3) {
 300                     $same = false;
 301                 }
 302             }
 303         }
 304         if ($same) {
 305             $address['text'] = $address['geoloc'];
 306             unset($address['geoloc']);
 307         }
 308     }
 309
 310     // Search for the lign from the given address that is the closest to the geocoded thoroughfareName
 311     // and replaces the corresponding lign in the geocoded text by it.
 312     static protected function fixStreetNumber(&$address)
 313     {
 314         if (isset($address['thoroughfareName'])) {
 315             $thoroughfareName  = $address['thoroughfareName'];
 316             $thoroughfareToken = strtoupper(trim(preg_replace(array("/[,\"'#~:;_\-]/", "/\r\n/"),
 317                                                               array("", "\n"), $thoroughfareName)));
 318             $geolocLines = explode("\n", $address['geoloc']);
 319             $textLines   = explode("\n", $address['text']);
 320             $mindist = strlen($thoroughfareToken);
 321             $minpos  = 0;
 322             $pos     = 0;
 323             foreach ($textLines as $i => $token) {
 324                 if (($l = levenshtein(strtoupper(trim(preg_replace(array("/[,\"'#~:;_\-]/", "/\r\n/"),
 325                                                                    array("", "\n"), $token))),
 326                                       $thoroughfareToken)) < $mindist) {
 327                     $mindist = $l;
 328                     $minpos  = $i;
 329                 }
 330             }
 331             foreach ($geolocLines as $i => $line) {
 332                 if (strtoupper(trim($thoroughfareName)) == strtoupper(trim($line))) {
 333                     $pos = $i;
 334                     break;
 335                 }
 336             }
 337             $geolocLines[$pos] = $textLines[$minpos];
 338             $address['geoloc'] = implode("\n", $geolocLines);
 339         }
 340     }
 341 }
 342
 343 // vim:set et sw=4 sts=4 sws=4 foldmethod=marker enc=utf-8:
 344 ?>