include/geocoding.inc.php

   1 <?php
   2 /***************************************************************************
   3  *  Copyright (C) 2003-2010 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     // Cleans the address from its geocoded data
  31     abstract public function stripGeocodingFromAddress(array $address);
  32
  33     // Updates geoloc_administrativeareas, geoloc_subadministrativeareas and
  34     // geoloc_localities databases with new geocoded data and returns the
  35     // corresponding id.
  36     static public function getAreaId(array &$address, $area)
  37     {
  38         static $databases = array(
  39             'administrativeArea'    => 'geoloc_administrativeareas',
  40             'subAdministrativeArea' => 'geoloc_subadministrativeareas',
  41             'locality'              => 'geoloc_localities',
  42             );
  43
  44         if (isset($address[$area . 'Name']) && isset($databases[$area])) {
  45             $res = XDB::query("SELECT  id
  46                                  FROM  " . $databases[$area] . "
  47                                 WHERE  name = {?}",
  48                               $address[$area . 'Name']);
  49             if ($res->numRows() == 0) {
  50                 XDB::execute('INSERT INTO  ' . $databases[$area] . ' (name, country)
  51                                    VALUES  ({?}, {?})',
  52                              $address[$area . 'Name'], $address['countryId']);
  53                 $address[$area . 'Id'] = XDB::insertId();
  54             } else {
  55                 $address[$area . 'Id'] = $res->fetchOneCell();
  56             }
  57         }
  58     }
  59
  60     // Returns the part of the text preceeding the line with the postal code
  61     // and the city name, within the limit of $limit number of lines.
  62     static public function getFirstLines($text, $postalCode, $limit)
  63     {
  64         $textArray  = explode("\n", $text);
  65         for ($i = 0; $i < count($textArray); ++$i) {
  66             if ($i > $limit || strpos($textLine, $postalCode) !== false) {
  67                 $limit = $i; break;
  68             }
  69         }
  70         return implode("\n", array_slice($textArray, 0, $limit));
  71     }
  72
  73     // Returns the number of non geocoded addresses for a user.
  74     static public function countNonGeocoded($pid)
  75     {
  76         $res = XDB::query("SELECT  COUNT(*)
  77                              FROM  profile_addresses
  78                             WHERE  pid = {?} AND FIND_IN_SET('home', type) AND accuracy = 0",
  79                           $pid);
  80         return $res->fetchOneCell();
  81     }
  82 }
  83
  84 // Implementation of a Geocoder using the Google Maps API. Please refer to
  85 // the following links for details:
  86 // http://code.google.com/apis/maps/documentation/services.html#Geocoding
  87 // http://code.google.com/intl/en/apis/maps/documentation/geocoding/
  88 // http://code.google.com/apis/maps/documentation/reference.html#GGeoAddressAccuracy
  89 //
  90 // It requires the properties gmaps_key and gmaps_url to be defined in section
  91 // Geocoder in plat/al's configuration (platal.ini & platal.conf).
  92 class GMapsGeocoder extends Geocoder {
  93
  94     // Maximum number of Geocoding calls to the Google Maps API.
  95     const MAX_GMAPS_RPC_CALLS = 5;
  96     // Maximum levenshtein distance authorized between input and geocoded text in a single line.
  97     const MAX_LINE_DISTANCE = 5;
  98     // Maximum levenshtein distance authorized between input and geocoded text in the whole text.
  99     const MAX_TOTAL_DISTANCE = 6;
 100
 101     public function getGeocodedAddress(array $address) {
 102         $address = $this->prepareAddress($address);
 103         $textAddress = $this->getTextToGeocode($address);
 104
 105         // Try to geocode the full address.
 106         if (($geocodedData = $this->getPlacemarkForAddress($textAddress))) {
 107             return $this->getUpdatedAddress($address, $geocodedData, null);
 108         }
 109
 110         // If the full geocoding failed, try to geocode only the final part of the address.
 111         // We start by geocoding everything but the first line, and continue until we get
 112         // a result. To respect the limit of GMaps calls, we ignore the first few lines
 113         // if there are too many address lines.
 114         $addressLines = explode("\n", $textAddress);
 115         $linesCount   = count($addressLines);
 116         for ($i = max(1, $linesCount - self::MAX_GMAPS_RPC_CALLS + 1); $i < $linesCount; ++$i) {
 117             $extraLines = implode("\n", array_slice($addressLines, 0, $i));
 118             $toGeocode  = implode("\n", array_slice($addressLines, $i));
 119             if (($geocodedData = $this->getPlacemarkForAddress($toGeocode))) {
 120                 return $this->getUpdatedAddress($address, $geocodedData, $extraLines);
 121             }
 122         }
 123
 124         // No geocoding could be done, the initial address is returned as-is.
 125         return $address;
 126     }
 127
 128     public function stripGeocodingFromAddress(array $address) {
 129         unset($address['geoloc'], $address['geoloc_choice'], $address['geocodedPostalText'],
 130               $address['countryId'], $address['country'], $address['administrativeAreaName'],
 131               $address['subAdministrativeAreaName'], $address['localityName'],
 132               $address['thoroughfareName'], $address['postalCode']);
 133         $address['accuracy'] = 0;
 134         return $address;
 135     }
 136
 137     // Updates the address with the geocoded information from Google Maps. Also
 138     // cleans up the final informations.
 139     private function getUpdatedAddress(array $address, array $geocodedData, $extraLines) {
 140         $this->fillAddressWithGeocoding(&$address, $geocodedData);
 141         $this->formatAddress($address, $extraLines);
 142         return $address;
 143     }
 144
 145     // Retrieves the Placemark object (see #getPlacemarkFromJson()) for the @p
 146     // address, by querying the Google Maps API. Returns the array on success,
 147     // and null otherwise.
 148     private function getPlacemarkForAddress($address) {
 149         $url     = $this->getGeocodingUrl($address);
 150         $geoData = $this->getGeoJsonFromUrl($url);
 151
 152         return ($geoData ? $this->getPlacemarkFromJson($geoData) : null);
 153     }
 154
 155     // Prepares address to be geocoded
 156     private function prepareAddress($address) {
 157         $address['text'] = preg_replace('/\s*\n\s*/m', "\n", trim($address['text']));
 158         $address['postalText'] = $this->getPostalAddress($address['text']);
 159         $address['updateTime'] = time();
 160         unset($address['changed']);
 161         return $address;
 162     }
 163
 164     // Builds the Google Maps geocoder url to fetch information about @p address.
 165     // Returns the built url.
 166     private function getGeocodingUrl($address) {
 167         global $globals;
 168
 169         $parameters = array(
 170             'key'    => $globals->geocoder->gmaps_key,
 171             'sensor' => 'false',   // The queried address wasn't obtained from a GPS sensor.
 172             'hl'     => 'fr',      // Output langage.
 173             'oe'     => 'utf8',    // Output encoding.
 174             'output' => 'json',    // Output format.
 175             'gl'     => 'fr',      // Location preferences (addresses are in France by default).
 176             'q'      => $address,  // The queries address.
 177         );
 178
 179         return $globals->geocoder->gmaps_url . '?' . http_build_query($parameters);
 180     }
 181
 182     // Fetches JSON-encoded data from a Google Maps API url, and decode them.
 183     // Returns the json array on success, and null otherwise.
 184     private function getGeoJsonFromUrl($url) {
 185         global $globals;
 186
 187         // Prepare a backtrace object to log errors.
 188         $bt = null;
 189         if ($globals->debug & DEBUG_BT) {
 190             if (!isset(PlBacktrace::$bt['Geoloc'])) {
 191                 new PlBacktrace('Geoloc');
 192             }
 193             $bt = &PlBacktrace::$bt['Geoloc'];
 194             $bt->start($url);
 195         }
 196
 197         // Fetch the geocoding data.
 198         $rawData = file_get_contents($url);
 199         if (!$rawData) {
 200             if ($bt) {
 201                 $bt->stop(0, "Could not retrieve geocoded address from GoogleMaps.");
 202             }
 203             return null;
 204         }
 205
 206         // Decode the JSON-encoded data, and check for their validity.
 207         $data = json_decode($rawData, true);
 208         if ($bt) {
 209             $bt->stop(count($data), null, $data);
 210         }
 211
 212         return $data;
 213     }
 214
 215     // Extracts the most appropriate placemark from the JSON data fetched from
 216     // Google Maps. Returns a Placemark array on success, and null otherwise. See
 217     // http://code.google.com/apis/maps/documentation/services.html#Geocoding_Structured
 218     // for details on the Placemark structure.
 219     private function getPlacemarkFromJson(array $data) {
 220         // Check for geocoding failures.
 221         if (!isset($data['Status']['code']) || $data['Status']['code'] != 200) {
 222             // TODO: handle non-200 codes in a better way, since the code might
 223             // indicate a temporary error on Google's side.
 224             return null;
 225         }
 226
 227         // Check that at least one placemark was found.
 228         if (count($data['Placemark']) == 0) {
 229             return null;
 230         }
 231
 232         // Extract the placemark with the best accuracy. This is not always the
 233         // best result (since the same address may yield two different placemarks).
 234         $result = $data['Placemark'][0];
 235         foreach ($data['Placemark'] as $place) {
 236             if ($place['AddressDetails']['Accuracy'] > $result['AddressDetails']['Accuracy']) {
 237                 $result = $place;
 238             }
 239         }
 240
 241         return $result;
 242     }
 243
 244     // Fills the address with the geocoded data
 245     private function fillAddressWithGeocoding(&$address, $geocodedData) {
 246         // The geocoded address three is
 247         // Country -> AdministrativeArea -> SubAdministrativeArea -> Locality -> Thoroughfare
 248         // with all the possible shortcuts
 249         // The address is formatted as xAL, or eXtensible Address Language, an international
 250         // standard for address formatting.
 251         // xAL documentation: http://www.oasis-open.org/committees/ciq/ciq.html#6
 252         $address['geoloc'] = str_replace(", ", "\n", $geocodedData['address']);
 253         if (isset($geocodedData['AddressDetails']['Accuracy'])) {
 254             $address['accuracy'] = $geocodedData['AddressDetails']['Accuracy'];
 255         }
 256
 257         $currentPosition = $geocodedData['AddressDetails'];
 258         if (isset($currentPosition['Country'])) {
 259             $currentPosition      = $currentPosition['Country'];
 260             $address['countryId'] = $currentPosition['CountryNameCode'];
 261             $address['country']   = $currentPosition['CountryName'];
 262         }
 263         if (isset($currentPosition['AdministrativeArea'])) {
 264             $currentPosition                   = $currentPosition['AdministrativeArea'];
 265             $address['administrativeAreaName'] = $currentPosition['AdministrativeAreaName'];
 266         }
 267         if (isset($currentPosition['SubAdministrativeArea'])) {
 268             $currentPosition                      = $currentPosition['SubAdministrativeArea'];
 269             $address['subAdministrativeAreaName'] = $currentPosition['SubAdministrativeAreaName'];
 270         }
 271         if (isset($currentPosition['Locality'])) {
 272             $currentPosition          = $currentPosition['Locality'];
 273             $address['localityName']  = $currentPosition['LocalityName'];
 274         }
 275         if (isset($currentPosition['Thoroughfare'])) {
 276             $address['thoroughfareName'] = $currentPosition['Thoroughfare']['ThoroughfareName'];
 277         }
 278         if (isset($currentPosition['PostalCode'])) {
 279             $address['postalCode'] = $currentPosition['PostalCode']['PostalCodeNumber'];
 280         }
 281
 282         // Gets coordinates.
 283         if (isset($geocodedData['Point']['coordinates'][0])) {
 284             $address['latitude'] = $geocodedData['Point']['coordinates'][0];
 285         }
 286         if (isset($geocodedData['Point']['coordinates'][1])) {
 287             $address['longitude'] = $geocodedData['Point']['coordinates'][1];
 288         }
 289         if (isset($geocodedData['ExtendedData']['LatLonBox']['north'])) {
 290             $address['north'] = $geocodedData['ExtendedData']['LatLonBox']['north'];
 291         }
 292         if (isset($geocodedData['ExtendedData']['LatLonBox']['south'])) {
 293             $address['south'] = $geocodedData['ExtendedData']['LatLonBox']['south'];
 294         }
 295         if (isset($geocodedData['ExtendedData']['LatLonBox']['east'])) {
 296             $address['east'] = $geocodedData['ExtendedData']['LatLonBox']['east'];
 297         }
 298         if (isset($geocodedData['ExtendedData']['LatLonBox']['west'])) {
 299             $address['west'] = $geocodedData['ExtendedData']['LatLonBox']['west'];
 300         }
 301     }
 302
 303     // Formats the text of the geocoded address using the unused data and
 304     // compares it to the given address. If they are too different, the user
 305     // will be asked to choose between them.
 306     private function formatAddress(&$address, $extraLines) {
 307         $same = true;
 308         if ($extraLines) {
 309             $address['geoloc'] = $extraLines . "\n" . $address['geoloc'];
 310         }
 311         $address['geocodedPostalText'] = $this->getPostalAddress($address['geoloc']);
 312         $geoloc = strtoupper(preg_replace(array("/[0-9,\"'#~:;_\- ]/", "/\r\n/"),
 313                                           array("", "\n"), $address['geoloc']));
 314         $text   = strtoupper(preg_replace(array("/[0-9,\"'#~:;_\- ]/", "/\r\n/"),
 315                                           array("", "\n"), $address['text']));
 316         $arrayGeoloc = explode("\n", $geoloc);
 317         $arrayText   = explode("\n", $text);
 318         $countGeoloc = count($arrayGeoloc);
 319         $countText   = count($arrayText);
 320
 321         $totalDistance = 0;
 322         if (($countText > $countGeoloc) || ($countText < $countGeoloc - 1)
 323             || (($countText == $countGeoloc - 1)
 324                 && ($arrayText[$countText - 1] == strtoupper($address['country'])))) {
 325             $same = false;
 326         } else {
 327             for ($i = 0; $i < $countGeoloc && $i < $countText; ++$i) {
 328                 $lineDistance = levenshtein($arrayText[$i], trim($arrayGeoloc[$i]));
 329                 $totalDistance += $lineDistance;
 330                 if ($lineDistance > self::MAX_LINE_DISTANCE || $totalDistance > self::MAX_TOTAL_DISTANCE) {
 331                     $same = false;
 332                     break;
 333                 }
 334             }
 335         }
 336
 337         if ($same) {
 338             unset($address['geoloc'], $address['geocodedPostalText']);
 339         } else {
 340             $address['geoloc'] = str_replace("\n", "\r\n", $address['geoloc']);
 341             $address['geocodedPostalText'] = str_replace("\n", "\r\n", $address['geocodedPostalText']);
 342         }
 343         $address['text'] = str_replace("\n", "\r\n", $address['text']);
 344         $address['postalText'] = str_replace("\n", "\r\n", $address['postalText']);
 345     }
 346
 347     // Returns the address formated for postal use.
 348     // The main rules are (cf AFNOR XPZ 10-011):
 349     // -everything in upper case;
 350     // -if there are more then than 38 characters in a lign, split it;
 351     // -if there are more then than 32 characters in the description of the "street", use abbreviations.
 352     private function getPostalAddress($text) {
 353          static $abbreviations = array(
 354              "IMPASSE"   => "IMP",
 355              "RUE"       => "R",
 356              "AVENUE"    => "AV",
 357              "BOULEVARD" => "BVD",
 358              "ROUTE"     => "R",
 359              "STREET"    => "ST",
 360              "ROAD"      => "RD",
 361              );
 362
 363         $text = strtoupper($text);
 364         $arrayText = explode("\n", $text);
 365         $postalText = "";
 366
 367         foreach ($arrayText as $i => $lign) {
 368             $postalText .= (($i == 0) ? "" : "\n");
 369             if (($length = strlen($lign)) > 32) {
 370                 $words = explode(" ", $lign);
 371                 $count = 0;
 372                 foreach ($words as $word) {
 373                     if (isset($abbreviations[$word])) {
 374                         $word = $abbreviations[$word];
 375                     }
 376                     if ($count + ($wordLength = strlen($word)) <= 38) {
 377                         $postalText .= (($count == 0) ? "" : " ") . $word;
 378                         $count += (($count == 0) ? 0 : 1) + $wordLength;
 379                     } else {
 380                         $postalText .= "\n" . $word;
 381                         $count = strlen($word);
 382                     }
 383                 }
 384             } else {
 385                 $postalText .= $lign;
 386             }
 387         }
 388         return $postalText;
 389     }
 390
 391     // Trims the name of the real country if it contains an ISO 3166-1 non-country
 392     // item. For that purpose, we compare the last but one line of the address with
 393     // all non-country items of ISO 3166-1.
 394     private function getTextToGeocode($address)
 395     {
 396         $res = XDB::iterator('SELECT  country, countryFR
 397                                 FROM  geoloc_countries
 398                                WHERE  belongsTo IS NOT NULL');
 399         $countries = array();
 400         foreach ($res as $item) {
 401             $countries[] = $item[0];
 402             $countries[] = $item[1];
 403         }
 404         $textLines  = explode("\n", $address['text']);
 405         $countLines = count($textLines);
 406         $needle     = strtoupper(trim($textLines[$countLines - 2]));
 407         $isPseudoCountry = false;
 408         foreach ($countries as $country) {
 409             if (strtoupper($country) == $needle) {
 410                 $isPseudoCountry = true;
 411                 break;
 412             }
 413         }
 414
 415         if ($isPseudoCountry) {
 416             return implode("\n", array_slice($textLines, 0, -1));
 417         }
 418         return $address['text'];
 419     }
 420 }
 421
 422 // vim:set et sw=4 sts=4 sws=4 foldmethod=marker enc=utf-8:
 423 ?>