| 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 | $address[$area . 'Id'] = XDB::execute("INSERT INTO " . $databases[$area] . " (name, country) |
| 51 | VALUES ({?}, {?})", |
| 52 | $address[$area . 'Name'], $address['countryId']); |
| 53 | } else { |
| 54 | $address[$area . 'Id'] = $res->fetchOneCell(); |
| 55 | } |
| 56 | } |
| 57 | } |
| 58 | |
| 59 | // Returns the part of the text preceeding the line with the postal code |
| 60 | // and the city name, within the limit of $limit number of lines. |
| 61 | static public function getFirstLines($text, $postalCode, $limit) |
| 62 | { |
| 63 | $textArray = explode("\n", $text); |
| 64 | for ($i = 0; $i < count($textArray); ++$i) { |
| 65 | if ($i > $limit || strpos($textLine, $postalCode) !== false) { |
| 66 | $limit = $i; break; |
| 67 | } |
| 68 | } |
| 69 | return implode("\n", array_slice($textArray, 0, $limit)); |
| 70 | } |
| 71 | |
| 72 | // Returns the number of non geocoded addresses for a user. |
| 73 | static public function countNonGeocoded($pid) |
| 74 | { |
| 75 | $res = XDB::query("SELECT COUNT(*) |
| 76 | FROM profile_addresses |
| 77 | WHERE pid = {?} AND FIND_IN_SET('home', type) AND accuracy = 0", |
| 78 | $pid); |
| 79 | return $res->fetchOneCell(); |
| 80 | } |
| 81 | } |
| 82 | |
| 83 | // Implementation of a Geocoder using the Google Maps API. Please refer to |
| 84 | // the following links for details: |
| 85 | // http://code.google.com/apis/maps/documentation/services.html#Geocoding |
| 86 | // http://code.google.com/intl/en/apis/maps/documentation/geocoding/ |
| 87 | // http://code.google.com/apis/maps/documentation/reference.html#GGeoAddressAccuracy |
| 88 | // |
| 89 | // It requires the properties gmaps_key and gmaps_url to be defined in section |
| 90 | // Geocoder in plat/al's configuration (platal.ini & platal.conf). |
| 91 | class GMapsGeocoder extends Geocoder { |
| 92 | |
| 93 | // Maximum number of Geocoding calls to the Google Maps API. |
| 94 | const MAX_GMAPS_RPC_CALLS = 5; |
| 95 | |
| 96 | public function getGeocodedAddress(array $address) { |
| 97 | $address = $this->prepareAddress($address); |
| 98 | $textAddress = $this->getTextToGeocode($address); |
| 99 | |
| 100 | // Try to geocode the full address. |
| 101 | if (($geocodedData = $this->getPlacemarkForAddress($textAddress))) { |
| 102 | return $this->getUpdatedAddress($address, $geocodedData, null); |
| 103 | } |
| 104 | |
| 105 | // If the full geocoding failed, try to geocode only the final part of the address. |
| 106 | // We start by geocoding everything but the first line, and continue until we get |
| 107 | // a result. To respect the limit of GMaps calls, we ignore the first few lines |
| 108 | // if there are too many address lines. |
| 109 | $addressLines = explode("\n", $textAddress); |
| 110 | $linesCount = count($addressLines); |
| 111 | for ($i = max(1, $linesCount - self::MAX_GMAPS_RPC_CALLS + 1); $i < $linesCount; ++$i) { |
| 112 | $extraLines = implode("\n", array_slice($addressLines, 0, $i)); |
| 113 | $toGeocode = implode("\n", array_slice($addressLines, $i)); |
| 114 | if (($geocodedData = $this->getPlacemarkForAddress($toGeocode))) { |
| 115 | return $this->getUpdatedAddress($address, $geocodedData, $extraLines); |
| 116 | } |
| 117 | } |
| 118 | |
| 119 | // No geocoding could be done, the initial address is returned as-is. |
| 120 | return $address; |
| 121 | } |
| 122 | |
| 123 | public function stripGeocodingFromAddress(array $address) { |
| 124 | unset($address['geoloc'], $address['geoloc_choice'], $address['geocodedPostalText'], |
| 125 | $address['countryId'], $address['country'], $address['administrativeAreaName'], |
| 126 | $address['subAdministrativeAreaName'], $address['localityName'], |
| 127 | $address['thoroughfareName'], $address['postalCode']); |
| 128 | $address['accuracy'] = 0; |
| 129 | return $address; |
| 130 | } |
| 131 | |
| 132 | // Updates the address with the geocoded information from Google Maps. Also |
| 133 | // cleans up the final informations. |
| 134 | private function getUpdatedAddress(array $address, array $geocodedData, $extraLines) { |
| 135 | $this->fillAddressWithGeocoding(&$address, $geocodedData); |
| 136 | |
| 137 | // If the accuracy is 6, it means only the street has been gecoded |
| 138 | // but not the number, thus we need to fix it. |
| 139 | if ($address['accuracy'] == 6) { |
| 140 | $this->fixStreetNumber($address); |
| 141 | } |
| 142 | |
| 143 | // We can now format the address. |
| 144 | $this->formatAddress($address, $extraLines); |
| 145 | |
| 146 | return $address; |
| 147 | } |
| 148 | |
| 149 | // Retrieves the Placemark object (see #getPlacemarkFromJson()) for the @p |
| 150 | // address, by querying the Google Maps API. Returns the array on success, |
| 151 | // and null otherwise. |
| 152 | private function getPlacemarkForAddress($address) { |
| 153 | $url = $this->getGeocodingUrl($address); |
| 154 | $geoData = $this->getGeoJsonFromUrl($url); |
| 155 | |
| 156 | return ($geoData ? $this->getPlacemarkFromJson($geoData) : null); |
| 157 | } |
| 158 | |
| 159 | // Prepares address to be geocoded |
| 160 | private function prepareAddress($address) { |
| 161 | $address['text'] = preg_replace('/\s*\n\s*/m', "\n", trim($address['text'])); |
| 162 | $address['postalText'] = $this->getPostalAddress($address['text']); |
| 163 | $address['updateTime'] = time(); |
| 164 | unset($address['changed']); |
| 165 | return $address; |
| 166 | } |
| 167 | |
| 168 | // Builds the Google Maps geocoder url to fetch information about @p address. |
| 169 | // Returns the built url. |
| 170 | private function getGeocodingUrl($address) { |
| 171 | global $globals; |
| 172 | |
| 173 | $parameters = array( |
| 174 | 'key' => $globals->geocoder->gmaps_key, |
| 175 | 'sensor' => 'false', // The queried address wasn't obtained from a GPS sensor. |
| 176 | 'hl' => 'fr', // Output langage. |
| 177 | 'oe' => 'utf8', // Output encoding. |
| 178 | 'output' => 'json', // Output format. |
| 179 | 'gl' => 'fr', // Location preferences (addresses are in France by default). |
| 180 | 'q' => $address, // The queries address. |
| 181 | ); |
| 182 | |
| 183 | return $globals->geocoder->gmaps_url . '?' . http_build_query($parameters); |
| 184 | } |
| 185 | |
| 186 | // Fetches JSON-encoded data from a Google Maps API url, and decode them. |
| 187 | // Returns the json array on success, and null otherwise. |
| 188 | private function getGeoJsonFromUrl($url) { |
| 189 | global $globals; |
| 190 | |
| 191 | // Prepare a backtrace object to log errors. |
| 192 | $bt = null; |
| 193 | if ($globals->debug & DEBUG_BT) { |
| 194 | if (!isset(PlBacktrace::$bt['Geoloc'])) { |
| 195 | new PlBacktrace('Geoloc'); |
| 196 | } |
| 197 | $bt = &PlBacktrace::$bt['Geoloc']; |
| 198 | $bt->start($url); |
| 199 | } |
| 200 | |
| 201 | // Fetch the geocoding data. |
| 202 | $rawData = file_get_contents($url); |
| 203 | if (!$rawData) { |
| 204 | if ($bt) { |
| 205 | $bt->stop(0, "Could not retrieve geocoded address from GoogleMaps."); |
| 206 | } |
| 207 | return null; |
| 208 | } |
| 209 | |
| 210 | // Decode the JSON-encoded data, and check for their validity. |
| 211 | $data = json_decode($rawData, true); |
| 212 | if ($bt) { |
| 213 | $bt->stop(count($data), null, $data); |
| 214 | } |
| 215 | |
| 216 | return $data; |
| 217 | } |
| 218 | |
| 219 | // Extracts the most appropriate placemark from the JSON data fetched from |
| 220 | // Google Maps. Returns a Placemark array on success, and null otherwise. See |
| 221 | // http://code.google.com/apis/maps/documentation/services.html#Geocoding_Structured |
| 222 | // for details on the Placemark structure. |
| 223 | private function getPlacemarkFromJson(array $data) { |
| 224 | // Check for geocoding failures. |
| 225 | if (!isset($data['Status']['code']) || $data['Status']['code'] != 200) { |
| 226 | // TODO: handle non-200 codes in a better way, since the code might |
| 227 | // indicate a temporary error on Google's side. |
| 228 | return null; |
| 229 | } |
| 230 | |
| 231 | // Check that at least one placemark was found. |
| 232 | if (count($data['Placemark']) == 0) { |
| 233 | return null; |
| 234 | } |
| 235 | |
| 236 | // Extract the placemark with the best accuracy. This is not always the |
| 237 | // best result (since the same address may yield two different placemarks). |
| 238 | $result = $data['Placemark'][0]; |
| 239 | foreach ($data['Placemark'] as $place) { |
| 240 | if ($place['AddressDetails']['Accuracy'] > $result['AddressDetails']['Accuracy']) { |
| 241 | $result = $place; |
| 242 | } |
| 243 | } |
| 244 | |
| 245 | return $result; |
| 246 | } |
| 247 | |
| 248 | // Fills the address with the geocoded data |
| 249 | private function fillAddressWithGeocoding(&$address, $geocodedData) { |
| 250 | // The geocoded address three is |
| 251 | // Country -> AdministrativeArea -> SubAdministrativeArea -> Locality -> Thoroughfare |
| 252 | // with all the possible shortcuts |
| 253 | // The address is formatted as xAL, or eXtensible Address Language, an international |
| 254 | // standard for address formatting. |
| 255 | // xAL documentation: http://www.oasis-open.org/committees/ciq/ciq.html#6 |
| 256 | $address['geoloc'] = str_replace(", ", "\n", $geocodedData['address']); |
| 257 | if (isset($geocodedData['AddressDetails']['Accuracy'])) { |
| 258 | $address['accuracy'] = $geocodedData['AddressDetails']['Accuracy']; |
| 259 | } |
| 260 | |
| 261 | $currentPosition = $geocodedData['AddressDetails']; |
| 262 | if (isset($currentPosition['Country'])) { |
| 263 | $currentPosition = $currentPosition['Country']; |
| 264 | $address['countryId'] = $currentPosition['CountryNameCode']; |
| 265 | $address['country'] = $currentPosition['CountryName']; |
| 266 | } |
| 267 | if (isset($currentPosition['AdministrativeArea'])) { |
| 268 | $currentPosition = $currentPosition['AdministrativeArea']; |
| 269 | $address['administrativeAreaName'] = $currentPosition['AdministrativeAreaName']; |
| 270 | } |
| 271 | if (isset($currentPosition['SubAdministrativeArea'])) { |
| 272 | $currentPosition = $currentPosition['SubAdministrativeArea']; |
| 273 | $address['subAdministrativeAreaName'] = $currentPosition['SubAdministrativeAreaName']; |
| 274 | } |
| 275 | if (isset($currentPosition['Locality'])) { |
| 276 | $currentPosition = $currentPosition['Locality']; |
| 277 | $address['localityName'] = $currentPosition['LocalityName']; |
| 278 | } |
| 279 | if (isset($currentPosition['Thoroughfare'])) { |
| 280 | $address['thoroughfareName'] = $currentPosition['Thoroughfare']['ThoroughfareName']; |
| 281 | } |
| 282 | if (isset($currentPosition['PostalCode'])) { |
| 283 | $address['postalCode'] = $currentPosition['PostalCode']['PostalCodeNumber']; |
| 284 | } |
| 285 | |
| 286 | // Gets coordinates. |
| 287 | if (isset($geocodedData['Point']['coordinates'][0])) { |
| 288 | $address['latitude'] = $geocodedData['Point']['coordinates'][0]; |
| 289 | } |
| 290 | if (isset($geocodedData['Point']['coordinates'][1])) { |
| 291 | $address['longitude'] = $geocodedData['Point']['coordinates'][1]; |
| 292 | } |
| 293 | if (isset($geocodedData['ExtendedData']['LatLonBox']['north'])) { |
| 294 | $address['north'] = $geocodedData['ExtendedData']['LatLonBox']['north']; |
| 295 | } |
| 296 | if (isset($geocodedData['ExtendedData']['LatLonBox']['south'])) { |
| 297 | $address['south'] = $geocodedData['ExtendedData']['LatLonBox']['south']; |
| 298 | } |
| 299 | if (isset($geocodedData['ExtendedData']['LatLonBox']['east'])) { |
| 300 | $address['east'] = $geocodedData['ExtendedData']['LatLonBox']['east']; |
| 301 | } |
| 302 | if (isset($geocodedData['ExtendedData']['LatLonBox']['west'])) { |
| 303 | $address['west'] = $geocodedData['ExtendedData']['LatLonBox']['west']; |
| 304 | } |
| 305 | } |
| 306 | |
| 307 | // Formats the text of the geocoded address using the unused data and |
| 308 | // compares it to the given address. If they are too different, the user |
| 309 | // will be asked to choose between them. |
| 310 | private function formatAddress(&$address, $extraLines) { |
| 311 | $same = true; |
| 312 | if ($extraLines) { |
| 313 | $address['geoloc'] = $extraLines . "\n" . $address['geoloc']; |
| 314 | } |
| 315 | $address['geocodedPostalText'] = $this->getPostalAddress($address['geoloc']); |
| 316 | $geoloc = strtoupper(preg_replace(array("/[0-9,\"'#~:;_\- ]/", "/\r\n/"), |
| 317 | array("", "\n"), $address['geoloc'])); |
| 318 | $text = strtoupper(preg_replace(array("/[0-9,\"'#~:;_\- ]/", "/\r\n/"), |
| 319 | array("", "\n"), $address['text'])); |
| 320 | $arrayGeoloc = explode("\n", $geoloc); |
| 321 | $arrayText = explode("\n", $text); |
| 322 | $countGeoloc = count($arrayGeoloc); |
| 323 | $countText = count($arrayText); |
| 324 | |
| 325 | if (($countText > $countGeoloc) || ($countText < $countGeoloc - 1) |
| 326 | || (($countText == $countGeoloc - 1) |
| 327 | && ($arrayText[$countText - 1] == strtoupper($address['country'])))) { |
| 328 | $same = false; |
| 329 | } else { |
| 330 | for ($i = 0; $i < $countGeoloc && $i < $countText; ++$i) { |
| 331 | if (levenshtein($arrayText[$i], trim($arrayGeoloc[$i])) > 3) { |
| 332 | $same = false; |
| 333 | } |
| 334 | } |
| 335 | } |
| 336 | if ($same) { |
| 337 | $address['text'] = $address['geoloc']; |
| 338 | $address['postalText'] = $address['geocodedPostalText']; |
| 339 | unset($address['geoloc'], $address['geocodedPostalText']); |
| 340 | } else { |
| 341 | $address['geoloc'] = str_replace("\n", "\r\n", $address['geoloc']); |
| 342 | $address['geocodedPostalText'] = str_replace("\n", "\r\n", $address['geocodedPostalText']); |
| 343 | } |
| 344 | $address['text'] = str_replace("\n", "\r\n", $address['text']); |
| 345 | $address['postalText'] = str_replace("\n", "\r\n", $address['postalText']); |
| 346 | } |
| 347 | |
| 348 | // Returns the address formated for postal use. |
| 349 | // The main rules are (cf AFNOR XPZ 10-011): |
| 350 | // -everything in upper case; |
| 351 | // -if there are more then than 38 characters in a lign, split it; |
| 352 | // -if there are more then than 32 characters in the description of the "street", use abbreviations. |
| 353 | private function getPostalAddress($text) { |
| 354 | static $abbreviations = array( |
| 355 | "IMPASSE" => "IMP", |
| 356 | "RUE" => "R", |
| 357 | "AVENUE" => "AV", |
| 358 | "BOULEVARD" => "BVD", |
| 359 | "ROUTE" => "R", |
| 360 | "STREET" => "ST", |
| 361 | "ROAD" => "RD", |
| 362 | ); |
| 363 | |
| 364 | $text = strtoupper($text); |
| 365 | $arrayText = explode("\n", $text); |
| 366 | $postalText = ""; |
| 367 | |
| 368 | foreach ($arrayText as $i => $lign) { |
| 369 | $postalText .= (($i == 0) ? "" : "\n"); |
| 370 | if (($length = strlen($lign)) > 32) { |
| 371 | $words = explode(" ", $lign); |
| 372 | $count = 0; |
| 373 | foreach ($words as $word) { |
| 374 | if (isset($abbreviations[$word])) { |
| 375 | $word = $abbreviations[$word]; |
| 376 | } |
| 377 | if ($count + ($wordLength = strlen($word)) <= 38) { |
| 378 | $postalText .= (($count == 0) ? "" : " ") . $word; |
| 379 | $count += (($count == 0) ? 0 : 1) + $wordLength; |
| 380 | } else { |
| 381 | $postalText .= "\n" . $word; |
| 382 | $count = strlen($word); |
| 383 | } |
| 384 | } |
| 385 | } else { |
| 386 | $postalText .= $lign; |
| 387 | } |
| 388 | } |
| 389 | return $postalText; |
| 390 | } |
| 391 | |
| 392 | // Trims the name of the real country if it contains an ISO 3166-1 non-country |
| 393 | // item. For that purpose, we compare the last but one line of the address with |
| 394 | // all non-country items of ISO 3166-1. |
| 395 | private function getTextToGeocode($address) |
| 396 | { |
| 397 | $res = XDB::iterator('SELECT country, countryFR |
| 398 | FROM geoloc_countries |
| 399 | WHERE belongsTo IS NOT NULL'); |
| 400 | $countries = array(); |
| 401 | foreach ($res as $item) { |
| 402 | $countries[] = $item[0]; |
| 403 | $countries[] = $item[1]; |
| 404 | } |
| 405 | $textLines = explode("\n", $address['text']); |
| 406 | $countLines = count($textLines); |
| 407 | $needle = strtoupper(trim($textLines[$countLines - 2])); |
| 408 | $isPseudoCountry = false; |
| 409 | foreach ($countries as $country) { |
| 410 | if (strtoupper($country) == $needle) { |
| 411 | $isPseudoCountry = true; |
| 412 | break; |
| 413 | } |
| 414 | } |
| 415 | |
| 416 | if ($isPseudoCountry) { |
| 417 | return $address['text']; |
| 418 | } |
| 419 | return implode("\n", array_slice($textLines, 0, -1)); |
| 420 | } |
| 421 | |
| 422 | // Search for the lign from the given address that is the closest to the geocoded thoroughfareName |
| 423 | // and replaces the corresponding lign in the geocoded text by it. |
| 424 | static protected function fixStreetNumber(&$address) |
| 425 | { |
| 426 | if (isset($address['thoroughfareName'])) { |
| 427 | $thoroughfareName = $address['thoroughfareName']; |
| 428 | $thoroughfareToken = strtoupper(trim(preg_replace(array("/[,\"'#~:;_\-]/", "/\r\n/"), |
| 429 | array("", "\n"), $thoroughfareName))); |
| 430 | $geolocLines = explode("\n", $address['geoloc']); |
| 431 | $textLines = explode("\n", $address['text']); |
| 432 | $mindist = strlen($thoroughfareToken); |
| 433 | $minpos = 0; |
| 434 | $pos = 0; |
| 435 | foreach ($textLines as $i => $token) { |
| 436 | if (($l = levenshtein(strtoupper(trim(preg_replace(array("/[,\"'#~:;_\-]/", "/\r\n/"), |
| 437 | array("", "\n"), $token))), |
| 438 | $thoroughfareToken)) < $mindist) { |
| 439 | $mindist = $l; |
| 440 | $minpos = $i; |
| 441 | } |
| 442 | } |
| 443 | foreach ($geolocLines as $i => $line) { |
| 444 | if (strtoupper(trim($thoroughfareName)) == strtoupper(trim($line))) { |
| 445 | $pos = $i; |
| 446 | break; |
| 447 | } |
| 448 | } |
| 449 | $geolocLines[$pos] = $textLines[$minpos]; |
| 450 | $address['geoloc'] = implode("\n", $geolocLines); |
| 451 | } |
| 452 | } |
| 453 | } |
| 454 | |
| 455 | // vim:set et sw=4 sts=4 sws=4 foldmethod=marker enc=utf-8: |
| 456 | ?> |