gnu-social/plugins/Geonames/GeonamesPlugin.php

513 lines
16 KiB
PHP
Raw Normal View History

2009-10-22 20:44:36 +01:00
<?php
/**
* StatusNet, the distributed open-source microblogging tool
*
* Plugin to convert string locations to Geonames IDs and vice versa
*
* PHP version 5
*
* LICENCE: This program is free software: you can redistribute it and/or modify
* it under the terms of the GNU Affero General Public License as published by
* the Free Software Foundation, either version 3 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 Affero General Public License for more details.
*
* You should have received a copy of the GNU Affero General Public License
* along with this program. If not, see <http://www.gnu.org/licenses/>.
*
* @category Action
* @package StatusNet
* @author Evan Prodromou <evan@status.net>
* @copyright 2009 StatusNet Inc.
* @license http://www.fsf.org/licensing/licenses/agpl-3.0.html GNU Affero General Public License version 3.0
* @link http://status.net/
*/
if (!defined('STATUSNET')) {
exit(1);
}
/**
* Plugin to convert string locations to Geonames IDs and vice versa
*
* This handles most of the events that Location class emits. It uses
* the geonames.org Web service to convert names like 'Montreal, Quebec, Canada'
* into IDs and lat/lon pairs.
*
* @category Plugin
* @package StatusNet
* @author Evan Prodromou <evan@status.net>
* @license http://www.fsf.org/licensing/licenses/agpl-3.0.html GNU Affero General Public License version 3.0
* @link http://status.net/
*
* @seeAlso Location
*/
class GeonamesPlugin extends Plugin
{
const PLUGIN_VERSION = '2.0.0';
const LOCATION_NS = 1;
2009-10-22 20:44:36 +01:00
public $host = 'ws.geonames.org';
public $username = null;
public $token = null;
public $expiry = 7776000; // 90-day expiry
public $timeout = 2; // Web service timeout in seconds.
public $timeoutWindow = 60; // Further lookups in this process will be disabled for N seconds after a timeout.
public $cachePrefix = null; // Optional shared memcache prefix override
// to share lookups between local instances.
protected $lastTimeout = null; // timestamp of last web service timeout
2009-10-22 20:44:36 +01:00
/**
* convert a name into a Location object
*
* @param string $name Name to convert
* @param string $language ISO code for anguage the name is in
* @param Location &$location Location object (may be null)
*
* @return boolean whether to continue (results in $location)
*/
function onLocationFromName($name, $language, &$location)
{
2009-11-20 15:58:28 +00:00
$loc = $this->getCache(array('name' => $name,
'language' => $language));
if ($loc !== false) {
2009-11-20 15:58:28 +00:00
$location = $loc;
return false;
}
try {
$geonames = $this->getGeonames('search',
array('maxRows' => 1,
'q' => $name,
'lang' => $language,
'type' => 'xml'));
} catch (Exception $e) {
$this->log(LOG_WARNING, "Error for $name: " . $e->getMessage());
return true;
}
2009-10-22 20:44:36 +01:00
if (count($geonames) == 0) {
// no results
$this->setCache(array('name' => $name,
'language' => $language),
null);
return true;
}
$n = $geonames[0];
2009-11-20 15:58:28 +00:00
$location = new Location();
$location->lat = $this->canonical($n->lat);
$location->lon = $this->canonical($n->lng);
$location->names[$language] = (string)$n->name;
$location->location_id = (string)$n->geonameId;
$location->location_ns = self::LOCATION_NS;
$this->setCache(array('name' => $name,
'language' => $language),
$location);
2009-10-22 20:44:36 +01:00
// handled, don't continue processing!
return false;
2009-10-22 20:44:36 +01:00
}
/**
* convert an id into a Location object
*
* @param string $id Name to convert
* @param string $ns Name to convert
* @param string $language ISO code for language for results
* @param Location &$location Location object (may be null)
*
* @return boolean whether to continue (results in $location)
*/
function onLocationFromId($id, $ns, $language, &$location)
{
if ($ns != self::LOCATION_NS) {
2009-10-22 20:44:36 +01:00
// It's not one of our IDs... keep processing
return true;
}
2009-11-20 15:58:28 +00:00
$loc = $this->getCache(array('id' => $id));
if ($loc !== false) {
2009-11-20 15:58:28 +00:00
$location = $loc;
return false;
}
try {
$geonames = $this->getGeonames('hierarchy',
array('geonameId' => $id,
'lang' => $language));
} catch (Exception $e) {
$this->log(LOG_WARNING, "Error for ID $id: " . $e->getMessage());
return false;
}
2009-10-22 20:44:36 +01:00
$parts = array();
2009-10-22 20:44:36 +01:00
foreach ($geonames as $level) {
if (in_array($level->fcode, array('PCLI', 'ADM1', 'PPL'))) {
$parts[] = (string)$level->name;
}
}
2009-10-22 20:44:36 +01:00
$last = $geonames[count($geonames)-1];
2009-10-22 20:44:36 +01:00
if (!in_array($level->fcode, array('PCLI', 'ADM1', 'PPL'))) {
$parts[] = (string)$last->name;
}
2009-10-22 20:44:36 +01:00
$location = new Location();
2009-10-22 20:44:36 +01:00
$location->location_id = (string)$last->geonameId;
$location->location_ns = self::LOCATION_NS;
$location->lat = $this->canonical($last->lat);
$location->lon = $this->canonical($last->lng);
$location->names[$language] = implode(', ', array_reverse($parts));
2009-11-20 15:58:28 +00:00
$this->setCache(array('id' => (string)$last->geonameId),
$location);
2009-10-22 20:44:36 +01:00
// We're responsible for this namespace; nobody else
2009-10-22 20:44:36 +01:00
// can resolve it
return false;
}
/**
* convert a lat/lon pair into a Location object
*
* Given a lat/lon, we try to find a Location that's around
* it or nearby. We prefer populated places (cities, towns, villages).
*
* @param string $lat Latitude
* @param string $lon Longitude
* @param string $language ISO code for language for results
* @param Location &$location Location object (may be null)
*
* @return boolean whether to continue (results in $location)
*/
function onLocationFromLatLon($lat, $lon, $language, &$location)
{
// Make sure they're canonical
$lat = $this->canonical($lat);
$lon = $this->canonical($lon);
2009-11-20 15:58:28 +00:00
$loc = $this->getCache(array('lat' => $lat,
'lon' => $lon));
if ($loc !== false) {
2009-11-20 15:58:28 +00:00
$location = $loc;
return false;
}
try {
$geonames = $this->getGeonames('findNearbyPlaceName',
array('lat' => $lat,
'lng' => $lon,
'lang' => $language));
} catch (Exception $e) {
$this->log(LOG_WARNING, "Error for coords $lat, $lon: " . $e->getMessage());
return true;
}
2009-10-22 20:44:36 +01:00
if (count($geonames) == 0) {
// no results
$this->setCache(array('lat' => $lat,
'lon' => $lon),
null);
return true;
}
$n = $geonames[0];
2009-10-22 20:44:36 +01:00
$parts = array();
2009-10-22 20:44:36 +01:00
$location = new Location();
2009-10-22 20:44:36 +01:00
$parts[] = (string)$n->name;
2009-10-22 20:44:36 +01:00
if (!empty($n->adminName1)) {
$parts[] = (string)$n->adminName1;
}
2009-10-22 20:44:36 +01:00
if (!empty($n->countryName)) {
$parts[] = (string)$n->countryName;
}
2009-10-22 20:44:36 +01:00
$location->location_id = (string)$n->geonameId;
$location->location_ns = self::LOCATION_NS;
$location->lat = $this->canonical($n->lat);
$location->lon = $this->canonical($n->lng);
2009-11-20 15:58:28 +00:00
$location->names[$language] = implode(', ', $parts);
2009-10-22 20:44:36 +01:00
$this->setCache(array('lat' => $lat,
'lon' => $lon),
$location);
2009-10-22 20:44:36 +01:00
// Success! We handled it, so no further processing
2009-10-22 20:44:36 +01:00
return false;
2009-10-22 20:44:36 +01:00
}
/**
* Human-readable name for a location
*
* Given a location, we try to retrieve a human-readable name
* in the target language.
*
* @param Location $location Location to get the name for
* @param string $language ISO code for language to find name in
* @param string &$name Place to put the name
*
* @return boolean whether to continue
*/
function onLocationNameLanguage($location, $language, &$name)
{
if ($location->location_ns != self::LOCATION_NS) {
2009-10-22 20:44:36 +01:00
// It's not one of our IDs... keep processing
return true;
}
$id = $location->location_id;
$n = $this->getCache(array('id' => $id,
2009-11-20 15:58:28 +00:00
'language' => $language));
if ($n !== false) {
2009-11-20 15:58:28 +00:00
$name = $n;
return false;
}
try {
$geonames = $this->getGeonames('hierarchy',
array('geonameId' => $id,
'lang' => $language));
} catch (Exception $e) {
$this->log(LOG_WARNING, "Error for ID $id: " . $e->getMessage());
return false;
}
2009-10-22 20:44:36 +01:00
if (count($geonames) == 0) {
$this->setCache(array('id' => $id,
'language' => $language),
null);
return false;
}
$parts = array();
2009-10-22 20:44:36 +01:00
foreach ($geonames as $level) {
if (in_array($level->fcode, array('PCLI', 'ADM1', 'PPL'))) {
$parts[] = (string)$level->name;
}
}
2009-10-22 20:44:36 +01:00
$last = $geonames[count($geonames)-1];
2009-10-22 20:44:36 +01:00
if (!in_array($level->fcode, array('PCLI', 'ADM1', 'PPL'))) {
$parts[] = (string)$last->name;
}
2009-10-22 20:44:36 +01:00
if (count($parts)) {
$name = implode(', ', array_reverse($parts));
$this->setCache(array('id' => $id,
'language' => $language),
$name);
2009-10-22 20:44:36 +01:00
}
return false;
2009-10-22 20:44:36 +01:00
}
2009-10-29 19:07:26 +00:00
/**
* Human-readable URL for a location
2009-10-29 19:07:26 +00:00
*
* Given a location, we try to retrieve a geonames.org URL.
*
* @param Location $location Location to get the url for
* @param string &$url Place to put the url
*
* @return boolean whether to continue
*/
function onLocationUrl($location, &$url)
{
if ($location->location_ns != self::LOCATION_NS) {
2009-10-29 19:07:26 +00:00
// It's not one of our IDs... keep processing
return true;
}
$url = 'http://www.geonames.org/' . $location->location_id;
// it's been filled, so don't process further.
2009-10-29 19:07:26 +00:00
return false;
}
2009-11-19 17:00:25 +00:00
/**
* Machine-readable name for a location
*
* Given a location, we try to retrieve a geonames.org URL.
*
* @param Location $location Location to get the url for
* @param string &$url Place to put the url
*
* @return boolean whether to continue
*/
function onLocationRdfUrl($location, &$url)
{
if ($location->location_ns != self::LOCATION_NS) {
// It's not one of our IDs... keep processing
return true;
}
$url = 'http://sws.geonames.org/' . $location->location_id . '/';
2009-11-19 17:00:25 +00:00
// it's been filled, so don't process further.
return false;
}
2009-11-20 15:58:28 +00:00
function getCache($attrs)
{
2010-09-06 14:56:45 +01:00
$c = Cache::instance();
2009-11-20 15:58:28 +00:00
if (empty($c)) {
2009-11-20 15:58:28 +00:00
return null;
}
$key = $this->cacheKey($attrs);
$value = $c->get($key);
return $value;
2009-11-20 15:58:28 +00:00
}
function setCache($attrs, $loc)
{
2010-09-06 14:56:45 +01:00
$c = Cache::instance();
2009-11-20 15:58:28 +00:00
if (empty($c)) {
2009-11-20 15:58:28 +00:00
return null;
}
$key = $this->cacheKey($attrs);
2009-11-20 15:58:28 +00:00
$result = $c->set($key, $loc, 0, time() + $this->expiry);
2009-11-20 15:58:28 +00:00
return $result;
2009-11-20 15:58:28 +00:00
}
function cacheKey($attrs)
{
$key = 'geonames:' .
implode(',', array_keys($attrs)) . ':'.
2010-09-06 15:03:51 +01:00
Cache::keyize(implode(',', array_values($attrs)));
if ($this->cachePrefix) {
return $this->cachePrefix . ':' . $key;
} else {
2010-09-06 15:07:43 +01:00
return Cache::key($key);
}
2009-11-20 15:58:28 +00:00
}
function wsUrl($method, $params)
{
if (!empty($this->username)) {
$params['username'] = $this->username;
}
if (!empty($this->token)) {
$params['token'] = $this->token;
}
$str = http_build_query($params, null, '&');
return 'http://'.$this->host.'/'.$method.'?'.$str;
}
function getGeonames($method, $params)
{
if ($this->lastTimeout && (time() - $this->lastTimeout < $this->timeoutWindow)) {
// TRANS: Exception thrown when a geo names service is not used because of a recent timeout.
throw new Exception(_m('Skipping due to recent web service timeout.'));
}
$client = HTTPClient::start();
$client->setConfig('connect_timeout', $this->timeout);
$client->setConfig('timeout', $this->timeout);
try {
$result = $client->get($this->wsUrl($method, $params));
} catch (Exception $e) {
common_log(LOG_ERR, __METHOD__ . ": " . $e->getMessage());
$this->lastTimeout = time();
throw $e;
}
if (!$result->isOk()) {
// TRANS: Exception thrown when a geo names service does not return an expected response.
// TRANS: %s is an HTTP error code.
throw new Exception(sprintf(_m('HTTP error code %s.'),$result->getStatus()));
}
$body = $result->getBody();
if (empty($body)) {
// TRANS: Exception thrown when a geo names service returns an empty body.
throw new Exception(_m('Empty HTTP body in response.'));
}
// This will throw an exception if the XML is mal-formed
$document = new SimpleXMLElement($body);
// No children, usually no results
$children = $document->children();
if (count($children) == 0) {
return array();
}
if (isset($document->status)) {
// TRANS: Exception thrown when a geo names service return a specific error number and error text.
// TRANS: %1$s is an error code, %2$s is an error message.
throw new Exception(sprintf(_m('Error #%1$s ("%2$s").'),$document->status['value'],$document->status['message']));
}
// Array of elements, >0 elements
return $document->geoname;
}
public function onPluginVersion(array &$versions): bool
{
$versions[] = array('name' => 'Geonames',
'version' => self::PLUGIN_VERSION,
'author' => 'Evan Prodromou',
'homepage' => GNUSOCIAL_ENGINE_REPO_URL . 'tree/master/plugins/Geonames',
'rawdescription' =>
// TRANS: Plugin description.
_m('Uses <a href="http://geonames.org/">Geonames</a> service to get human-readable '.
'names for locations based on user-provided lat/long pairs.'));
return true;
}
function canonical($coord)
{
$coord = rtrim($coord, "0");
$coord = rtrim($coord, ".");
return $coord;
}
2009-10-22 20:44:36 +01:00
}