X-Git-Url: https://git.openstreetmap.org./nominatim.git/blobdiff_plain/af150bc63eb6893ea06577a1534beb85a5dbf29e..03039ebfa85a859bf80b5ed3125e34077d9a1df1:/utils/update.php?ds=sidebyside diff --git a/utils/update.php b/utils/update.php index cc275463..6caa7e4b 100755 --- a/utils/update.php +++ b/utils/update.php @@ -1,590 +1,409 @@ #!/usr/bin/php -Cq 1) - { - showUsage($aCMDOptions, true, 'Select either import of hourly or daily'); - } - - if (!isset($aResult['index-instances'])) $aResult['index-instances'] = 1; - if (!isset($aResult['index-rank'])) $aResult['index-rank'] = 0; - -/* - // Lock to prevent multiple copies running - if (exec('/bin/ps uww | grep '.basename(__FILE__).' | grep -v /dev/null | grep -v grep -c', $aOutput2, $iResult) > 1) - { - fail("Copy already running\n"); - } - if (!isset($aResult['max-load'])) $aResult['max-load'] = 1.9; - if (!isset($aResult['max-blocking'])) $aResult['max-blocking'] = 3; - if (getBlockingProcesses() > $aResult['max-blocking']) - { - fail("Too many blocking processes for import\n"); - } -*/ - - // Assume osm2pgsql is in the folder above - $sBasePath = dirname(dirname(__FILE__)); - - date_default_timezone_set('Etc/UTC'); - - $oDB =& getDB(); - - $aDSNInfo = DB::parseDSN(CONST_Database_DSN); - if (!isset($aDSNInfo['port']) || !$aDSNInfo['port']) $aDSNInfo['port'] = 5432; - - // cache memory to be used by osm2pgsql, should not be more than the available memory - $iCacheMemory = (isset($aResult['osm2pgsql-cache'])?$aResult['osm2pgsql-cache']:2000); - if ($iCacheMemory + 500 > getTotalMemoryMB()) - { - $iCacheMemory = getCacheMemoryMB(); - echo "WARNING: resetting cache memory to $iCacheMemory\n"; - } - $sOsm2pgsqlCmd = CONST_Osm2pgsql_Binary.' -klas --number-processes 1 -C '.$iCacheMemory.' -O gazetteer -d '.$aDSNInfo['database'].' -P '.$aDSNInfo['port']; - if (!is_null(CONST_Osm2pgsql_Flatnode_File)) - { - $sOsm2pgsqlCmd .= ' --flat-nodes '.CONST_Osm2pgsql_Flatnode_File; - } - - - $bFirst = true; - $bContinue = $aResult['import-all']; - while ($bContinue || $bFirst) - { - $bFirst = false; - - if ($aResult['import-hourly']) - { - // Mirror the hourly diffs - exec('wget --quiet --mirror -l 1 -P '.$sMirrorDir.' http://planet.openstreetmap.org/hourly'); - $sNextFile = $oDB->getOne('select TO_CHAR(lastimportdate,\'YYYYMMDDHH24\')||\'-\'||TO_CHAR(lastimportdate+\'1 hour\'::interval,\'YYYYMMDDHH24\')||\'.osc.gz\' from import_status'); - $sNextFile = $sMirrorDir.'planet.openstreetmap.org/hourly/'.$sNextFile; - $sUpdateSQL = 'update import_status set lastimportdate = lastimportdate+\'1 hour\'::interval'; - } - - if ($aResult['import-daily']) - { - // Mirror the daily diffs - exec('wget --quiet --mirror -l 1 -P '.$sMirrorDir.' http://planet.openstreetmap.org/daily'); - $sNextFile = $oDB->getOne('select TO_CHAR(lastimportdate,\'YYYYMMDD\')||\'-\'||TO_CHAR(lastimportdate+\'1 day\'::interval,\'YYYYMMDD\')||\'.osc.gz\' from import_status'); - $sNextFile = $sMirrorDir.'planet.openstreetmap.org/daily/'.$sNextFile; - $sUpdateSQL = 'update import_status set lastimportdate = lastimportdate::date + 1'; - } - - if (isset($aResult['import-diff'])) - { - // import diff directly (e.g. from osmosis --rri) - $sNextFile = $aResult['import-diff']; - if (!file_exists($sNextFile)) - { - fail("Cannot open $sNextFile\n"); - } - // Don't update the import status - we don't know what this file contains - $sUpdateSQL = 'update import_status set lastimportdate = now() where false'; - } - - // Missing file is not an error - it might not be created yet - if (($aResult['import-hourly'] || $aResult['import-daily'] || isset($aResult['import-diff'])) && file_exists($sNextFile)) - { - // Import the file - $sCMD = $sOsm2pgsqlCmd.' '.$sNextFile; - echo $sCMD."\n"; - exec($sCMD, $sJunk, $iErrorLevel); - - if ($iErrorLevel) - { - fail("Error from osm2pgsql, $iErrorLevel\n"); - } - - // Move the date onwards - $oDB->query($sUpdateSQL); - } - else - { - $bContinue = false; - } - } - - $bModifyXML = false; - $sModifyXMLstr = ''; - $bUseOSMApi = isset($aResult['import-from-main-api']) && $aResult['import-from-main-api']; - if (isset($aResult['import-file']) && $aResult['import-file']) - { - $bModifyXML = true; - } - if (isset($aResult['import-node']) && $aResult['import-node']) - { - $bModifyXML = true; - if ($bUseOSMApi) - { - $sModifyXMLstr = file_get_contents('http://www.openstreetmap.org/api/0.6/node/'.$aResult['import-node']); - } - else - { - $sModifyXMLstr = file_get_contents('http://overpass-api.de/api/interpreter?data=node('.$aResult['import-node'].');out%20meta;'); - } - } - if (isset($aResult['import-way']) && $aResult['import-way']) - { - $bModifyXML = true; - if ($bUseOSMApi) - { - $sCmd = 'http://www.openstreetmap.org/api/0.6/way/'.$aResult['import-way'].'/full'; - } - else - { - $sCmd = 'http://overpass-api.de/api/interpreter?data=(way('.$aResult['import-way'].');node(w););out%20meta;'; - } - $sModifyXMLstr = file_get_contents($sCmd); - } - if (isset($aResult['import-relation']) && $aResult['import-relation']) - { - $bModifyXML = true; - if ($bUseOSMApi) - { - $sModifyXMLstr = file_get_contents('http://www.openstreetmap.org/api/0.6/relation/'.$aResult['import-relation'].'/full'); - } - else - { - $sModifyXMLstr = file_get_contents('http://overpass-api.de/api/interpreter?data=((rel('.$aResult['import-relation'].');way(r);node(w));node(r));out%20meta;'); - } - } - if ($bModifyXML) - { - // derive change from normal osm file with osmosis - $sTemporaryFile = CONST_BasePath.'/data/osmosischange.osc'; - if (isset($aResult['import-file']) && $aResult['import-file']) - { - $sCMD = CONST_Osmosis_Binary.' --read-xml \''.$aResult['import-file'].'\' --read-empty --derive-change --write-xml-change '.$sTemporaryFile; - echo $sCMD."\n"; - exec($sCMD, $sJunk, $iErrorLevel); - if ($iErrorLevel) - { - fail("Error converting osm to osc, osmosis returned: $iErrorLevel\n"); - } - } - else - { - $aSpec = array( - 0 => array("pipe", "r"), // stdin - 1 => array("pipe", "w"), // stdout - 2 => array("pipe", "w") // stderr - ); - $sCMD = CONST_Osmosis_Binary.' --read-xml - --read-empty --derive-change --write-xml-change '.$sTemporaryFile; - echo $sCMD."\n"; - $hProc = proc_open($sCMD, $aSpec, $aPipes); - if (!is_resource($hProc)) - { - fail("Error converting osm to osc, osmosis failed\n"); - } - fwrite($aPipes[0], $sModifyXMLstr); - fclose($aPipes[0]); - $sOut = stream_get_contents($aPipes[1]); - if ($aResult['verbose']) echo $sOut; - fclose($aPipes[1]); - $sErrors = stream_get_contents($aPipes[2]); - if ($aResult['verbose']) echo $sErrors; - fclose($aPipes[2]); - if ($iError = proc_close($hProc)) - { - echo "Error converting osm to osc, osmosis returned: $iError\n"; - echo $sOut; - echo $sErrors; - exit(-1); - } - } - - // import generated change file - $sCMD = $sOsm2pgsqlCmd.' '.$sTemporaryFile; - echo $sCMD."\n"; - exec($sCMD, $sJunk, $iErrorLevel); - if ($iErrorLevel) - { - fail("osm2pgsql exited with error level $iErrorLevel\n"); - } - } - - if ($aResult['deduplicate']) - { - - $pgver = (float) CONST_Postgresql_Version; - if ($pgver < 9.3) { - fail("ERROR: deduplicate is only currently supported in postgresql 9.3"); - } - - $oDB =& getDB(); - $sSQL = 'select partition from country_name order by country_code'; - $aPartitions = $oDB->getCol($sSQL); - if (PEAR::isError($aPartitions)) - { - fail($aPartitions->getMessage()); +require_once(dirname(dirname(__FILE__)).'/settings/settings.php'); +require_once(CONST_BasePath.'/lib/init-cmd.php'); +ini_set('memory_limit', '800M'); + +$aCMDOptions += array( + 'Import / update / index osm data', + array('help', 'h', 0, 1, 0, 0, false, 'Show Help'), + array('quiet', 'q', 0, 1, 0, 0, 'bool', 'Quiet output'), + array('verbose', 'v', 0, 1, 0, 0, 'bool', 'Verbose output'), + + array('init-updates', '', 0, 1, 0, 0, 'bool', 'Set up database for updating'), + array('check-for-updates', '', 0, 1, 0, 0, 'bool', 'Check if new updates are available'), + array('import-osmosis', '', 0, 1, 0, 0, 'bool', 'Import updates once'), + array('import-osmosis-all', '', 0, 1, 0, 0, 'bool', 'Import updates forever'), + array('no-index', '', 0, 1, 0, 0, 'bool', 'Do not index the new data'), + + array('calculate-postcodes', '', 0, 1, 0, 0, 'bool', 'Update postcode centroid table'), + + array('import-file', '', 0, 1, 1, 1, 'realpath', 'Re-import data from an OSM file'), + array('import-diff', '', 0, 1, 1, 1, 'realpath', 'Import a diff (osc) file from local file system'), + array('osm2pgsql-cache', '', 0, 1, 1, 1, 'int', 'Cache size used by osm2pgsql'), + + array('import-node', '', 0, 1, 1, 1, 'int', 'Re-import node'), + array('import-way', '', 0, 1, 1, 1, 'int', 'Re-import way'), + array('import-relation', '', 0, 1, 1, 1, 'int', 'Re-import relation'), + array('import-from-main-api', '', 0, 1, 0, 0, 'bool', 'Use OSM API instead of Overpass to download objects'), + + array('index', '', 0, 1, 0, 0, 'bool', 'Index'), + array('index-rank', '', 0, 1, 1, 1, 'int', 'Rank to start indexing from'), + array('index-instances', '', 0, 1, 1, 1, 'int', 'Number of indexing instances (threads)'), + + array('deduplicate', '', 0, 1, 0, 0, 'bool', 'Deduplicate tokens'), + array('recompute-word-counts', '', 0, 1, 0, 0, 'bool', 'Compute frequency of full-word search terms'), + array('no-npi', '', 0, 1, 0, 0, 'bool', '(obsolete)'), + ); +getCmdOpt($_SERVER['argv'], $aCMDOptions, $aResult, true, true); + +if (!isset($aResult['index-instances'])) $aResult['index-instances'] = 1; +if (!isset($aResult['index-rank'])) $aResult['index-rank'] = 0; + +date_default_timezone_set('Etc/UTC'); + +$oDB =& getDB(); + +$aDSNInfo = DB::parseDSN(CONST_Database_DSN); +if (!isset($aDSNInfo['port']) || !$aDSNInfo['port']) $aDSNInfo['port'] = 5432; + +// cache memory to be used by osm2pgsql, should not be more than the available memory +$iCacheMemory = (isset($aResult['osm2pgsql-cache'])?$aResult['osm2pgsql-cache']:2000); +if ($iCacheMemory + 500 > getTotalMemoryMB()) { + $iCacheMemory = getCacheMemoryMB(); + echo "WARNING: resetting cache memory to $iCacheMemory\n"; +} +$sOsm2pgsqlCmd = CONST_Osm2pgsql_Binary.' -klas --number-processes 1 -C '.$iCacheMemory.' -O gazetteer -d '.$aDSNInfo['database'].' -P '.$aDSNInfo['port']; +if (!is_null(CONST_Osm2pgsql_Flatnode_File) && CONST_Osm2pgsql_Flatnode_File) { + $sOsm2pgsqlCmd .= ' --flat-nodes '.CONST_Osm2pgsql_Flatnode_File; +} + +if ($aResult['init-updates']) { + // sanity check that the replication URL is correct + $sBaseState = file_get_contents(CONST_Replication_Url.'/state.txt'); + if ($sBaseState === false) { + echo "\nCannot find state.txt file at the configured replication URL.\n"; + echo "Does the URL point to a directory containing OSM update data?\n\n"; + fail('replication URL not reachable.'); + } + // sanity check for pyosmium-get-changes + if (!CONST_Pyosmium_Binary) { + echo "\nCONST_Pyosmium_Binary not configured.\n"; + echo "You need to install pyosmium and set up the path to pyosmium-get-changes\n"; + echo "in your local settings file.\n\n"; + fail('CONST_Pyosmium_Binary not configured'); + } + $aOutput = 0; + $sCmd = CONST_Pyosmium_Binary.' --help'; + exec($sCmd, $aOutput, $iRet); + if ($iRet != 0) { + echo "Cannot execute pyosmium-get-changes.\n"; + echo "Make sure you have pyosmium installed correctly\n"; + echo "and have set up CONST_Pyosmium_Binary to point to pyosmium-get-changes.\n"; + fail('pyosmium-get-changes not found or not usable'); + } + $sSetup = CONST_InstallPath.'/utils/setup.php'; + $iRet = -1; + passthru($sSetup.' --create-functions --enable-diff-updates', $iRet); + if ($iRet != 0) { + fail('Error running setup script'); + } + + $sDatabaseDate = getDatabaseDate($oDB); + if ($sDatabaseDate === false) { + fail('Cannot determine date of database.'); + } + $sWindBack = strftime('%Y-%m-%dT%H:%M:%SZ', strtotime($sDatabaseDate) - (3*60*60)); + + // get the appropriate state id + $aOutput = 0; + $sCmd = CONST_Pyosmium_Binary.' -D '.$sWindBack.' --server '.CONST_Replication_Url; + exec($sCmd, $aOutput, $iRet); + if ($iRet != 0 || $aOutput[0] == 'None') { + fail('Error running pyosmium tools'); + } + + pg_query($oDB->connection, 'TRUNCATE import_status'); + $sSQL = "INSERT INTO import_status (lastimportdate, sequence_id, indexed) VALUES('"; + $sSQL .= $sDatabaseDate."',".$aOutput[0].', true)'; + if (!pg_query($oDB->connection, $sSQL)) { + fail('Could not enter sequence into database.'); + } + + echo "Done. Database updates will start at sequence $aOutput[0] ($sWindBack)\n"; +} + +if ($aResult['check-for-updates']) { + $aLastState = chksql($oDB->getRow('SELECT sequence_id FROM import_status')); + + if (!$aLastState['sequence_id']) { + fail('Updates not set up. Please run ./utils/update.php --init-updates.'); + } + + system(CONST_BasePath.'/utils/check_server_for_updates.py '.CONST_Replication_Url.' '.$aLastState['sequence_id'], $iRet); + exit($iRet); +} + +if (isset($aResult['import-diff']) || isset($aResult['import-file'])) { + // import diffs and files directly (e.g. from osmosis --rri) + $sNextFile = isset($aResult['import-diff']) ? $aResult['import-diff'] : $aResult['import-file']; + + if (!file_exists($sNextFile)) { + fail("Cannot open $sNextFile\n"); + } + + // Import the file + $sCMD = $sOsm2pgsqlCmd.' '.$sNextFile; + echo $sCMD."\n"; + exec($sCMD, $sJunk, $iErrorLevel); + + if ($iErrorLevel) { + fail("Error from osm2pgsql, $iErrorLevel\n"); + } + + // Don't update the import status - we don't know what this file contains +} + +if ($aResult['calculate-postcodes']) { + info('Update postcodes centroids'); + $sTemplate = file_get_contents(CONST_BasePath.'/sql/update-postcodes.sql'); + runSQLScript($sTemplate, true, true); +} + +$sTemporaryFile = CONST_BasePath.'/data/osmosischange.osc'; +$bHaveDiff = false; +$bUseOSMApi = isset($aResult['import-from-main-api']) && $aResult['import-from-main-api']; +$sContentURL = ''; +if (isset($aResult['import-node']) && $aResult['import-node']) { + if ($bUseOSMApi) { + $sContentURL = 'https://www.openstreetmap.org/api/0.6/node/'.$aResult['import-node']; + } else { + $sContentURL = 'https://overpass-api.de/api/interpreter?data=node('.$aResult['import-node'].');out%20meta;'; + } +} + +if (isset($aResult['import-way']) && $aResult['import-way']) { + if ($bUseOSMApi) { + $sContentURL = 'https://www.openstreetmap.org/api/0.6/way/'.$aResult['import-way'].'/full'; + } else { + $sContentURL = 'https://overpass-api.de/api/interpreter?data=(way('.$aResult['import-way'].');node(w););out%20meta;'; + } +} + +if (isset($aResult['import-relation']) && $aResult['import-relation']) { + if ($bUseOSMApi) { + $sContentURLsModifyXMLstr = 'https://www.openstreetmap.org/api/0.6/relation/'.$aResult['import-relation'].'/full'; + } else { + $sContentURL = 'https://overpass-api.de/api/interpreter?data=((rel('.$aResult['import-relation'].');way(r);node(w));node(r));out%20meta;'; + } +} + +if ($sContentURL) { + file_put_contents($sTemporaryFile, file_get_contents($sContentURL)); + $bHaveDiff = true; +} + +if ($bHaveDiff) { + // import generated change file + $sCMD = $sOsm2pgsqlCmd.' '.$sTemporaryFile; + echo $sCMD."\n"; + exec($sCMD, $sJunk, $iErrorLevel); + if ($iErrorLevel) { + fail("osm2pgsql exited with error level $iErrorLevel\n"); + } +} + +if ($aResult['deduplicate']) { + $oDB =& getDB(); + + if (getPostgresVersion($oDB) < 9.3) { + fail('ERROR: deduplicate is only currently supported in postgresql 9.3'); + } + + $sSQL = 'select partition from country_name order by country_code'; + $aPartitions = chksql($oDB->getCol($sSQL)); + $aPartitions[] = 0; + + // we don't care about empty search_name_* partitions, they can't contain mentions of duplicates + foreach ($aPartitions as $i => $sPartition) { + $sSQL = 'select count(*) from search_name_'.$sPartition; + $nEntries = chksql($oDB->getOne($sSQL)); + if ($nEntries == 0) { + unset($aPartitions[$i]); + } + } + + $sSQL = "select word_token,count(*) from word where substr(word_token, 1, 1) = ' '"; + $sSQL .= ' and class is null and type is null and country_code is null'; + $sSQL .= ' group by word_token having count(*) > 1 order by word_token'; + $aDuplicateTokens = chksql($oDB->getAll($sSQL)); + foreach ($aDuplicateTokens as $aToken) { + if (trim($aToken['word_token']) == '' || trim($aToken['word_token']) == '-') continue; + echo 'Deduping '.$aToken['word_token']."\n"; + $sSQL = 'select word_id,'; + $sSQL .= ' (select count(*) from search_name where nameaddress_vector @> ARRAY[word_id]) as num'; + $sSQL .= " from word where word_token = '".$aToken['word_token']; + $sSQL .= "' and class is null and type is null and country_code is null order by num desc"; + $aTokenSet = chksql($oDB->getAll($sSQL)); + + $aKeep = array_shift($aTokenSet); + $iKeepID = $aKeep['word_id']; + + foreach ($aTokenSet as $aRemove) { + $sSQL = 'update search_name set'; + $sSQL .= ' name_vector = array_replace(name_vector,'.$aRemove['word_id'].','.$iKeepID.'),'; + $sSQL .= ' nameaddress_vector = array_replace(nameaddress_vector,'.$aRemove['word_id'].','.$iKeepID.')'; + $sSQL .= ' where name_vector @> ARRAY['.$aRemove['word_id'].']'; + chksql($oDB->query($sSQL)); + + $sSQL = 'update search_name set'; + $sSQL .= ' nameaddress_vector = array_replace(nameaddress_vector,'.$aRemove['word_id'].','.$iKeepID.')'; + $sSQL .= ' where nameaddress_vector @> ARRAY['.$aRemove['word_id'].']'; + chksql($oDB->query($sSQL)); + + $sSQL = 'update location_area_country set'; + $sSQL .= ' keywords = array_replace(keywords,'.$aRemove['word_id'].','.$iKeepID.')'; + $sSQL .= ' where keywords @> ARRAY['.$aRemove['word_id'].']'; + chksql($oDB->query($sSQL)); + + foreach ($aPartitions as $sPartition) { + $sSQL = 'update search_name_'.$sPartition.' set'; + $sSQL .= ' name_vector = array_replace(name_vector,'.$aRemove['word_id'].','.$iKeepID.')'; + $sSQL .= ' where name_vector @> ARRAY['.$aRemove['word_id'].']'; + chksql($oDB->query($sSQL)); + + $sSQL = 'update location_area_country set'; + $sSQL .= ' keywords = array_replace(keywords,'.$aRemove['word_id'].','.$iKeepID.')'; + $sSQL .= ' where keywords @> ARRAY['.$aRemove['word_id'].']'; + chksql($oDB->query($sSQL)); + } + + $sSQL = 'delete from word where word_id = '.$aRemove['word_id']; + chksql($oDB->query($sSQL)); + } + } +} + +if ($aResult['recompute-word-counts']) { + info('Recompute frequency of full-word search terms'); + $sTemplate = file_get_contents(CONST_BasePath.'/sql/words_from_search_name.sql'); + runSQLScript($sTemplate, true, true); +} + +if ($aResult['index']) { + passthru(CONST_InstallPath.'/nominatim/nominatim -i -d '.$aDSNInfo['database'].' -P '.$aDSNInfo['port'].' -t '.$aResult['index-instances'].' -r '.$aResult['index-rank']); +} + +if ($aResult['import-osmosis'] || $aResult['import-osmosis-all']) { + // + if (strpos(CONST_Replication_Url, 'download.geofabrik.de') !== false && CONST_Replication_Update_Interval < 86400) { + fail('Error: Update interval too low for download.geofabrik.de. ' . + "Please check install documentation (http://nominatim.org/release-docs/latest/Import-and-Update#setting-up-the-update-process)\n"); + } + + $sImportFile = CONST_InstallPath.'/osmosischange.osc'; + $sCMDDownload = CONST_Pyosmium_Binary.' --server '.CONST_Replication_Url.' -o '.$sImportFile.' -s '.CONST_Replication_Max_Diff_size; + $sCMDImport = $sOsm2pgsqlCmd.' '.$sImportFile; + $sCMDIndex = CONST_InstallPath.'/nominatim/nominatim -i -d '.$aDSNInfo['database'].' -P '.$aDSNInfo['port'].' -t '.$aResult['index-instances']; + + while (true) { + $fStartTime = time(); + $aLastState = chksql($oDB->getRow('SELECT *, EXTRACT (EPOCH FROM lastimportdate) as unix_ts FROM import_status')); + + if (!$aLastState['sequence_id']) { + echo "Updates not set up. Please run ./utils/update.php --init-updates.\n"; + exit(1); + } + + echo 'Currently at sequence '.$aLastState['sequence_id'].' ('.$aLastState['lastimportdate'].') - '.$aLastState['indexed']." indexed\n"; + + $sBatchEnd = $aLastState['lastimportdate']; + $iEndSequence = $aLastState['sequence_id']; + + if ($aLastState['indexed'] == 't') { + // Sleep if the update interval has not yet been reached. + $fNextUpdate = $aLastState['unix_ts'] + CONST_Replication_Update_Interval; + if ($fNextUpdate > $fStartTime) { + $iSleepTime = $fNextUpdate - $fStartTime; + echo "Waiting for next update for $iSleepTime sec."; + sleep($iSleepTime); + } + + // Download the next batch of changes. + do { + $fCMDStartTime = time(); + $iNextSeq = (int) $aLastState['sequence_id']; + unset($aOutput); + echo "$sCMDDownload -I $iNextSeq\n"; + if (file_exists($sImportFile)) { + unlink($sImportFile); } - $aPartitions[] = 0; - - $sSQL = "select word_token,count(*) from word where substr(word_token, 1, 1) = ' ' and class is null and type is null and country_code is null group by word_token having count(*) > 1 order by word_token"; - $aDuplicateTokens = $oDB->getAll($sSQL); - foreach($aDuplicateTokens as $aToken) - { - if (trim($aToken['word_token']) == '' || trim($aToken['word_token']) == '-') continue; - echo "Deduping ".$aToken['word_token']."\n"; - $sSQL = "select word_id,(select count(*) from search_name where nameaddress_vector @> ARRAY[word_id]) as num from word where word_token = '".$aToken['word_token']."' and class is null and type is null and country_code is null order by num desc"; - $aTokenSet = $oDB->getAll($sSQL); - if (PEAR::isError($aTokenSet)) - { - var_dump($aTokenSet, $sSQL); - exit(1); - } - - $aKeep = array_shift($aTokenSet); - $iKeepID = $aKeep['word_id']; - - foreach($aTokenSet as $aRemove) - { - $sSQL = "update search_name set"; - $sSQL .= " name_vector = array_replace(name_vector,".$aRemove['word_id'].",".$iKeepID."),"; - $sSQL .= " nameaddress_vector = array_replace(nameaddress_vector,".$aRemove['word_id'].",".$iKeepID.")"; - $sSQL .= " where name_vector @> ARRAY[".$aRemove['word_id']."]"; - $x = $oDB->query($sSQL); - if (PEAR::isError($x)) - { - var_dump($x); - exit(1); - } - - $sSQL = "update search_name set"; - $sSQL .= " nameaddress_vector = array_replace(nameaddress_vector,".$aRemove['word_id'].",".$iKeepID.")"; - $sSQL .= " where nameaddress_vector @> ARRAY[".$aRemove['word_id']."]"; - $x = $oDB->query($sSQL); - if (PEAR::isError($x)) - { - var_dump($x); - exit(1); - } - - $sSQL = "update location_area_country set"; - $sSQL .= " keywords = array_replace(keywords,".$aRemove['word_id'].",".$iKeepID.")"; - $sSQL .= " where keywords @> ARRAY[".$aRemove['word_id']."]"; - $x = $oDB->query($sSQL); - if (PEAR::isError($x)) - { - var_dump($x); - exit(1); - } - - foreach ($aPartitions as $sPartition) - { - $sSQL = "update search_name_".$sPartition." set"; - $sSQL .= " name_vector = array_replace(name_vector,".$aRemove['word_id'].",".$iKeepID.")"; - $sSQL .= " where name_vector @> ARRAY[".$aRemove['word_id']."]"; - $x = $oDB->query($sSQL); - if (PEAR::isError($x)) - { - var_dump($x); - exit(1); - } - - $sSQL = "update location_area_country set"; - $sSQL .= " keywords = array_replace(keywords,".$aRemove['word_id'].",".$iKeepID.")"; - $sSQL .= " where keywords @> ARRAY[".$aRemove['word_id']."]"; - $x = $oDB->query($sSQL); - if (PEAR::isError($x)) - { - var_dump($x); - exit(1); - } - } - - $sSQL = "delete from word where word_id = ".$aRemove['word_id']; - $x = $oDB->query($sSQL); - if (PEAR::isError($x)) - { - var_dump($x); - exit(1); - } - } - - } - } - - if ($aResult['index']) - { - passthru(CONST_BasePath.'/nominatim/nominatim -i -d '.$aDSNInfo['database'].' -P '.$aDSNInfo['port'].' -t '.$aResult['index-instances'].' -r '.$aResult['index-rank']); - } - - if ($aResult['import-osmosis'] || $aResult['import-osmosis-all']) - { - - if (strpos(CONST_Replication_Url, 'download.geofabrik.de') !== false && CONST_Replication_Update_Interval < 86400) { - fail("Error: Update interval too low for download.geofabrik.de. Please check install documentation (http://wiki.openstreetmap.org/wiki/Nominatim/Installation#Updates)\n"); - } - - $sImportFile = CONST_BasePath.'/data/osmosischange.osc'; - $sOsmosisCMD = CONST_Osmosis_Binary; - $sOsmosisConfigDirectory = CONST_BasePath.'/settings'; - $sCMDDownload = $sOsmosisCMD.' --read-replication-interval workingDirectory='.$sOsmosisConfigDirectory.' --simplify-change --write-xml-change '.$sImportFile; - $sCMDCheckReplicationLag = $sOsmosisCMD.' -q --read-replication-lag workingDirectory='.$sOsmosisConfigDirectory; - $sCMDImport = $sOsm2pgsqlCmd.' '.$sImportFile; - $sCMDIndex = $sBasePath.'/nominatim/nominatim -i -d '.$aDSNInfo['database'].' -P '.$aDSNInfo['port'].' -t '.$aResult['index-instances']; - if (!$aResult['no-npi']) { - $sCMDIndex .= '-F '; - } - while(true) - { - $fStartTime = time(); - $iFileSize = 1001; - - // Logic behind this is that osm2pgsql locks the database quite a bit - // So it is better to import lots of small files - // But indexing works most efficiently on large amounts of data - // So do lots of small imports and a BIG index - -// while($aResult['import-osmosis-all'] && $iFileSize > 1000) -// { - if (!file_exists($sImportFile)) - { - // First check if there are new updates published (except for minutelies - there's always new diffs to process) - if ( CONST_Replication_Update_Interval > 60 ) - { - - unset($aReplicationLag); - exec($sCMDCheckReplicationLag, $aReplicationLag, $iErrorLevel); - while ($iErrorLevel > 0 || $aReplicationLag[0] < 1) - { - if ($iErrorLevel) - { - echo "Error: $iErrorLevel. "; - echo "Re-trying: ".$sCMDCheckReplicationLag." in ".CONST_Replication_Recheck_Interval." secs\n"; - } - else - { - echo "."; - } - sleep(CONST_Replication_Recheck_Interval); - unset($aReplicationLag); - exec($sCMDCheckReplicationLag, $aReplicationLag, $iErrorLevel); - } - // There are new replication files - use osmosis to download the file - echo "\n".date('Y-m-d H:i:s')." Replication Delay is ".$aReplicationLag[0]."\n"; - } - $fStartTime = time(); - $fCMDStartTime = time(); - echo $sCMDDownload."\n"; - exec($sCMDDownload, $sJunk, $iErrorLevel); - while ($iErrorLevel > 0) - { - echo "Error: $iErrorLevel\n"; - sleep(60); - echo 'Re-trying: '.$sCMDDownload."\n"; - exec($sCMDDownload, $sJunk, $iErrorLevel); - } - $iFileSize = filesize($sImportFile); - $sBatchEnd = getosmosistimestamp($sOsmosisConfigDirectory); - $sSQL = "INSERT INTO import_osmosis_log values ('$sBatchEnd',$iFileSize,'".date('Y-m-d H:i:s',$fCMDStartTime)."','".date('Y-m-d H:i:s')."','osmosis')"; - var_Dump($sSQL); - $oDB->query($sSQL); - echo date('Y-m-d H:i:s')." Completed osmosis step for $sBatchEnd in ".round((time()-$fCMDStartTime)/60,2)." minutes\n"; - } - - $iFileSize = filesize($sImportFile); - $sBatchEnd = getosmosistimestamp($sOsmosisConfigDirectory); - - // Import the file - $fCMDStartTime = time(); - echo $sCMDImport."\n"; - exec($sCMDImport, $sJunk, $iErrorLevel); - if ($iErrorLevel) - { - echo "Error: $iErrorLevel\n"; - exit($iErrorLevel); - } - $sSQL = "INSERT INTO import_osmosis_log values ('$sBatchEnd',$iFileSize,'".date('Y-m-d H:i:s',$fCMDStartTime)."','".date('Y-m-d H:i:s')."','osm2pgsql')"; - var_Dump($sSQL); - $oDB->query($sSQL); - echo date('Y-m-d H:i:s')." Completed osm2pgsql step for $sBatchEnd in ".round((time()-$fCMDStartTime)/60,2)." minutes\n"; - - // Archive for debug? - unlink($sImportFile); -// } - - $sBatchEnd = getosmosistimestamp($sOsmosisConfigDirectory); - - // Index file - $sThisIndexCmd = $sCMDIndex; - $fCMDStartTime = time(); - - if (!$aResult['no-npi']) - { - $iFileID = $oDB->getOne('select nextval(\'file\')'); - if (PEAR::isError($iFileID)) - { - echo $iFileID->getMessage()."\n"; - exit(-1); - } - $sFileDir = CONST_BasePath.'/export/diff/'; - $sFileDir .= str_pad(floor($iFileID/1000000), 3, '0', STR_PAD_LEFT); - $sFileDir .= '/'.str_pad(floor($iFileID/1000) % 1000, 3, '0', STR_PAD_LEFT); - - if (!is_dir($sFileDir)) mkdir($sFileDir, 0777, true); - $sThisIndexCmd .= $sFileDir; - $sThisIndexCmd .= '/'.str_pad($iFileID % 1000, 3, '0', STR_PAD_LEFT); - $sThisIndexCmd .= ".npi.out"; - - preg_match('#^([0-9]{4})-([0-9]{2})-([0-9]{2})#', $sBatchEnd, $aBatchMatch); - $sFileDir = CONST_BasePath.'/export/index/'; - $sFileDir .= $aBatchMatch[1].'/'.$aBatchMatch[2]; - - if (!is_dir($sFileDir)) mkdir($sFileDir, 0777, true); - file_put_contents($sFileDir.'/'.$aBatchMatch[3].'.idx', "$sBatchEnd\t$iFileID\n", FILE_APPEND); - } - - if (!$aResult['no-index']) - { - echo "$sThisIndexCmd\n"; - exec($sThisIndexCmd, $sJunk, $iErrorLevel); - if ($iErrorLevel) - { - echo "Error: $iErrorLevel\n"; - exit($iErrorLevel); - } - - if (!$aResult['no-npi']) - { - $sFileDir = CONST_BasePath.'/export/diff/'; - $sFileDir .= str_pad(floor($iFileID/1000000), 3, '0', STR_PAD_LEFT); - $sFileDir .= '/'.str_pad(floor($iFileID/1000) % 1000, 3, '0', STR_PAD_LEFT); - - $sThisIndexCmd = 'bzip2 -z9 '.$sFileDir.'/'.str_pad($iFileID % 1000, 3, '0', STR_PAD_LEFT).".npi.out"; - echo "$sThisIndexCmd\n"; - exec($sThisIndexCmd, $sJunk, $iErrorLevel); - if ($iErrorLevel) - { - echo "Error: $iErrorLevel\n"; - exit($iErrorLevel); - } - - rename($sFileDir.'/'.str_pad($iFileID % 1000, 3, '0', STR_PAD_LEFT).".npi.out.bz2", - $sFileDir.'/'.str_pad($iFileID % 1000, 3, '0', STR_PAD_LEFT).".npi.bz2"); - } - } - - $sSQL = "INSERT INTO import_osmosis_log values ('$sBatchEnd',$iFileSize,'".date('Y-m-d H:i:s',$fCMDStartTime)."','".date('Y-m-d H:i:s')."','index')"; - var_Dump($sSQL); - $oDB->query($sSQL); - echo date('Y-m-d H:i:s')." Completed index step for $sBatchEnd in ".round((time()-$fCMDStartTime)/60,2)." minutes\n"; - - $sSQL = "update import_status set lastimportdate = '$sBatchEnd'"; - $oDB->query($sSQL); - - $fDuration = time() - $fStartTime; - echo date('Y-m-d H:i:s')." Completed all for $sBatchEnd in ".round($fDuration/60,2)." minutes\n"; - if (!$aResult['import-osmosis-all']) exit(0); - - if ( CONST_Replication_Update_Interval > 60 ) - { - $iSleep = max(0,(strtotime($sBatchEnd)+CONST_Replication_Update_Interval-time())); - } - else - { - $iSleep = max(0,CONST_Replication_Update_Interval-$fDuration); - } - echo date('Y-m-d H:i:s')." Sleeping $iSleep seconds\n"; - sleep($iSleep); - } - - } - - if ($aResult['import-npi-all']) - { - $iNPIID = $oDB->getOne('select max(npiid) from import_npi_log'); - if (PEAR::isError($iNPIID)) - { - var_dump($iNPIID); - exit(1); - } - $sConfigDirectory = CONST_BasePath.'/settings'; - $sCMDImportTemplate = $sBasePath.'/nominatim/nominatim -d gazetteer -P 5433 -I -T '.$sBasePath.'/nominatim/partitionedtags.def -F '; - while(true) - { - $fStartTime = time(); - - $iNPIID++; - - $sImportFile = CONST_BasePath.'/export/diff/'; - $sImportFile .= str_pad(floor($iNPIID/1000000), 3, '0', STR_PAD_LEFT); - $sImportFile .= '/'.str_pad(floor($iNPIID/1000) % 1000, 3, '0', STR_PAD_LEFT); - $sImportFile .= '/'.str_pad($iNPIID % 1000, 3, '0', STR_PAD_LEFT); - $sImportFile .= ".npi"; - while(!file_exists($sImportFile) && !file_exists($sImportFile.'.bz2')) - { - echo "sleep (waiting for $sImportFile)\n"; - sleep(10); - } - if (file_exists($sImportFile.'.bz2')) $sImportFile .= '.bz2'; - - $iFileSize = filesize($sImportFile); - - // Import the file - $fCMDStartTime = time(); - $sCMDImport = $sCMDImportTemplate . $sImportFile; - echo $sCMDImport."\n"; - exec($sCMDImport, $sJunk, $iErrorLevel); - if ($iErrorLevel) - { - fail("Error: $iErrorLevel\n"); - } - $sBatchEnd = $iNPIID; - echo "Completed for $sBatchEnd in ".round((time()-$fCMDStartTime)/60,2)." minutes\n"; - $sSQL = "INSERT INTO import_npi_log values ($iNPIID, null, $iFileSize,'".date('Y-m-d H:i:s',$fCMDStartTime)."','".date('Y-m-d H:i:s')."','import')"; - var_Dump($sSQL); - $oDB->query($sSQL); - } - - } - - function getosmosistimestamp($sOsmosisConfigDirectory) - { - $sStateFile = file_get_contents($sOsmosisConfigDirectory.'/state.txt'); - preg_match('#timestamp=(.+)#', $sStateFile, $aResult); - return str_replace('\:',':',$aResult[1]); - } + exec($sCMDDownload.' -I '.$iNextSeq, $aOutput, $iResult); + + if ($iResult == 3) { + echo 'No new updates. Sleeping for '.CONST_Replication_Recheck_Interval." sec.\n"; + sleep(CONST_Replication_Recheck_Interval); + } elseif ($iResult != 0) { + echo 'ERROR: updates failed.'; + exit($iResult); + } else { + $iEndSequence = (int)$aOutput[0]; + } + } while ($iResult); + + // get the newest object from the diff file + $sBatchEnd = 0; + $iRet = 0; + exec(CONST_BasePath.'/utils/osm_file_date.py '.$sImportFile, $sBatchEnd, $iRet); + if ($iRet == 5) { + echo "Diff file is empty. skipping import.\n"; + if (!$aResult['import-osmosis-all']) { + exit(0); + } else { + continue; + } + } + if ($iRet != 0) { + fail('Error getting date from diff file.'); + } + $sBatchEnd = $sBatchEnd[0]; + + // Import the file + $fCMDStartTime = time(); + echo $sCMDImport."\n"; + unset($sJunk); + exec($sCMDImport, $sJunk, $iErrorLevel); + if ($iErrorLevel) { + echo "Error executing osm2pgsql: $iErrorLevel\n"; + exit($iErrorLevel); + } + + // write the update logs + $iFileSize = filesize($sImportFile); + $sSQL = 'INSERT INTO import_osmosis_log'; + $sSQL .= '(batchend, batchseq, batchsize, starttime, endtime, event)'; + $sSQL .= " values ('$sBatchEnd',$iEndSequence,$iFileSize,'"; + $sSQL .= date('Y-m-d H:i:s', $fCMDStartTime)."','"; + $sSQL .= date('Y-m-d H:i:s')."','import')"; + var_Dump($sSQL); + chksql($oDB->query($sSQL)); + + // update the status + $sSQL = "UPDATE import_status SET lastimportdate = '$sBatchEnd', indexed=false, sequence_id = $iEndSequence"; + var_Dump($sSQL); + chksql($oDB->query($sSQL)); + echo date('Y-m-d H:i:s')." Completed download step for $sBatchEnd in ".round((time()-$fCMDStartTime)/60, 2)." minutes\n"; + } + + // Index file + if (!$aResult['no-index']) { + $sThisIndexCmd = $sCMDIndex; + $fCMDStartTime = time(); + + echo "$sThisIndexCmd\n"; + exec($sThisIndexCmd, $sJunk, $iErrorLevel); + if ($iErrorLevel) { + echo "Error: $iErrorLevel\n"; + exit($iErrorLevel); + } + + $sSQL = 'INSERT INTO import_osmosis_log'; + $sSQL .= '(batchend, batchseq, batchsize, starttime, endtime, event)'; + $sSQL .= " values ('$sBatchEnd',$iEndSequence,$iFileSize,'"; + $sSQL .= date('Y-m-d H:i:s', $fCMDStartTime)."','"; + $sSQL .= date('Y-m-d H:i:s')."','index')"; + var_Dump($sSQL); + $oDB->query($sSQL); + echo date('Y-m-d H:i:s')." Completed index step for $sBatchEnd in ".round((time()-$fCMDStartTime)/60, 2)." minutes\n"; + + $sSQL = 'update import_status set indexed = true'; + $oDB->query($sSQL); + } + + $fDuration = time() - $fStartTime; + echo date('Y-m-d H:i:s')." Completed all for $sBatchEnd in ".round($fDuration/60, 2)." minutes\n"; + if (!$aResult['import-osmosis-all']) exit(0); + } +}