X-Git-Url: https://git.openstreetmap.org./nominatim.git/blobdiff_plain/7471b09ed376f0870fad08234e89990e5cb74e02..98cb868ec119fca6eeec9145afe79bb9a621b4a4:/utils/update.php?ds=sidebyside diff --git a/utils/update.php b/utils/update.php index ac7b5be8..a561aaa3 100755 --- a/utils/update.php +++ b/utils/update.php @@ -47,7 +47,6 @@ 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; /* @@ -74,6 +73,7 @@ $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); @@ -82,7 +82,7 @@ $iCacheMemory = getCacheMemoryMB(); echo "WARNING: resetting cache memory to $iCacheMemory\n"; } - $sOsm2pgsqlCmd = CONST_Osm2pgsql_Binary.' -klas -C '.$iCacheMemory.' -O gazetteer -d '.$aDSNInfo['database']; + $sOsm2pgsqlCmd = CONST_Osm2pgsql_Binary.' -klas -C '.$iCacheMemory.' -O gazetteer -d '.$aDSNInfo['database'].' -P '.$aDSNInfo['port']; if (!is_null(CONST_Osm2pgsql_Flatnode_File)) { $sOsm2pgsqlCmd .= ' --flat-nodes '.CONST_Osm2pgsql_Flatnode_File; @@ -253,6 +253,13 @@ if ($aResult['deduplicate']) { + + $pgver = (float) CONST_Postgresql_Version; + if ($pgver < 9.3) { + echo "ERROR: deduplicate is only currently supported in postgresql 9.3"; + exit; + } + $oDB =& getDB(); $sSQL = 'select partition from country_name order by country_code'; $aPartitions = $oDB->getCol($sSQL); @@ -275,15 +282,15 @@ var_dump($aTokenSet, $sSQL); exit; } - + $aKeep = array_shift($aTokenSet); $iKeepID = $aKeep['word_id']; foreach($aTokenSet as $aRemove) { $sSQL = "update search_name set"; - $sSQL .= " name_vector = (name_vector - ".$aRemove['word_id'].")+".$iKeepID.","; - $sSQL .= " nameaddress_vector = (nameaddress_vector - ".$aRemove['word_id'].")+".$iKeepID; + $sSQL .= " name_vector = array_replace(name_vector,".$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)) @@ -293,7 +300,7 @@ } $sSQL = "update search_name set"; - $sSQL .= " nameaddress_vector = (nameaddress_vector - ".$aRemove['word_id'].")+".$iKeepID; + $sSQL .= " nameaddress_vector = array_replace(nameaddress_vector,".$aRemove['word_id'].",".$iKeepID.")"; $sSQL .= " where nameaddress_vector @> ARRAY[".$aRemove['word_id']."]"; $x = $oDB->query($sSQL); if (PEAR::isError($x)) @@ -303,7 +310,7 @@ } $sSQL = "update location_area_country set"; - $sSQL .= " keywords = (keywords - ".$aRemove['word_id'].")+".$iKeepID; + $sSQL .= " keywords = array_replace(keywords,".$aRemove['word_id'].",".$iKeepID.")"; $sSQL .= " where keywords @> ARRAY[".$aRemove['word_id']."]"; $x = $oDB->query($sSQL); if (PEAR::isError($x)) @@ -315,8 +322,7 @@ foreach ($aPartitions as $sPartition) { $sSQL = "update search_name_".$sPartition." set"; - $sSQL .= " name_vector = (name_vector - ".$aRemove['word_id'].")+".$iKeepID.","; - $sSQL .= " nameaddress_vector = (nameaddress_vector - ".$aRemove['word_id'].")+".$iKeepID; + $sSQL .= " name_vector = array_replace(name_vector,".$aRemove['word_id'].",".$iKeepID.")"; $sSQL .= " where name_vector @> ARRAY[".$aRemove['word_id']."]"; $x = $oDB->query($sSQL); if (PEAR::isError($x)) @@ -325,18 +331,8 @@ exit; } - $sSQL = "update search_name_".$sPartition." set"; - $sSQL .= " nameaddress_vector = (nameaddress_vector - ".$aRemove['word_id'].")+".$iKeepID; - $sSQL .= " where nameaddress_vector @> ARRAY[".$aRemove['word_id']."]"; - $x = $oDB->query($sSQL); - if (PEAR::isError($x)) - { - var_dump($x); - exit; - } - $sSQL = "update location_area_country set"; - $sSQL .= " keywords = (keywords - ".$aRemove['word_id'].")+".$iKeepID; + $sSQL .= " keywords = array_replace(keywords,".$aRemove['word_id'].",".$iKeepID.")"; $sSQL .= " where keywords @> ARRAY[".$aRemove['word_id']."]"; $x = $oDB->query($sSQL); if (PEAR::isError($x)) @@ -360,7 +356,8 @@ if ($aResult['index']) { - passthru(CONST_BasePath.'/nominatim/nominatim -i -d '.$aDSNInfo['database'].' -t '.$aResult['index-instances'].' -r '.$aResult['index-rank']); + if (!isset($aResult['index-instances'])) $aResult['index-instances'] = 1; + passthru(CONST_BasePath.'/nominatim/nominatim -i -d '.$aDSNInfo['database'].' -P '.$aDSNInfo['port'].' -t '.$aResult['index-instances'].' -r '.$aResult['index-rank']); } if ($aResult['import-osmosis'] || $aResult['import-osmosis-all']) @@ -377,7 +374,7 @@ $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'].' -t '.$aResult['index-instances']; + $sCMDIndex = $sBasePath.'/nominatim/nominatim -i -d '.$aDSNInfo['database'].' -P '.$aDSNInfo['port']; if (!$aResult['no-npi']) { $sCMDIndex .= '-F '; } @@ -417,8 +414,9 @@ exec($sCMDCheckReplicationLag, $aReplicationLag, $iErrorLevel); } // There are new replication files - use osmosis to download the file - echo "\nReplication Delay is ".$aReplicationLag[0]."\n"; + 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); @@ -431,9 +429,10 @@ } $iFileSize = filesize($sImportFile); $sBatchEnd = getosmosistimestamp($sOsmosisConfigDirectory); - echo "Completed for $sBatchEnd in ".round((time()-$fCMDStartTime)/60,2)." minutes\n"; $sSQL = "INSERT INTO import_osmosis_log values ('$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); @@ -448,10 +447,10 @@ echo "Error: $iErrorLevel\n"; exit($iErrorLevel); } - echo "Completed for $sBatchEnd in ".round((time()-$fCMDStartTime)/60,2)." minutes\n"; $sSQL = "INSERT INTO import_osmosis_log values ('$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); @@ -460,12 +459,20 @@ $sBatchEnd = getosmosistimestamp($sOsmosisConfigDirectory); // Index file - $sThisIndexCmd = $sCMDIndex; - $fCMDStartTime = time(); + if (!isset($aResult['index-instances'])) + { + if (getLoadAverage() < 24) + $iIndexInstances = 2; + else + $iIndexInstances = 1; + } else + $iIndexInstances = $aResult['index-instances']; + + $sThisIndexCmd = $sCMDIndex.' -t '.$iIndexInstances; + $fCMDStartTime = time(); if (!$aResult['no-npi']) { - $fCMDStartTime = time(); $iFileID = $oDB->getOne('select nextval(\'file\')'); if (PEAR::isError($iFileID)) { @@ -519,15 +526,16 @@ } } - echo "Completed for $sBatchEnd in ".round((time()-$fCMDStartTime)/60,2)." minutes\n"; $sSQL = "INSERT INTO import_osmosis_log values ('$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 "Completed for $sBatchEnd in ".round($fDuration/60,2)."\n"; + 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; if ( CONST_Replication_Update_Interval > 60 ) @@ -538,7 +546,7 @@ { $iSleep = max(0,CONST_Replication_Update_Interval-$fDuration); } - echo "Sleeping $iSleep seconds\n"; + echo date('Y-m-d H:i:s')." Sleeping $iSleep seconds\n"; sleep($iSleep); }