X-Git-Url: https://git.openstreetmap.org./nominatim.git/blobdiff_plain/7dbbc43a85dd6c46110493e63256149cbc89491c..fc96bee063c26ade2cdc89f255de580de9d43009:/utils/setup.php?ds=inline diff --git a/utils/setup.php b/utils/setup.php index a203db7b..01bf1134 100755 --- a/utils/setup.php +++ b/utils/setup.php @@ -22,21 +22,24 @@ array('create-functions', '', 0, 1, 0, 0, 'bool', 'Create functions'), array('enable-diff-updates', '', 0, 1, 0, 0, 'bool', 'Turn on the code required to make diff updates work'), array('enable-debug-statements', '', 0, 1, 0, 0, 'bool', 'Include debug warning statements in pgsql commands'), + array('ignore-errors', '', 0, 1, 0, 0, 'bool', 'Continue import even when errors in SQL are present (EXPERT)'), array('create-minimal-tables', '', 0, 1, 0, 0, 'bool', 'Create minimal main tables'), array('create-tables', '', 0, 1, 0, 0, 'bool', 'Create main tables'), - array('create-partitions', '', 0, 1, 0, 0, 'bool', 'Create required partition tables and triggers'), + array('create-partition-tables', '', 0, 1, 0, 0, 'bool', 'Create required partition tables'), + array('create-partition-functions', '', 0, 1, 0, 0, 'bool', 'Create required partition triggers'), + array('no-partitions', '', 0, 1, 0, 0, 'bool', "Do not partition search indices (speeds up import of single country extracts)"), array('import-wikipedia-articles', '', 0, 1, 0, 0, 'bool', 'Import wikipedia article dump'), array('load-data', '', 0, 1, 0, 0, 'bool', 'Copy data to live tables from import table'), array('disable-token-precalc', '', 0, 1, 0, 0, 'bool', 'Disable name precalculation (EXPERT)'), array('import-tiger-data', '', 0, 1, 0, 0, 'bool', 'Import tiger data (not included in \'all\')'), array('calculate-postcodes', '', 0, 1, 0, 0, 'bool', 'Calculate postcode centroids'), - array('create-roads', '', 0, 1, 0, 0, 'bool', 'Calculate postcode centroids'), array('osmosis-init', '', 0, 1, 0, 0, 'bool', 'Generate default osmosis configuration'), array('index', '', 0, 1, 0, 0, 'bool', 'Index the data'), array('index-noanalyse', '', 0, 1, 0, 0, 'bool', 'Do not perform analyse operations during index (EXPERT)'), array('index-output', '', 0, 1, 1, 1, 'string', 'File to dump index information to'), array('create-search-indices', '', 0, 1, 0, 0, 'bool', 'Create additional indices required for search and update'), array('create-website', '', 0, 1, 1, 1, 'realpath', 'Create symlinks to setup web directory'), + array('drop', '', 0, 1, 0, 0, 'bool', 'Drop tables needed for updates, making the database readonly (EXPERIMENTAL)'), ); getCmdOpt($_SERVER['argv'], $aCMDOptions, $aCMDResult, true, true); @@ -76,16 +79,20 @@ } // Assume we can steal all the cache memory in the box (unless told otherwise) - $iCacheMemory = (isset($aCMDResult['osm2pgsql-cache'])?$aCMDResult['osm2pgsql-cache']:getCacheMemoryMB()); - if ($iCacheMemory > getTotalMemoryMB()) + if (isset($aCMDResult['osm2pgsql-cache'])) + { + $iCacheMemory = $aCMDResult['osm2pgsql-cache']; + } + else { $iCacheMemory = getCacheMemoryMB(); - echo "WARNING: resetting cache memory to $iCacheMemory\n"; } $aDSNInfo = DB::parseDSN(CONST_Database_DSN); if (!isset($aDSNInfo['port']) || !$aDSNInfo['port']) $aDSNInfo['port'] = 5432; + $fPostgisVersion = (float) CONST_Postgis_Version; + if ($aCMDResult['create-db'] || $aCMDResult['all']) { echo "Create DB\n"; @@ -95,7 +102,7 @@ { fail('database already exists ('.CONST_Database_DSN.')'); } - passthru('createdb -E UTF-8 '.$aDSNInfo['database']); + passthruCheckReturn('createdb -E UTF-8 -p '.$aDSNInfo['port'].' '.$aDSNInfo['database']); } if ($aCMDResult['setup-db'] || $aCMDResult['all']) @@ -105,23 +112,69 @@ // TODO: path detection, detection memory, etc. $oDB =& getDB(); - passthru('createlang plpgsql '.$aDSNInfo['database']); + + $sVersionString = $oDB->getOne('select version()'); + preg_match('#PostgreSQL ([0-9]+)[.]([0-9]+)[^0-9]#', $sVersionString, $aMatches); + if (CONST_Postgresql_Version != $aMatches[1].'.'.$aMatches[2]) + { + echo "ERROR: PostgreSQL version is not correct. Expected ".CONST_Postgresql_Version." found ".$aMatches[1].'.'.$aMatches[2]."\n"; + exit; + } + + passthru('createlang plpgsql -p '.$aDSNInfo['port'].' '.$aDSNInfo['database']); $pgver = (float) CONST_Postgresql_Version; if ($pgver < 9.1) { pgsqlRunScriptFile(CONST_Path_Postgresql_Contrib.'/hstore.sql'); + pgsqlRunScriptFile(CONST_BasePath.'/sql/hstore_compatability_9_0.sql'); } else { pgsqlRunScript('CREATE EXTENSION hstore'); } - pgsqlRunScriptFile(CONST_Path_Postgresql_Postgis.'/postgis.sql'); - pgsqlRunScriptFile(CONST_Path_Postgresql_Postgis.'/spatial_ref_sys.sql'); + + if ($fPostgisVersion < 2.0) { + pgsqlRunScriptFile(CONST_Path_Postgresql_Postgis.'/postgis.sql'); + pgsqlRunScriptFile(CONST_Path_Postgresql_Postgis.'/spatial_ref_sys.sql'); + } else { + pgsqlRunScript('CREATE EXTENSION IF NOT EXISTS postgis'); + } + if ($fPostgisVersion < 2.1) { + // Function was renamed in 2.1 and throws an annoying deprecation warning + pgsqlRunScript('ALTER FUNCTION st_line_interpolate_point(geometry, double precision) RENAME TO ST_LineInterpolatePoint'); + } + $sVersionString = $oDB->getOne('select postgis_full_version()'); + preg_match('#POSTGIS="([0-9]+)[.]([0-9]+)[.]([0-9]+)( r([0-9]+))?"#', $sVersionString, $aMatches); + if (CONST_Postgis_Version != $aMatches[1].'.'.$aMatches[2]) + { + echo "ERROR: PostGIS version is not correct. Expected ".CONST_Postgis_Version." found ".$aMatches[1].'.'.$aMatches[2]."\n"; + exit; + } + pgsqlRunScriptFile(CONST_BasePath.'/data/country_name.sql'); pgsqlRunScriptFile(CONST_BasePath.'/data/country_naturalearthdata.sql'); pgsqlRunScriptFile(CONST_BasePath.'/data/country_osm_grid.sql'); - pgsqlRunScriptFile(CONST_BasePath.'/data/gb_postcode.sql'); + pgsqlRunScriptFile(CONST_BasePath.'/data/gb_postcode_table.sql'); + if (file_exists(CONST_BasePath.'/data/gb_postcode_data.sql.gz')) + { + pgsqlRunScriptFile(CONST_BasePath.'/data/gb_postcode_data.sql.gz'); + } + else + { + echo "WARNING: external UK postcode table not found.\n"; + } pgsqlRunScriptFile(CONST_BasePath.'/data/us_statecounty.sql'); pgsqlRunScriptFile(CONST_BasePath.'/data/us_state.sql'); pgsqlRunScriptFile(CONST_BasePath.'/data/us_postcode.sql'); - pgsqlRunScriptFile(CONST_BasePath.'/data/worldboundaries.sql'); + + if ($aCMDResult['no-partitions']) + { + pgsqlRunScript('update country_name set partition = 0'); + } + + // the following will be needed by create_functions later but + // is only defined in the subsequently called create_tables. + // Create dummies here that will be overwritten by the proper + // versions in create-tables. + pgsqlRunScript('CREATE TABLE place_boundingbox ()'); + pgsqlRunScript('create type wikipedia_article_match as ()'); } if ($aCMDResult['import-data'] || $aCMDResult['all']) @@ -135,8 +188,22 @@ echo "Please download and build osm2pgsql.\nIf it is already installed, check the path in your local settings (settings/local.php) file.\n"; fail("osm2pgsql not found in '$osm2pgsql'"); } - $osm2pgsql .= ' -lsc -O gazetteer --hstore'; - $osm2pgsql .= ' -C '.$iCacheMemory; + + if (!is_null(CONST_Osm2pgsql_Flatnode_File)) + { + $osm2pgsql .= ' --flat-nodes '.CONST_Osm2pgsql_Flatnode_File; + } + if (CONST_Tablespace_Osm2pgsql_Data) + $osm2pgsql .= ' --tablespace-slim-data '.CONST_Tablespace_Osm2pgsql_Data; + if (CONST_Tablespace_Osm2pgsql_Index) + $osm2pgsql .= ' --tablespace-slim-index '.CONST_Tablespace_Osm2pgsql_Index; + if (CONST_Tablespace_Place_Data) + $osm2pgsql .= ' --tablespace-main-data '.CONST_Tablespace_Place_Data; + if (CONST_Tablespace_Place_Index) + $osm2pgsql .= ' --tablespace-main-index '.CONST_Tablespace_Place_Index; + $osm2pgsql .= ' -lsc -O gazetteer --hstore --number-processes 1'; + $osm2pgsql .= ' -C 25000'; + $osm2pgsql .= ' -P '.$aDSNInfo['port']; $osm2pgsql .= ' -d '.$aDSNInfo['database'].' '.$aCMDResult['osm-file']; passthruCheckReturn($osm2pgsql); @@ -157,6 +224,16 @@ $sTemplate = str_replace('{modulepath}', CONST_BasePath.'/module', $sTemplate); if ($aCMDResult['enable-diff-updates']) $sTemplate = str_replace('RETURN NEW; -- @DIFFUPDATES@', '--', $sTemplate); if ($aCMDResult['enable-debug-statements']) $sTemplate = str_replace('--DEBUG:', '', $sTemplate); + if (CONST_Limit_Reindexing) $sTemplate = str_replace('--LIMIT INDEXING:', '', $sTemplate); + pgsqlRunScript($sTemplate); + + if ($fPostgisVersion < 2.0) { + echo "Helper functions for postgis < 2.0\n"; + $sTemplate = file_get_contents(CONST_BasePath.'/sql/postgis_15_aux.sql'); + } else { + echo "Helper functions for postgis >= 2.0\n"; + $sTemplate = file_get_contents(CONST_BasePath.'/sql/postgis_20_aux.sql'); + } pgsqlRunScript($sTemplate); } @@ -192,30 +269,88 @@ if ($aCMDResult['create-tables'] || $aCMDResult['all']) { - echo "Tables\n"; $bDidSomething = true; - pgsqlRunScriptFile(CONST_BasePath.'/sql/tables.sql'); + + echo "Tables\n"; + $sTemplate = file_get_contents(CONST_BasePath.'/sql/tables.sql'); + $sTemplate = str_replace('{www-user}', CONST_Database_Web_User, $sTemplate); + $sTemplate = replace_tablespace('{ts:address-data}', + CONST_Tablespace_Address_Data, $sTemplate); + $sTemplate = replace_tablespace('{ts:address-index}', + CONST_Tablespace_Address_Index, $sTemplate); + $sTemplate = replace_tablespace('{ts:search-data}', + CONST_Tablespace_Search_Data, $sTemplate); + $sTemplate = replace_tablespace('{ts:search-index}', + CONST_Tablespace_Search_Index, $sTemplate); + $sTemplate = replace_tablespace('{ts:aux-data}', + CONST_Tablespace_Aux_Data, $sTemplate); + $sTemplate = replace_tablespace('{ts:aux-index}', + CONST_Tablespace_Aux_Index, $sTemplate); + pgsqlRunScript($sTemplate, false); // re-run the functions + echo "Functions\n"; $sTemplate = file_get_contents(CONST_BasePath.'/sql/functions.sql'); - $sTemplate = str_replace('{modulepath}',CONST_BasePath.'/module', $sTemplate); + $sTemplate = str_replace('{modulepath}', + CONST_BasePath.'/module', $sTemplate); + pgsqlRunScript($sTemplate); + } + + if ($aCMDResult['create-partition-tables'] || $aCMDResult['all']) + { + echo "Partition Tables\n"; + $bDidSomething = true; + $oDB =& getDB(); + $sSQL = 'select distinct partition from country_name'; + $aPartitions = $oDB->getCol($sSQL); + if (PEAR::isError($aPartitions)) + { + fail($aPartitions->getMessage()); + } + if (!$aCMDResult['no-partitions']) $aPartitions[] = 0; + + $sTemplate = file_get_contents(CONST_BasePath.'/sql/partition-tables.src.sql'); + $sTemplate = replace_tablespace('{ts:address-data}', + CONST_Tablespace_Address_Data, $sTemplate); + $sTemplate = replace_tablespace('{ts:address-index}', + CONST_Tablespace_Address_Index, $sTemplate); + $sTemplate = replace_tablespace('{ts:search-data}', + CONST_Tablespace_Search_Data, $sTemplate); + $sTemplate = replace_tablespace('{ts:search-index}', + CONST_Tablespace_Search_Index, $sTemplate); + $sTemplate = replace_tablespace('{ts:aux-data}', + CONST_Tablespace_Aux_Data, $sTemplate); + $sTemplate = replace_tablespace('{ts:aux-index}', + CONST_Tablespace_Aux_Index, $sTemplate); + preg_match_all('#^-- start(.*?)^-- end#ms', $sTemplate, $aMatches, PREG_SET_ORDER); + foreach($aMatches as $aMatch) + { + $sResult = ''; + foreach($aPartitions as $sPartitionName) + { + $sResult .= str_replace('-partition-', $sPartitionName, $aMatch[1]); + } + $sTemplate = str_replace($aMatch[0], $sResult, $sTemplate); + } + pgsqlRunScript($sTemplate); } - if ($aCMDResult['create-partitions'] || $aCMDResult['all']) + + if ($aCMDResult['create-partition-functions'] || $aCMDResult['all']) { - echo "Partitions\n"; + echo "Partition Functions\n"; $bDidSomething = true; $oDB =& getDB(); - $sSQL = 'select partition from country_name order by country_code'; + $sSQL = 'select distinct partition from country_name'; $aPartitions = $oDB->getCol($sSQL); if (PEAR::isError($aPartitions)) { fail($aPartitions->getMessage()); } - $aPartitions[] = 0; + if (!$aCMDResult['no-partitions']) $aPartitions[] = 0; - $sTemplate = file_get_contents(CONST_BasePath.'/sql/partitions.src.sql'); + $sTemplate = file_get_contents(CONST_BasePath.'/sql/partition-functions.src.sql'); preg_match_all('#^-- start(.*?)^-- end#ms', $sTemplate, $aMatches, PREG_SET_ORDER); foreach($aMatches as $aMatch) { @@ -283,13 +418,13 @@ if (!pg_query($oDB->connection, 'CREATE SEQUENCE seq_place start 100000')) fail(pg_last_error($oDB->connection)); echo '.'; - $sSQL = 'select partition from country_name order by country_code'; + $sSQL = 'select distinct partition from country_name'; $aPartitions = $oDB->getCol($sSQL); if (PEAR::isError($aPartitions)) { fail($aPartitions->getMessage()); } - $aPartitions[] = 0; + if (!$aCMDResult['no-partitions']) $aPartitions[] = 0; foreach($aPartitions as $sPartition) { if (!pg_query($oDB->connection, 'TRUNCATE location_road_'.$sPartition)) fail(pg_last_error($oDB->connection)); @@ -313,7 +448,7 @@ { $aDBInstances[$i] =& getDB(true); $sSQL = 'insert into placex (osm_type, osm_id, class, type, name, admin_level, '; - $sSQL .= 'housenumber, street, isin, postcode, country_code, extratags, '; + $sSQL .= 'housenumber, street, addr_place, isin, postcode, country_code, extratags, '; $sSQL .= 'geometry) select * from place where osm_id % '.$iInstances.' = '.$i; if ($aCMDResult['verbose']) echo "$sSQL\n"; if (!pg_send_query($aDBInstances[$i]->connection, $sSQL)) fail(pg_last_error($oDB->connection)); @@ -334,40 +469,17 @@ pgsqlRunScript('ANALYSE'); } - if ($aCMDResult['create-roads']) - { - $bDidSomething = true; - - $oDB =& getDB(); - $aDBInstances = array(); - for($i = 0; $i < $iInstances; $i++) - { - $aDBInstances[$i] =& getDB(true); - if (!pg_query($aDBInstances[$i]->connection, 'set enable_bitmapscan = off')) fail(pg_last_error($oDB->connection)); - $sSQL = 'select count(*) from (select insertLocationRoad(partition, place_id, calculated_country_code, geometry) from '; - $sSQL .= 'placex where osm_id % '.$iInstances.' = '.$i.' and rank_search between 26 and 27 and class = \'highway\') as x '; - if ($aCMDResult['verbose']) echo "$sSQL\n"; - if (!pg_send_query($aDBInstances[$i]->connection, $sSQL)) fail(pg_last_error($oDB->connection)); - } - $bAnyBusy = true; - while($bAnyBusy) - { - $bAnyBusy = false; - for($i = 0; $i < $iInstances; $i++) - { - if (pg_connection_busy($aDBInstances[$i]->connection)) $bAnyBusy = true; - } - sleep(1); - echo '.'; - } - echo "\n"; - } - if ($aCMDResult['import-tiger-data']) { $bDidSomething = true; - pgsqlRunScriptFile(CONST_BasePath.'/sql/tiger_import_start.sql'); + $sTemplate = file_get_contents(CONST_BasePath.'/sql/tiger_import_start.sql'); + $sTemplate = str_replace('{www-user}', CONST_Database_Web_User, $sTemplate); + $sTemplate = replace_tablespace('{ts:aux-data}', + CONST_Tablespace_Aux_Data, $sTemplate); + $sTemplate = replace_tablespace('{ts:aux-index}', + CONST_Tablespace_Aux_Index, $sTemplate); + pgsqlRunScript($sTemplate, false); $aDBInstances = array(); for($i = 0; $i < $iInstances; $i++) @@ -375,7 +487,7 @@ $aDBInstances[$i] =& getDB(true); } - foreach(glob(CONST_BasePath.'/data/tiger2011/*.sql') as $sFile) + foreach(glob(CONST_Tiger_Data_Path.'/*.sql') as $sFile) { echo $sFile.': '; $hFile = fopen($sFile, "r"); @@ -419,7 +531,13 @@ } echo "Creating indexes\n"; - pgsqlRunScriptFile(CONST_BasePath.'/sql/tiger_import_finish.sql'); + $sTemplate = file_get_contents(CONST_BasePath.'/sql/tiger_import_finish.sql'); + $sTemplate = str_replace('{www-user}', CONST_Database_Web_User, $sTemplate); + $sTemplate = replace_tablespace('{ts:aux-data}', + CONST_Tablespace_Aux_Data, $sTemplate); + $sTemplate = replace_tablespace('{ts:aux-index}', + CONST_Tablespace_Aux_Index, $sTemplate); + pgsqlRunScript($sTemplate, false); } if ($aCMDResult['calculate-postcodes'] || $aCMDResult['all']) @@ -431,7 +549,7 @@ $sSQL .= "select 'P',nextval('seq_postcodes'),'place','postcode',postcode,calculated_country_code,"; $sSQL .= "ST_SetSRID(ST_Point(x,y),4326) as geometry from (select calculated_country_code,postcode,"; $sSQL .= "avg(st_x(st_centroid(geometry))) as x,avg(st_y(st_centroid(geometry))) as y "; - $sSQL .= "from placex where postcode is not null group by calculated_country_code,postcode) as x"; + $sSQL .= "from placex where postcode is not null and calculated_country_code not in ('ie') group by calculated_country_code,postcode) as x"; if (!pg_query($oDB->connection, $sSQL)) fail(pg_last_error($oDB->connection)); $sSQL = "insert into placex (osm_type,osm_id,class,type,postcode,calculated_country_code,geometry) "; @@ -440,77 +558,98 @@ if (!pg_query($oDB->connection, $sSQL)) fail(pg_last_error($oDB->connection)); } - if ($aCMDResult['osmosis-init'] || $aCMDResult['all']) + if ($aCMDResult['osmosis-init'] || ($aCMDResult['all'] && !$aCMDResult['drop'])) // no use doing osmosis-init when dropping update tables { $bDidSomething = true; $oDB =& getDB(); - if (!file_exists(CONST_Osmosis_Binary)) fail("please download osmosis"); - if (file_exists(CONST_BasePath.'/settings/configuration.txt')) + if (!file_exists(CONST_Osmosis_Binary)) { - echo "settings/configuration.txt already exists\n"; + echo "Please download osmosis.\nIf it is already installed, check the path in your local settings (settings/local.php) file.\n"; + if (!$aCMDResult['all']) + { + fail("osmosis not found in '".CONST_Osmosis_Binary."'"); + } } else { - passthru(CONST_Osmosis_Binary.' --read-replication-interval-init '.CONST_BasePath.'/settings'); - // server layout changed afer license change, fix path to minutely diffs - passthru("sed -i 's:minute-replicate:replication/minute:' ".CONST_BasePath.'/settings/configuration.txt'); - } - - // Find the last node in the DB - $iLastOSMID = $oDB->getOne("select max(osm_id) as osm_id from place where osm_type = 'N'"); - - // Lookup the timestamp that node was created (less 3 hours for margin for changsets to be closed) - $sLastNodeURL = 'http://www.openstreetmap.org/api/0.6/node/'.$iLastOSMID; - $sLastNodeXML = file_get_contents($sLastNodeURL); - preg_match('#timestamp="(([0-9]{4})-([0-9]{2})-([0-9]{2})T([0-9]{2}):([0-9]{2}):([0-9]{2})Z)"#', $sLastNodeXML, $aLastNodeDate); - $iLastNodeTimestamp = strtotime($aLastNodeDate[1]) - (3*60*60); - + if (file_exists(CONST_BasePath.'/settings/configuration.txt')) + { + echo "settings/configuration.txt already exists\n"; + } + else + { + passthru(CONST_Osmosis_Binary.' --read-replication-interval-init '.CONST_BasePath.'/settings'); + // update osmosis configuration.txt with our settings + passthru("sed -i 's!baseUrl=.*!baseUrl=".CONST_Replication_Url."!' ".CONST_BasePath.'/settings/configuration.txt'); + passthru("sed -i 's:maxInterval = .*:maxInterval = ".CONST_Replication_MaxInterval.":' ".CONST_BasePath.'/settings/configuration.txt'); + } - // Search for the correct state file - uses file timestamps - $sRepURL = 'http://planet.openstreetmap.org/replication/minute/'; - $sRep = file_get_contents($sRepURL); - preg_match_all('#([0-9]{3}/) *(([0-9]{2})-([A-z]{3})-([0-9]{4}) ([0-9]{2}):([0-9]{2}))#', $sRep, $aRepMatches, PREG_SET_ORDER); - $aPrevRepMatch = false; - foreach($aRepMatches as $aRepMatch) - { - if (strtotime($aRepMatch[2]) < $iLastNodeTimestamp) break; - $aPrevRepMatch = $aRepMatch; - } - if ($aPrevRepMatch) $aRepMatch = $aPrevRepMatch; + // Find the last node in the DB + $iLastOSMID = $oDB->getOne("select max(osm_id) from place where osm_type = 'N'"); + + // Lookup the timestamp that node was created (less 3 hours for margin for changsets to be closed) + $sLastNodeURL = 'http://www.openstreetmap.org/api/0.6/node/'.$iLastOSMID."/1"; + $sLastNodeXML = file_get_contents($sLastNodeURL); + preg_match('#timestamp="(([0-9]{4})-([0-9]{2})-([0-9]{2})T([0-9]{2}):([0-9]{2}):([0-9]{2})Z)"#', $sLastNodeXML, $aLastNodeDate); + $iLastNodeTimestamp = strtotime($aLastNodeDate[1]) - (3*60*60); + + // Search for the correct state file - uses file timestamps so need to sort by date descending + $sRepURL = CONST_Replication_Url."/"; + $sRep = file_get_contents($sRepURL."?C=M;O=D;F=1"); + // download.geofabrik.de: 000/