X-Git-Url: https://git.openstreetmap.org/nominatim.git/blobdiff_plain/f05ea577f42c5ad4a62ff25ec62fbb60c550ee26..cfb0f3f94ced30b8b44390367e07bc9b0cec787d:/utils/setup.php diff --git a/utils/setup.php b/utils/setup.php index 4d1427ee..06216587 100755 --- a/utils/setup.php +++ b/utils/setup.php @@ -5,40 +5,42 @@ require_once(dirname(dirname(__FILE__)).'/settings/settings.php'); require_once(CONST_BasePath.'/lib/init-cmd.php'); ini_set('memory_limit', '800M'); -$aCMDOptions = array( - "Create and setup nominatim search system", - 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('osm-file', '', 0, 1, 1, 1, 'realpath', 'File to import'), - array('threads', '', 0, 1, 1, 1, 'int', 'Number of threads (where possible)'), - - array('all', '', 0, 1, 0, 0, 'bool', 'Do the complete process'), - - array('create-db', '', 0, 1, 0, 0, 'bool', 'Create nominatim db'), - array('setup-db', '', 0, 1, 0, 0, 'bool', 'Build a blank nominatim db'), - array('import-data', '', 0, 1, 0, 0, 'bool', 'Import a osm file'), - array('osm2pgsql-cache', '', 0, 1, 1, 1, 'int', 'Cache size used by osm2pgsql'), - 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-tables', '', 0, 1, 0, 0, 'bool', 'Create main tables'), - 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('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('create-search-indices', '', 0, 1, 0, 0, 'bool', 'Create additional indices required for search and update'), - array('drop', '', 0, 1, 0, 0, 'bool', 'Drop tables needed for updates, making the database readonly (EXPERIMENTAL)'), -); +$aCMDOptions += array( + "Create and setup nominatim search system", + 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('osm-file', '', 0, 1, 1, 1, 'realpath', 'File to import'), + array('threads', '', 0, 1, 1, 1, 'int', 'Number of threads (where possible)'), + + array('all', '', 0, 1, 0, 0, 'bool', 'Do the complete process'), + + array('create-db', '', 0, 1, 0, 0, 'bool', 'Create nominatim db'), + array('setup-db', '', 0, 1, 0, 0, 'bool', 'Build a blank nominatim db'), + array('import-data', '', 0, 1, 0, 0, 'bool', 'Import a osm file'), + array('osm2pgsql-cache', '', 0, 1, 1, 1, 'int', 'Cache size used by osm2pgsql'), + 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-tables', '', 0, 1, 0, 0, 'bool', 'Create main tables'), + 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('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('create-search-indices', '', 0, 1, 0, 0, 'bool', 'Create additional indices required for search and update'), + array('create-country-names', '', 0, 1, 0, 0, 'bool', 'Create default list of searchable country names'), + array('drop', '', 0, 1, 0, 0, 'bool', 'Drop tables needed for updates, making the database readonly (EXPERIMENTAL)'), + ); getCmdOpt($_SERVER['argv'], $aCMDOptions, $aCMDResult, true, true); $bDidSomething = false; @@ -93,8 +95,9 @@ if ($aCMDResult['create-db'] || $aCMDResult['all']) { if ($aCMDResult['setup-db'] || $aCMDResult['all']) { echo "Setup DB\n"; $bDidSomething = true; - // TODO: path detection, detection memory, etc. + // TODO: path detection, detection memory, etc. + // $oDB =& getDB(); $fPostgresVersion = getPostgresVersion($oDB); @@ -121,8 +124,12 @@ if ($aCMDResult['setup-db'] || $aCMDResult['all']) { echo 'Postgis version found: '.$fPostgisVersion."\n"; if ($fPostgisVersion < 2.1) { - // Function was renamed in 2.1 and throws an annoying deprecation warning + // Functions were renamed in 2.1 and throw an annoying deprecation warning pgsqlRunScript('ALTER FUNCTION st_line_interpolate_point(geometry, double precision) RENAME TO ST_LineInterpolatePoint'); + pgsqlRunScript('ALTER FUNCTION ST_Line_Locate_Point(geometry, geometry) RENAME TO ST_LineLocatePoint'); + } + if ($fPostgisVersion < 2.2) { + pgsqlRunScript('ALTER FUNCTION ST_Distance_Spheroid(geometry, geometry, spheroid) RENAME TO ST_DistanceSpheroid'); } pgsqlRunScriptFile(CONST_BasePath.'/data/country_name.sql'); @@ -156,7 +163,8 @@ if ($aCMDResult['import-data'] || $aCMDResult['all']) { $osm2pgsql = CONST_Osm2pgsql_Binary; if (!file_exists($osm2pgsql)) { - echo "Please download and build osm2pgsql.\nIf it is already installed, check the path in your local settings (settings/local.php) file.\n"; + echo "Check CONST_Osm2pgsql_Binary in your local settings file.\n"; + echo "Normally you should not need to set this manually.\n"; fail("osm2pgsql not found in '$osm2pgsql'"); } @@ -196,18 +204,36 @@ if ($aCMDResult['create-tables'] || $aCMDResult['all']) { 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); + $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 @@ -220,18 +246,36 @@ if ($aCMDResult['create-partition-tables'] || $aCMDResult['all']) { $bDidSomething = true; $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); + $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, + $sTemplate + ); pgsqlRunPartitionScript($sTemplate); } @@ -302,7 +346,11 @@ if ($aCMDResult['load-data'] || $aCMDResult['all']) { } // used by getorcreate_word_id to ignore frequent partial words - if (!pg_query($oDB->connection, 'CREATE OR REPLACE FUNCTION get_maxwordfreq() RETURNS integer AS $$ SELECT '.CONST_Max_Word_Frequency.' as maxwordfreq; $$ LANGUAGE SQL IMMUTABLE')) fail(pg_last_error($oDB->connection)); + $sSQL = 'CREATE OR REPLACE FUNCTION get_maxwordfreq() RETURNS integer AS '; + $sSQL .= '$$ SELECT '.CONST_Max_Word_Frequency.' as maxwordfreq; $$ LANGUAGE SQL IMMUTABLE'; + if (!pg_query($oDB->connection, $sSQL)) { + fail(pg_last_error($oDB->connection)); + } echo ".\n"; // pre-create the word list @@ -350,10 +398,16 @@ if ($aCMDResult['import-tiger-data']) { $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); + $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); $aDBInstances = array(); @@ -400,10 +454,16 @@ if ($aCMDResult['import-tiger-data']) { echo "Creating indexes\n"; $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); + $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); } @@ -521,16 +581,51 @@ if ($aCMDResult['create-search-indices'] || $aCMDResult['all']) { $bDidSomething = true; $sTemplate = file_get_contents(CONST_BasePath.'/sql/indices.src.sql'); - $sTemplate = replace_tablespace('{ts:address-index}', - CONST_Tablespace_Address_Index, $sTemplate); - $sTemplate = replace_tablespace('{ts:search-index}', - CONST_Tablespace_Search_Index, $sTemplate); - $sTemplate = replace_tablespace('{ts:aux-index}', - CONST_Tablespace_Aux_Index, $sTemplate); + $sTemplate = replace_tablespace( + '{ts:address-index}', + CONST_Tablespace_Address_Index, + $sTemplate + ); + $sTemplate = replace_tablespace( + '{ts:search-index}', + CONST_Tablespace_Search_Index, + $sTemplate + ); + $sTemplate = replace_tablespace( + '{ts:aux-index}', + CONST_Tablespace_Aux_Index, + $sTemplate + ); pgsqlRunScript($sTemplate); } +if ($aCMDResult['create-country-names'] || $aCMDResult['all']) { + echo 'Creating search index for default country names'; + $bDidSomething = true; + + pgsqlRunScript("select getorcreate_country(make_standard_name('uk'), 'gb')"); + pgsqlRunScript("select getorcreate_country(make_standard_name('united states'), 'us')"); + pgsqlRunScript("select count(*) from (select getorcreate_country(make_standard_name(country_code), country_code) from country_name where country_code is not null) as x"); + pgsqlRunScript("select count(*) from (select getorcreate_country(make_standard_name(name->'name'), country_code) from country_name where name ? 'name') as x"); + + $sSQL = 'select count(*) from (select getorcreate_country(make_standard_name(v), country_code) from (select country_code, skeys(name) as k, svals(name) as v from country_name) x where k '; + if (CONST_Languages) { + $sSQL .= 'in '; + $sDelim = '('; + foreach (explode(',', CONST_Languages) as $sLang) { + $sSQL .= $sDelim."'name:$sLang'"; + $sDelim = ','; + } + $sSQL .= ')'; + } else { + // all include all simple name tags + $sSQL .= "like 'name:%'"; + } + $sSQL .= ') v'; + pgsqlRunScript($sSQL); +} + if ($aCMDResult['drop']) { // The implementation is potentially a bit dangerous because it uses // a positive selection of tables to keep, and deletes everything else. @@ -540,22 +635,22 @@ if ($aCMDResult['drop']) { // tables we want to keep. everything else goes. $aKeepTables = array( - "*columns", - "import_polygon_*", - "import_status", - "place_addressline", - "location_property*", - "placex", - "search_name", - "seq_*", - "word", - "query_log", - "new_query_log", - "gb_postcode", - "spatial_ref_sys", - "country_name", - "place_classtype_*" - ); + "*columns", + "import_polygon_*", + "import_status", + "place_addressline", + "location_property*", + "placex", + "search_name", + "seq_*", + "word", + "query_log", + "new_query_log", + "gb_postcode", + "spatial_ref_sys", + "country_name", + "place_classtype_*" + ); $oDB =& getDB(); $aDropTables = array(); @@ -591,6 +686,7 @@ if (!$bDidSomething) { echo "Setup finished.\n"; } + function pgsqlRunScriptFile($sFilename) { if (!file_exists($sFilename)) fail('unable to find '.$sFilename); @@ -603,10 +699,10 @@ function pgsqlRunScriptFile($sFilename) $ahGzipPipes = null; if (preg_match('/\\.gz$/', $sFilename)) { $aDescriptors = array( - 0 => array('pipe', 'r'), - 1 => array('pipe', 'w'), - 2 => array('file', '/dev/null', 'a') - ); + 0 => array('pipe', 'r'), + 1 => array('pipe', 'w'), + 2 => array('file', '/dev/null', 'a') + ); $hGzipProcess = proc_open('zcat '.$sFilename, $aDescriptors, $ahGzipPipes); if (!is_resource($hGzipProcess)) fail('unable to start zcat'); $aReadPipe = $ahGzipPipes[1]; @@ -617,10 +713,10 @@ function pgsqlRunScriptFile($sFilename) } $aDescriptors = array( - 0 => $aReadPipe, - 1 => array('pipe', 'w'), - 2 => array('file', '/dev/null', 'a') - ); + 0 => $aReadPipe, + 1 => array('pipe', 'w'), + 2 => array('file', '/dev/null', 'a') + ); $ahPipes = null; $hProcess = proc_open($sCMD, $aDescriptors, $ahPipes); if (!is_resource($hProcess)) fail('unable to start pgsql'); @@ -640,7 +736,6 @@ function pgsqlRunScriptFile($sFilename) fclose($ahGzipPipes[1]); proc_close($hGzipProcess); } - } function pgsqlRunScript($sScript, $bfatal = true) @@ -653,10 +748,10 @@ function pgsqlRunScript($sScript, $bfatal = true) if ($bfatal && !$aCMDResult['ignore-errors']) $sCMD .= ' -v ON_ERROR_STOP=1'; $aDescriptors = array( - 0 => array('pipe', 'r'), - 1 => STDOUT, - 2 => STDERR - ); + 0 => array('pipe', 'r'), + 1 => STDOUT, + 2 => STDERR + ); $ahPipes = null; $hProcess = @proc_open($sCMD, $aDescriptors, $ahPipes); if (!is_resource($hProcess)) fail('unable to start pgsql'); @@ -702,10 +797,10 @@ function pgsqlRunRestoreData($sDumpFile) $sCMD = 'pg_restore -p '.$aDSNInfo['port'].' -d '.$aDSNInfo['database'].' -Fc -a '.$sDumpFile; $aDescriptors = array( - 0 => array('pipe', 'r'), - 1 => array('pipe', 'w'), - 2 => array('file', '/dev/null', 'a') - ); + 0 => array('pipe', 'r'), + 1 => array('pipe', 'w'), + 2 => array('file', '/dev/null', 'a') + ); $ahPipes = null; $hProcess = proc_open($sCMD, $aDescriptors, $ahPipes); if (!is_resource($hProcess)) fail('unable to start pg_restore'); @@ -729,10 +824,10 @@ function pgsqlRunDropAndRestore($sDumpFile) $sCMD = 'pg_restore -p '.$aDSNInfo['port'].' -d '.$aDSNInfo['database'].' -Fc --clean '.$sDumpFile; $aDescriptors = array( - 0 => array('pipe', 'r'), - 1 => array('pipe', 'w'), - 2 => array('file', '/dev/null', 'a') - ); + 0 => array('pipe', 'r'), + 1 => array('pipe', 'w'), + 2 => array('file', '/dev/null', 'a') + ); $ahPipes = null; $hProcess = proc_open($sCMD, $aDescriptors, $ahPipes); if (!is_resource($hProcess)) fail('unable to start pg_restore'); @@ -758,8 +853,7 @@ function passthruCheckReturn($cmd) function replace_tablespace($sTemplate, $sTablespace, $sSql) { if ($sTablespace) { - $sSql = str_replace($sTemplate, 'TABLESPACE "'.$sTablespace.'"', - $sSql); + $sSql = str_replace($sTemplate, 'TABLESPACE "'.$sTablespace.'"', $sSql); } else { $sSql = str_replace($sTemplate, '', $sSql); } @@ -787,6 +881,4 @@ function create_sql_functions($aCMDResult) $sTemplate = str_replace('-- %NOAUXDATA% ', '', $sTemplate); } pgsqlRunScript($sTemplate); - } -