Nominatim/utils/setup.php

719 lines
25 KiB
PHP
Raw Normal View History

2010-10-24 03:12:37 +04:00
#!/usr/bin/php -Cq
<?php
2010-10-25 16:22:22 +04:00
require_once(dirname(dirname(__FILE__)).'/lib/init-cmd.php');
2010-10-24 03:12:37 +04:00
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'),
2010-11-15 19:34:27 +03:00
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'),
2010-10-27 18:05:42 +04:00
array('create-db', '', 0, 1, 0, 0, 'bool', 'Create nominatim db'),
array('setup-db', '', 0, 1, 0, 0, 'bool', 'Build a blank nominatim db'),
2010-11-15 19:34:27 +03:00
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'),
2010-10-27 18:05:42 +04:00
array('create-functions', '', 0, 1, 0, 0, 'bool', 'Create functions'),
DB Scheme changes: alter table placex add column calculated_country_code varchar(2); function changes: ----------------- Move to ST_PointOnSurface from ST_Centroid in various places to avoid looking up a point outside the polygon Move to ST_Covers from ST_Contains to include points on admin boundaries Re-order preference for get_country_code now our data is better. country_osm_grid is now the preffered source. Fix code to calculate country code in placex_insert, rank_search test was too early Add extra field to placex 'calculated_country_code' to improve structure of code Move split_geometery function out of add_location into its own function Rewrite split_geometery to be more efficient. Change place_insert to do more updates and less delete/inserts (delete is slow) Include wikipedia links in details.php ouput Cleanup no longer used geometry validation (adding overhead) Include debug statements in function.sql (--DEBUG: ) and add flag to setup.php to turn them on setup.php: ---------- add flag --disable-token-precalc to speed up debuging add flag --index-noanalyse to disable analysising DB at rank 4 and 26 (previously removed, but on my local DB it seems to be required) add flag --enable-diff-updates (modifier to --create-functions) to turn on the code required for diff updates without having to modify functions.sql add flag --enable-debug-statements (modifier to --create-functions) to turn on debug warning statements update.php: ----------- added flag --no-index to import osmosis changes without indexing them extend the hack to allow import of JOSM generated osm files country_grid.sql - reference copy of the sql used to generate the country_osm_grid table, needs cleanup
2012-05-22 18:27:42 +04:00
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('create-minimal-tables', '', 0, 1, 0, 0, 'bool', 'Create minimal main tables'),
2010-10-27 18:05:42 +04:00
array('create-tables', '', 0, 1, 0, 0, 'bool', 'Create main tables'),
2010-10-24 03:12:37 +04:00
array('create-partitions', '', 0, 1, 0, 0, 'bool', 'Create required partition tables and triggers'),
2012-04-30 04:05:06 +04:00
array('import-wikipedia-articles', '', 0, 1, 0, 0, 'bool', 'Import wikipedia article dump'),
2010-10-27 18:05:42 +04:00
array('load-data', '', 0, 1, 0, 0, 'bool', 'Copy data to live tables from import table'),
DB Scheme changes: alter table placex add column calculated_country_code varchar(2); function changes: ----------------- Move to ST_PointOnSurface from ST_Centroid in various places to avoid looking up a point outside the polygon Move to ST_Covers from ST_Contains to include points on admin boundaries Re-order preference for get_country_code now our data is better. country_osm_grid is now the preffered source. Fix code to calculate country code in placex_insert, rank_search test was too early Add extra field to placex 'calculated_country_code' to improve structure of code Move split_geometery function out of add_location into its own function Rewrite split_geometery to be more efficient. Change place_insert to do more updates and less delete/inserts (delete is slow) Include wikipedia links in details.php ouput Cleanup no longer used geometry validation (adding overhead) Include debug statements in function.sql (--DEBUG: ) and add flag to setup.php to turn them on setup.php: ---------- add flag --disable-token-precalc to speed up debuging add flag --index-noanalyse to disable analysising DB at rank 4 and 26 (previously removed, but on my local DB it seems to be required) add flag --enable-diff-updates (modifier to --create-functions) to turn on the code required for diff updates without having to modify functions.sql add flag --enable-debug-statements (modifier to --create-functions) to turn on debug warning statements update.php: ----------- added flag --no-index to import osmosis changes without indexing them extend the hack to allow import of JOSM generated osm files country_grid.sql - reference copy of the sql used to generate the country_osm_grid table, needs cleanup
2012-05-22 18:27:42 +04:00
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'),
2012-07-10 17:36:17 +04:00
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'),
2010-10-24 03:12:37 +04:00
);
getCmdOpt($_SERVER['argv'], $aCMDOptions, $aCMDResult, true, true);
2010-10-26 19:22:41 +04:00
$bDidSomething = false;
// Check if osm-file is set and points to a valid file if --all or --import-data is given
if ($aCMDResult['import-data'] || $aCMDResult['all'])
{
if (!isset($aCMDResult['osm-file']))
{
fail('missing --osm-file for data import');
}
if (!file_exists($aCMDResult['osm-file']))
{
fail('the path supplied to --osm-file does not exist');
}
if (!is_readable($aCMDResult['osm-file']))
{
fail('osm-file "'.$aCMDResult['osm-file'].'" not readable');
}
}
2012-07-10 17:36:17 +04:00
// This is a pretty hard core default - the number of processors in the box - 1
$iInstances = isset($aCMDResult['threads'])?$aCMDResult['threads']:(getProcessorCount()-1);
if ($iInstances < 1)
{
$iInstances = 1;
echo "WARNING: resetting threads to $iInstances\n";
}
if ($iInstances > getProcessorCount())
{
$iInstances = getProcessorCount();
echo "WARNING: resetting threads to $iInstances\n";
}
// 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())
{
$iCacheMemory = getCacheMemoryMB();
echo "WARNING: resetting cache memory to $iCacheMemory\n";
}
$aDSNInfo = DB::parseDSN(CONST_Database_DSN);
if (!isset($aDSNInfo['port']) || !$aDSNInfo['port']) $aDSNInfo['port'] = 5432;
2010-11-15 19:34:27 +03:00
if ($aCMDResult['create-db'] || $aCMDResult['all'])
2010-10-27 18:05:42 +04:00
{
echo "Create DB\n";
2010-10-27 18:05:42 +04:00
$bDidSomething = true;
$oDB =& DB::connect(CONST_Database_DSN, false);
if (!PEAR::isError($oDB))
{
fail('database already exists ('.CONST_Database_DSN.')');
2010-10-27 18:05:42 +04:00
}
passthruCheckReturn('createdb -E UTF-8 -p '.$aDSNInfo['port'].' '.$aDSNInfo['database']);
2010-10-27 18:05:42 +04:00
}
if ($aCMDResult['setup-db'] || $aCMDResult['all'])
2010-10-25 16:22:22 +04:00
{
echo "Setup DB\n";
2010-10-26 19:22:41 +04:00
$bDidSomething = true;
2010-10-25 16:22:22 +04:00
// TODO: path detection, detection memory, etc.
2010-10-27 18:05:42 +04:00
$oDB =& getDB();
$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');
}
2010-10-27 18:05:42 +04:00
pgsqlRunScriptFile(CONST_Path_Postgresql_Postgis.'/postgis.sql');
$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;
}
2010-10-27 18:05:42 +04:00
pgsqlRunScriptFile(CONST_Path_Postgresql_Postgis.'/spatial_ref_sys.sql');
pgsqlRunScriptFile(CONST_BasePath.'/data/country_name.sql');
pgsqlRunScriptFile(CONST_BasePath.'/data/country_naturalearthdata.sql');
pgsqlRunScriptFile(CONST_BasePath.'/data/country_osm_grid.sql');
2010-10-27 18:05:42 +04:00
pgsqlRunScriptFile(CONST_BasePath.'/data/gb_postcode.sql');
pgsqlRunScriptFile(CONST_BasePath.'/data/us_statecounty.sql');
pgsqlRunScriptFile(CONST_BasePath.'/data/us_state.sql');
pgsqlRunScriptFile(CONST_BasePath.'/data/us_postcode.sql');
2010-10-27 18:05:42 +04:00
pgsqlRunScriptFile(CONST_BasePath.'/data/worldboundaries.sql');
}
2010-11-15 19:34:27 +03:00
if ($aCMDResult['import-data'] || $aCMDResult['all'])
2010-10-27 18:05:42 +04:00
{
echo "Import\n";
2010-10-27 18:05:42 +04:00
$bDidSomething = true;
$osm2pgsql = CONST_Osm2pgsql_Binary;
2012-05-24 02:26:16 +04:00
if (!file_exists($osm2pgsql))
{
echo "Please download and build osm2pgsql.\nIf it is already installed, check the path in your local settings (settings/local.php) file.\n";
fail("osm2pgsql not found in '$osm2pgsql'");
}
$osm2pgsql .= ' -lsc -O gazetteer --hstore';
$osm2pgsql .= ' -C '.$iCacheMemory;
$osm2pgsql .= ' -P '.$aDSNInfo['port'];
$osm2pgsql .= ' -d '.$aDSNInfo['database'].' '.$aCMDResult['osm-file'];
passthruCheckReturn($osm2pgsql);
$oDB =& getDB();
$x = $oDB->getRow('select * from place limit 1');
2012-04-26 15:30:49 +04:00
if (PEAR::isError($x)) {
fail($x->getMessage());
}
if (!$x) fail('No Data');
2010-10-27 18:05:42 +04:00
}
2010-11-15 19:34:27 +03:00
if ($aCMDResult['create-functions'] || $aCMDResult['all'])
2010-10-27 18:05:42 +04:00
{
echo "Functions\n";
2010-10-27 18:05:42 +04:00
$bDidSomething = true;
2010-11-15 19:34:27 +03:00
if (!file_exists(CONST_BasePath.'/module/nominatim.so')) fail("nominatim module not built");
2010-10-27 18:05:42 +04:00
$sTemplate = file_get_contents(CONST_BasePath.'/sql/functions.sql');
DB Scheme changes: alter table placex add column calculated_country_code varchar(2); function changes: ----------------- Move to ST_PointOnSurface from ST_Centroid in various places to avoid looking up a point outside the polygon Move to ST_Covers from ST_Contains to include points on admin boundaries Re-order preference for get_country_code now our data is better. country_osm_grid is now the preffered source. Fix code to calculate country code in placex_insert, rank_search test was too early Add extra field to placex 'calculated_country_code' to improve structure of code Move split_geometery function out of add_location into its own function Rewrite split_geometery to be more efficient. Change place_insert to do more updates and less delete/inserts (delete is slow) Include wikipedia links in details.php ouput Cleanup no longer used geometry validation (adding overhead) Include debug statements in function.sql (--DEBUG: ) and add flag to setup.php to turn them on setup.php: ---------- add flag --disable-token-precalc to speed up debuging add flag --index-noanalyse to disable analysising DB at rank 4 and 26 (previously removed, but on my local DB it seems to be required) add flag --enable-diff-updates (modifier to --create-functions) to turn on the code required for diff updates without having to modify functions.sql add flag --enable-debug-statements (modifier to --create-functions) to turn on debug warning statements update.php: ----------- added flag --no-index to import osmosis changes without indexing them extend the hack to allow import of JOSM generated osm files country_grid.sql - reference copy of the sql used to generate the country_osm_grid table, needs cleanup
2012-05-22 18:27:42 +04:00
$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:', '', $sTemplate);
2010-10-27 18:05:42 +04:00
pgsqlRunScript($sTemplate);
2010-10-25 16:22:22 +04:00
}
if ($aCMDResult['create-minimal-tables'])
{
echo "Minimal Tables\n";
$bDidSomething = true;
pgsqlRunScriptFile(CONST_BasePath.'/sql/tables-minimal.sql');
$sScript = '';
// Backstop the import process - easliest possible import id
$sScript .= "insert into import_npi_log values (18022);\n";
$hFile = @fopen(CONST_BasePath.'/settings/partitionedtags.def', "r");
if (!$hFile) fail('unable to open list of partitions: '.CONST_BasePath.'/settings/partitionedtags.def');
while (($sLine = fgets($hFile, 4096)) !== false && $sLine && substr($sLine,0,1) !='#')
{
list($sClass, $sType) = explode(' ', trim($sLine));
$sScript .= "create table place_classtype_".$sClass."_".$sType." as ";
$sScript .= "select place_id as place_id,geometry as centroid from placex limit 0;\n";
$sScript .= "CREATE INDEX idx_place_classtype_".$sClass."_".$sType."_centroid ";
$sScript .= "ON place_classtype_".$sClass."_".$sType." USING GIST (centroid);\n";
$sScript .= "CREATE INDEX idx_place_classtype_".$sClass."_".$sType."_place_id ";
$sScript .= "ON place_classtype_".$sClass."_".$sType." USING btree(place_id);\n";
}
fclose($hFile);
pgsqlRunScript($sScript);
}
2010-11-15 19:34:27 +03:00
if ($aCMDResult['create-tables'] || $aCMDResult['all'])
2010-10-25 16:22:22 +04:00
{
echo "Tables\n";
2010-10-26 19:22:41 +04:00
$bDidSomething = true;
2010-10-27 18:05:42 +04:00
pgsqlRunScriptFile(CONST_BasePath.'/sql/tables.sql');
// re-run the functions
$sTemplate = file_get_contents(CONST_BasePath.'/sql/functions.sql');
$sTemplate = str_replace('{modulepath}',CONST_BasePath.'/module', $sTemplate);
pgsqlRunScript($sTemplate);
2010-10-25 16:22:22 +04:00
}
2010-11-15 19:34:27 +03:00
if ($aCMDResult['create-partitions'] || $aCMDResult['all'])
2010-10-24 03:12:37 +04:00
{
echo "Partitions\n";
2010-10-26 19:22:41 +04:00
$bDidSomething = true;
2010-10-27 18:05:42 +04:00
$oDB =& getDB();
$sSQL = 'select partition from country_name order by country_code';
2010-10-24 03:12:37 +04:00
$aPartitions = $oDB->getCol($sSQL);
2010-10-25 16:22:22 +04:00
if (PEAR::isError($aPartitions))
{
fail($aPartitions->getMessage());
}
$aPartitions[] = 0;
2010-10-24 03:12:37 +04:00
$sTemplate = file_get_contents(CONST_BasePath.'/sql/partitions.src.sql');
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, $sTemplate);
}
2010-10-27 18:05:42 +04:00
pgsqlRunScript($sTemplate);
}
2012-04-30 04:05:06 +04:00
if ($aCMDResult['import-wikipedia-articles'] || $aCMDResult['all'])
{
$bDidSomething = true;
$sWikiArticlesFile = CONST_BasePath.'/data/wikipedia_article.sql.bin';
$sWikiRedirectsFile = CONST_BasePath.'/data/wikipedia_redirect.sql.bin';
if (file_exists($sWikiArticlesFile))
{
echo "Importing wikipedia articles...";
pgsqlRunDropAndRestore($sWikiArticlesFile);
2012-04-30 04:05:06 +04:00
echo "...done\n";
}
else
{
echo "WARNING: wikipedia article dump file not found - places will have default importance\n";
}
if (file_exists($sWikiRedirectsFile))
{
echo "Importing wikipedia redirects...";
pgsqlRunDropAndRestore($sWikiRedirectsFile);
2012-04-30 04:05:06 +04:00
echo "...done\n";
}
else
{
echo "WARNING: wikipedia redirect dump file not found - some place importance values may be missing\n";
}
}
2010-11-15 19:34:27 +03:00
if ($aCMDResult['load-data'] || $aCMDResult['all'])
2010-10-27 18:05:42 +04:00
{
echo "Drop old Data\n";
2010-10-27 18:05:42 +04:00
$bDidSomething = true;
$oDB =& getDB();
if (!pg_query($oDB->connection, 'TRUNCATE word')) fail(pg_last_error($oDB->connection));
echo '.';
if (!pg_query($oDB->connection, 'TRUNCATE placex')) fail(pg_last_error($oDB->connection));
echo '.';
if (!pg_query($oDB->connection, 'TRUNCATE place_addressline')) fail(pg_last_error($oDB->connection));
echo '.';
if (!pg_query($oDB->connection, 'TRUNCATE place_boundingbox')) fail(pg_last_error($oDB->connection));
echo '.';
if (!pg_query($oDB->connection, 'TRUNCATE location_area')) fail(pg_last_error($oDB->connection));
echo '.';
if (!pg_query($oDB->connection, 'TRUNCATE search_name')) fail(pg_last_error($oDB->connection));
echo '.';
if (!pg_query($oDB->connection, 'TRUNCATE search_name_blank')) fail(pg_last_error($oDB->connection));
echo '.';
if (!pg_query($oDB->connection, 'DROP SEQUENCE seq_place')) fail(pg_last_error($oDB->connection));
echo '.';
if (!pg_query($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';
$aPartitions = $oDB->getCol($sSQL);
if (PEAR::isError($aPartitions))
{
fail($aPartitions->getMessage());
}
$aPartitions[] = 0;
foreach($aPartitions as $sPartition)
{
if (!pg_query($oDB->connection, 'TRUNCATE location_road_'.$sPartition)) fail(pg_last_error($oDB->connection));
echo '.';
}
// 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($oDB->connection));
echo ".\n";
// pre-create the word list
DB Scheme changes: alter table placex add column calculated_country_code varchar(2); function changes: ----------------- Move to ST_PointOnSurface from ST_Centroid in various places to avoid looking up a point outside the polygon Move to ST_Covers from ST_Contains to include points on admin boundaries Re-order preference for get_country_code now our data is better. country_osm_grid is now the preffered source. Fix code to calculate country code in placex_insert, rank_search test was too early Add extra field to placex 'calculated_country_code' to improve structure of code Move split_geometery function out of add_location into its own function Rewrite split_geometery to be more efficient. Change place_insert to do more updates and less delete/inserts (delete is slow) Include wikipedia links in details.php ouput Cleanup no longer used geometry validation (adding overhead) Include debug statements in function.sql (--DEBUG: ) and add flag to setup.php to turn them on setup.php: ---------- add flag --disable-token-precalc to speed up debuging add flag --index-noanalyse to disable analysising DB at rank 4 and 26 (previously removed, but on my local DB it seems to be required) add flag --enable-diff-updates (modifier to --create-functions) to turn on the code required for diff updates without having to modify functions.sql add flag --enable-debug-statements (modifier to --create-functions) to turn on debug warning statements update.php: ----------- added flag --no-index to import osmosis changes without indexing them extend the hack to allow import of JOSM generated osm files country_grid.sql - reference copy of the sql used to generate the country_osm_grid table, needs cleanup
2012-05-22 18:27:42 +04:00
if (!$aCMDResult['disable-token-precalc'])
{
echo "Loading word list\n";
pgsqlRunScriptFile(CONST_BasePath.'/data/words.sql');
DB Scheme changes: alter table placex add column calculated_country_code varchar(2); function changes: ----------------- Move to ST_PointOnSurface from ST_Centroid in various places to avoid looking up a point outside the polygon Move to ST_Covers from ST_Contains to include points on admin boundaries Re-order preference for get_country_code now our data is better. country_osm_grid is now the preffered source. Fix code to calculate country code in placex_insert, rank_search test was too early Add extra field to placex 'calculated_country_code' to improve structure of code Move split_geometery function out of add_location into its own function Rewrite split_geometery to be more efficient. Change place_insert to do more updates and less delete/inserts (delete is slow) Include wikipedia links in details.php ouput Cleanup no longer used geometry validation (adding overhead) Include debug statements in function.sql (--DEBUG: ) and add flag to setup.php to turn them on setup.php: ---------- add flag --disable-token-precalc to speed up debuging add flag --index-noanalyse to disable analysising DB at rank 4 and 26 (previously removed, but on my local DB it seems to be required) add flag --enable-diff-updates (modifier to --create-functions) to turn on the code required for diff updates without having to modify functions.sql add flag --enable-debug-statements (modifier to --create-functions) to turn on debug warning statements update.php: ----------- added flag --no-index to import osmosis changes without indexing them extend the hack to allow import of JOSM generated osm files country_grid.sql - reference copy of the sql used to generate the country_osm_grid table, needs cleanup
2012-05-22 18:27:42 +04:00
}
echo "Load Data\n";
$aDBInstances = array();
for($i = 0; $i < $iInstances; $i++)
{
$aDBInstances[$i] =& getDB(true);
$sSQL = 'insert into placex (osm_type, osm_id, class, type, name, admin_level, ';
$sSQL .= 'housenumber, street, isin, postcode, country_code, extratags, ';
$sSQL .= 'geometry) select * from place where osm_id % '.$iInstances.' = '.$i;
2010-11-15 19:34:27 +03:00
if ($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)) $bAnyBusy = true;
}
sleep(1);
echo '.';
}
echo "\n";
echo "Reanalysing database...\n";
pgsqlRunScript('ANALYSE');
2010-10-24 03:12:37 +04:00
}
if ($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');
$aDBInstances = array();
for($i = 0; $i < $iInstances; $i++)
{
$aDBInstances[$i] =& getDB(true);
}
2012-02-18 00:34:28 +04:00
foreach(glob(CONST_BasePath.'/data/tiger2011/*.sql') as $sFile)
{
echo $sFile.': ';
$hFile = fopen($sFile, "r");
$sSQL = fgets($hFile, 100000);
$iLines = 0;
while(true)
{
for($i = 0; $i < $iInstances; $i++)
{
if (!pg_connection_busy($aDBInstances[$i]->connection))
{
while(pg_get_result($aDBInstances[$i]->connection));
$sSQL = fgets($hFile, 100000);
if (!$sSQL) break 2;
if (!pg_send_query($aDBInstances[$i]->connection, $sSQL)) fail(pg_last_error($oDB->connection));
$iLines++;
if ($iLines == 1000)
{
echo ".";
$iLines = 0;
}
}
}
usleep(10);
}
fclose($hFile);
2012-04-30 04:05:06 +04:00
$bAnyBusy = true;
while($bAnyBusy)
{
$bAnyBusy = false;
for($i = 0; $i < $iInstances; $i++)
{
if (pg_connection_busy($aDBInstances[$i]->connection)) $bAnyBusy = true;
}
usleep(10);
}
echo "\n";
}
echo "Creating indexes\n";
pgsqlRunScriptFile(CONST_BasePath.'/sql/tiger_import_finish.sql');
}
if ($aCMDResult['calculate-postcodes'] || $aCMDResult['all'])
{
$bDidSomething = true;
$oDB =& getDB();
if (!pg_query($oDB->connection, 'DELETE from placex where osm_type=\'P\'')) fail(pg_last_error($oDB->connection));
$sSQL = "insert into placex (osm_type,osm_id,class,type,postcode,calculated_country_code,geometry) ";
$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 ";
$sSQL .= "from placex where postcode is not null group by calculated_country_code,postcode) as x";
if (!pg_query($oDB->connection, $sSQL)) fail(pg_last_error($oDB->connection));
2012-10-10 00:20:05 +04:00
$sSQL = "insert into placex (osm_type,osm_id,class,type,postcode,calculated_country_code,geometry) ";
$sSQL .= "select 'P',nextval('seq_postcodes'),'place','postcode',postcode,'us',";
$sSQL .= "ST_SetSRID(ST_Point(x,y),4326) as geometry from us_postcode";
if (!pg_query($oDB->connection, $sSQL)) fail(pg_last_error($oDB->connection));
}
if ($aCMDResult['osmosis-init'] || $aCMDResult['all'])
{
$bDidSomething = true;
$oDB =& getDB();
if (!file_exists(CONST_Osmosis_Binary)) fail("please download osmosis");
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');
// 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);
// Search for the correct state file - uses file timestamps
$sRepURL = 'http://planet.openstreetmap.org/replication/minute/';
$sRep = file_get_contents($sRepURL);
preg_match_all('#<a href="[0-9]{3}/">([0-9]{3}/)</a> *(([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;
$sRepURL .= $aRepMatch[1];
$sRep = file_get_contents($sRepURL);
preg_match_all('#<a href="[0-9]{3}/">([0-9]{3}/)</a> *(([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;
$sRepURL .= $aRepMatch[1];
$sRep = file_get_contents($sRepURL);
preg_match_all('#<a href="[0-9]{3}.state.txt">([0-9]{3}).state.txt</a> *(([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;
$sRepURL .= $aRepMatch[1].'.state.txt';
echo "Getting state file: $sRepURL\n";
$sStateFile = file_get_contents($sRepURL);
if (!$sStateFile || strlen($sStateFile) > 1000) fail("unable to obtain state file");
file_put_contents(CONST_BasePath.'/settings/state.txt', $sStateFile);
echo "Updating DB status\n";
pg_query($oDB->connection, 'TRUNCATE import_status');
$sSQL = "INSERT INTO import_status VALUES('".$aRepMatch[2]."')";
pg_query($oDB->connection, $sSQL);
}
if ($aCMDResult['index'] || $aCMDResult['all'])
{
$bDidSomething = true;
$sOutputFile = '';
if (isset($aCMDResult['index-output'])) $sOutputFile = ' -F '.$aCMDResult['index-output'];
$sBaseCmd = CONST_BasePath.'/nominatim/nominatim -i -d '.$aDSNInfo['database'].' -P '.$aDSNInfo['port'].' -t '.$iInstances.$sOutputFile;
passthruCheckReturn($sBaseCmd.' -R 4');
DB Scheme changes: alter table placex add column calculated_country_code varchar(2); function changes: ----------------- Move to ST_PointOnSurface from ST_Centroid in various places to avoid looking up a point outside the polygon Move to ST_Covers from ST_Contains to include points on admin boundaries Re-order preference for get_country_code now our data is better. country_osm_grid is now the preffered source. Fix code to calculate country code in placex_insert, rank_search test was too early Add extra field to placex 'calculated_country_code' to improve structure of code Move split_geometery function out of add_location into its own function Rewrite split_geometery to be more efficient. Change place_insert to do more updates and less delete/inserts (delete is slow) Include wikipedia links in details.php ouput Cleanup no longer used geometry validation (adding overhead) Include debug statements in function.sql (--DEBUG: ) and add flag to setup.php to turn them on setup.php: ---------- add flag --disable-token-precalc to speed up debuging add flag --index-noanalyse to disable analysising DB at rank 4 and 26 (previously removed, but on my local DB it seems to be required) add flag --enable-diff-updates (modifier to --create-functions) to turn on the code required for diff updates without having to modify functions.sql add flag --enable-debug-statements (modifier to --create-functions) to turn on debug warning statements update.php: ----------- added flag --no-index to import osmosis changes without indexing them extend the hack to allow import of JOSM generated osm files country_grid.sql - reference copy of the sql used to generate the country_osm_grid table, needs cleanup
2012-05-22 18:27:42 +04:00
if (!$aCMDResult['index-noanalyse']) pgsqlRunScript('ANALYSE');
passthruCheckReturn($sBaseCmd.' -r 5 -R 25');
DB Scheme changes: alter table placex add column calculated_country_code varchar(2); function changes: ----------------- Move to ST_PointOnSurface from ST_Centroid in various places to avoid looking up a point outside the polygon Move to ST_Covers from ST_Contains to include points on admin boundaries Re-order preference for get_country_code now our data is better. country_osm_grid is now the preffered source. Fix code to calculate country code in placex_insert, rank_search test was too early Add extra field to placex 'calculated_country_code' to improve structure of code Move split_geometery function out of add_location into its own function Rewrite split_geometery to be more efficient. Change place_insert to do more updates and less delete/inserts (delete is slow) Include wikipedia links in details.php ouput Cleanup no longer used geometry validation (adding overhead) Include debug statements in function.sql (--DEBUG: ) and add flag to setup.php to turn them on setup.php: ---------- add flag --disable-token-precalc to speed up debuging add flag --index-noanalyse to disable analysising DB at rank 4 and 26 (previously removed, but on my local DB it seems to be required) add flag --enable-diff-updates (modifier to --create-functions) to turn on the code required for diff updates without having to modify functions.sql add flag --enable-debug-statements (modifier to --create-functions) to turn on debug warning statements update.php: ----------- added flag --no-index to import osmosis changes without indexing them extend the hack to allow import of JOSM generated osm files country_grid.sql - reference copy of the sql used to generate the country_osm_grid table, needs cleanup
2012-05-22 18:27:42 +04:00
if (!$aCMDResult['index-noanalyse']) pgsqlRunScript('ANALYSE');
passthruCheckReturn($sBaseCmd.' -r 26');
}
if ($aCMDResult['create-search-indices'] || $aCMDResult['all'])
{
echo "Search indices\n";
$bDidSomething = true;
$oDB =& getDB();
$sSQL = 'select partition from country_name order by country_code';
$aPartitions = $oDB->getCol($sSQL);
if (PEAR::isError($aPartitions))
{
fail($aPartitions->getMessage());
}
$aPartitions[] = 0;
$sTemplate = file_get_contents(CONST_BasePath.'/sql/indices.src.sql');
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 (isset($aCMDResult['create-website']))
{
$bDidSomething = true;
$sTargetDir = $aCMDResult['create-website'];
2012-05-24 02:26:16 +04:00
if (!is_dir($sTargetDir))
{
echo "You must create the website directory before calling this function.\n";
fail("Target directory does not exist.");
}
@symlink(CONST_BasePath.'/website/details.php', $sTargetDir.'/details.php');
@symlink(CONST_BasePath.'/website/reverse.php', $sTargetDir.'/reverse.php');
@symlink(CONST_BasePath.'/website/search.php', $sTargetDir.'/search.php');
@symlink(CONST_BasePath.'/website/search.php', $sTargetDir.'/index.php');
@symlink(CONST_BasePath.'/website/deletable.php', $sTargetDir.'/deletable.php');
@symlink(CONST_BasePath.'/website/polygons.php', $sTargetDir.'/polygons.php');
@symlink(CONST_BasePath.'/website/images', $sTargetDir.'/images');
@symlink(CONST_BasePath.'/website/js', $sTargetDir.'/js');
2012-11-19 20:00:19 +04:00
@symlink(CONST_BasePath.'/website/css', $sTargetDir.'/css');
echo "Symlinks created\n";
}
2010-10-26 19:22:41 +04:00
if (!$bDidSomething)
{
showUsage($aCMDOptions, true);
}
2010-10-27 18:05:42 +04:00
function pgsqlRunScriptFile($sFilename)
{
if (!file_exists($sFilename)) fail('unable to find '.$sFilename);
2012-07-10 17:36:17 +04:00
// Convert database DSN to psql parameters
2010-10-27 18:05:42 +04:00
$aDSNInfo = DB::parseDSN(CONST_Database_DSN);
2012-04-30 04:05:06 +04:00
if (!isset($aDSNInfo['port']) || !$aDSNInfo['port']) $aDSNInfo['port'] = 5432;
$sCMD = 'psql -p '.$aDSNInfo['port'].' -d '.$aDSNInfo['database'].' -f '.$sFilename;
2010-10-27 18:05:42 +04:00
$aDescriptors = array(
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 pgsql');
fclose($ahPipes[0]);
// TODO: error checking
while(!feof($ahPipes[1]))
{
echo fread($ahPipes[1], 4096);
}
fclose($ahPipes[1]);
proc_close($hProcess);
}
function pgsqlRunScript($sScript)
{
2012-07-10 17:36:17 +04:00
// Convert database DSN to psql parameters
2010-10-27 18:05:42 +04:00
$aDSNInfo = DB::parseDSN(CONST_Database_DSN);
if (!isset($aDSNInfo['port']) || !$aDSNInfo['port']) $aDSNInfo['port'] = 5432;
2012-04-30 04:05:06 +04:00
$sCMD = 'psql -p '.$aDSNInfo['port'].' -d '.$aDSNInfo['database'];
2010-10-27 18:05:42 +04:00
$aDescriptors = array(
0 => array('pipe', 'r'),
1 => STDOUT,
2 => STDERR
2010-10-27 18:05:42 +04:00
);
$ahPipes = null;
2011-06-14 17:42:46 +04:00
$hProcess = @proc_open($sCMD, $aDescriptors, $ahPipes);
2010-10-27 18:05:42 +04:00
if (!is_resource($hProcess)) fail('unable to start pgsql');
while(strlen($sScript))
2010-10-27 18:05:42 +04:00
{
$written = fwrite($ahPipes[0], $sScript);
$sScript = substr($sScript, $written);
2010-10-27 18:05:42 +04:00
}
fclose($ahPipes[0]);
2010-10-27 18:05:42 +04:00
proc_close($hProcess);
}
2012-04-30 04:05:06 +04:00
function pgsqlRunRestoreData($sDumpFile)
{
2012-07-10 17:36:17 +04:00
// Convert database DSN to psql parameters
2012-04-30 04:05:06 +04:00
$aDSNInfo = DB::parseDSN(CONST_Database_DSN);
if (!isset($aDSNInfo['port']) || !$aDSNInfo['port']) $aDSNInfo['port'] = 5432;
$sCMD = 'pg_restore -p '.$aDSNInfo['port'].' -d '.$aDSNInfo['database'].' -Fc -a '.$sDumpFile;
$aDescriptors = array(
0 => array('pipe', 'r'),
1 => array('pipe', 'w'),
2 => array('file', '/dev/null', 'a')
);
$ahPipes = null;
$hProcess = proc_open($sCMD, $aDescriptors, $ahPipes);
if (!is_resource($hProcess)) fail('unable to start pg_restore');
fclose($ahPipes[0]);
// TODO: error checking
while(!feof($ahPipes[1]))
{
echo fread($ahPipes[1], 4096);
}
fclose($ahPipes[1]);
proc_close($hProcess);
}
function pgsqlRunDropAndRestore($sDumpFile)
{
2012-07-10 17:36:17 +04:00
// Convert database DSN to psql parameters
$aDSNInfo = DB::parseDSN(CONST_Database_DSN);
if (!isset($aDSNInfo['port']) || !$aDSNInfo['port']) $aDSNInfo['port'] = 5432;
$sCMD = 'pg_restore -p '.$aDSNInfo['port'].' -d '.$aDSNInfo['database'].' -Fc --clean '.$sDumpFile;
$aDescriptors = array(
0 => array('pipe', 'r'),
1 => array('pipe', 'w'),
2 => array('file', '/dev/null', 'a')
);
$ahPipes = null;
$hProcess = proc_open($sCMD, $aDescriptors, $ahPipes);
if (!is_resource($hProcess)) fail('unable to start pg_restore');
fclose($ahPipes[0]);
// TODO: error checking
while(!feof($ahPipes[1]))
{
echo fread($ahPipes[1], 4096);
}
fclose($ahPipes[1]);
proc_close($hProcess);
}
function passthruCheckReturn($cmd)
{
$result = -1;
passthru($cmd, $result);
if ($result != 0) fail('Error executing external command: '.$cmd);
}