Files
Nominatim/utils/setup.php

290 lines
8.8 KiB
PHP
Executable File

#!/usr/bin/php -Cq
<?php
require_once(dirname(dirname(__FILE__)).'/lib/init-cmd.php');
ini_set('memory_limit', '800M');
$aCMDOptions = array(
"Create and setup nominatim search system",
array('help', 'h', 0, 1, 0, 0, false, 'Show Help'),
array('quiet', 'q', 0, 1, 0, 0, 'bool', 'Quiet output'),
array('verbose', 'v', 0, 1, 0, 0, 'bool', 'Verbose output'),
array('osm-file', '', 0, 1, 1, 1, 'realpath', 'File to import'),
array('threads', '', 0, 1, 1, 1, 'int', 'Number of threads (where possible)'),
array('all', '', 0, 1, 0, 0, 'bool', 'Do the complete process'),
array('create-db', '', 0, 1, 0, 0, 'bool', 'Create nominatim db'),
array('setup-db', '', 0, 1, 0, 0, 'bool', 'Build a blank nominatim db'),
array('import-data', '', 0, 1, 0, 0, 'bool', 'Import a osm file'),
array('create-functions', '', 0, 1, 0, 0, 'bool', 'Create functions'),
array('create-tables', '', 0, 1, 0, 0, 'bool', 'Create main tables'),
array('create-partitions', '', 0, 1, 0, 0, 'bool', 'Create required partition tables and triggers'),
array('load-data', '', 0, 1, 0, 0, 'bool', 'Copy data to live tables from import table'),
array('import-tiger-data', '', 0, 1, 0, 0, 'bool', 'Import tiger data'),
);
getCmdOpt($_SERVER['argv'], $aCMDOptions, $aCMDResult, true, true);
$bDidSomething = false;
// This is a pretty hard core defult - 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 $iInstances\n";
}
if ($aCMDResult['create-db'] || $aCMDResult['all'])
{
$bDidSomething = true;
$oDB =& DB::connect(CONST_Database_DSN, false);
if (!PEAR::isError($oDB))
{
fail('database already exists');
}
passthru('createdb nominatim');
}
if ($aCMDResult['create-db'] || $aCMDResult['all'])
{
$bDidSomething = true;
// TODO: path detection, detection memory, etc.
$oDB =& getDB();
passthru('createlang plpgsql nominatim');
pgsqlRunScriptFile(CONST_Path_Postgresql_Contrib.'/_int.sql');
pgsqlRunScriptFile(CONST_Path_Postgresql_Contrib.'/hstore.sql');
pgsqlRunScriptFile(CONST_Path_Postgresql_Postgis.'/postgis.sql');
pgsqlRunScriptFile(CONST_Path_Postgresql_Postgis.'/spatial_ref_sys.sql');
pgsqlRunScriptFile(CONST_BasePath.'/data/country_name.sql');
pgsqlRunScriptFile(CONST_BasePath.'/data/country_naturaleathdata.sql');
pgsqlRunScriptFile(CONST_BasePath.'/data/country_osm_grid.sql');
pgsqlRunScriptFile(CONST_BasePath.'/data/gb_postcode.sql');
pgsqlRunScriptFile(CONST_BasePath.'/data/us_statecounty.sql');
pgsqlRunScriptFile(CONST_BasePath.'/data/us_state.sql');
pgsqlRunScriptFile(CONST_BasePath.'/data/worldboundaries.sql');
}
if ($aCMDResult['import-data'] || $aCMDResult['all'])
{
$bDidSomething = true;
passthru(CONST_BasePath.'/osm2pgsql/osm2pgsql -lsc -O gazetteer -C 10000 --hstore -d nominatim '.$aCMDResult['osm-file']);
}
if ($aCMDResult['create-functions'] || $aCMDResult['all'])
{
$bDidSomething = true;
if (!file_exists(CONST_BasePath.'/module/nominatim.so')) fail("nominatim module not built");
$sTemplate = file_get_contents(CONST_BasePath.'/sql/functions.sql');
$sTemplate = str_replace('{modulepath}',CONST_BasePath.'/module', $sTemplate);
pgsqlRunScript($sTemplate);
}
if ($aCMDResult['create-tables'] || $aCMDResult['all'])
{
$bDidSomething = true;
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);
}
if ($aCMDResult['create-partitions'] || $aCMDResult['all'])
{
$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/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, $sTemplate);
}
pgsqlRunScript($sTemplate);
}
if ($aCMDResult['load-data'] || $aCMDResult['all'])
{
$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 '.';
$aDBInstances = array();
for($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.' = '.$i;
if ($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'] || $aCMDResult['all'])
{
$bDidSomething = true;
$aDBInstances = array();
for($i = 0; $i < $iInstances; $i++)
{
$aDBInstances[$i] =& getDB(true);
}
foreach(glob(CONST_BasePath.'/data/tiger2009/*.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);
$bAnyBusy = true;
while($bAnyBusy)
{
$bAnyBusy = false;
for($i = 0; $i < $iInstances; $i++)
{
if (pg_connection_busy($aDBInstances[$i]->connection)) $bAnyBusy = true;
}
usleep(10);
}
echo "\n";
}
}
if (!$bDidSomething)
{
showUsage($aCMDOptions, true);
}
function pgsqlRunScriptFile($sFilename)
{
if (!file_exists($sFilename)) fail('unable to find '.$sFilename);
// Convert database DSN to psql paramaters
$aDSNInfo = DB::parseDSN(CONST_Database_DSN);
$sCMD = 'psql -f '.$sFilename.' '.$aDSNInfo['database'];
$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)
{
// Convert database DSN to psql paramaters
$aDSNInfo = DB::parseDSN(CONST_Database_DSN);
$sCMD = 'psql '.$aDSNInfo['database'];
$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');
fwrite($ahPipes[0], $sScript);
fclose($ahPipes[0]);
// TODO: error checking
while(!feof($ahPipes[1]))
{
echo fread($ahPipes[1], 4096);
}
fclose($ahPipes[1]);
proc_close($hProcess);
}