X-Git-Url: https://git.openstreetmap.org/nominatim.git/blobdiff_plain/c32fa2172d7d688e3b816957cd1e92f6ca129244..8fc2f9551bb056f2dcdd8f85f168ca956015e266:/utils/setup.php

diff --git a/utils/setup.php b/utils/setup.php
index b40bad3a..bfc7c720 100755
--- a/utils/setup.php
+++ b/utils/setup.php
@@ -20,17 +20,21 @@
 		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('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('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('osmosis-init-date', '', 0, 1, 1, 1, 'string', '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'),
@@ -39,7 +43,27 @@
 
 	$bDidSomething = false;
 
-	// This is a pretty hard core defult - the number of processors in the box - 1
+	// 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');
+		}
+	}
+
+
+	// 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)
 	{
@@ -85,9 +109,9 @@
 		passthru('createdb -E UTF-8 '.$aDSNInfo['database']);
 	}
 
-	if ($aCMDResult['create-db'] || $aCMDResult['all'])
+	if ($aCMDResult['setup-db'] || $aCMDResult['all'])
 	{
-		echo "Create DB (2)\n";
+		echo "Setup DB\n";
 		$bDidSomething = true;
 		// TODO: path detection, detection memory, etc.
 
@@ -117,11 +141,15 @@
 		$bDidSomething = true;
 
 		$osm2pgsql = CONST_Osm2pgsql_Binary;
-		if (!file_exists($osm2pgsql)) fail("please download and build osm2pgsql");
+		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;
 		$osm2pgsql .= ' -d '.$aDSNInfo['database'].' '.$aCMDResult['osm-file'];
-		passthru($osm2pgsql);
+		passthruCheckReturn($osm2pgsql);
 
 		$oDB =& getDB();
 		$x = $oDB->getRow('select * from place limit 1');
@@ -137,7 +165,9 @@
 		$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);
+		$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:', '', $sTemplate);
 		pgsqlRunScript($sTemplate);
 	}
 
@@ -219,7 +249,7 @@
 		if (file_exists($sWikiArticlesFile))
 		{
 			echo "Importing wikipedia articles...";
-			pgsqlRunRestoreData($sWikiArticlesFile);
+			pgsqlRunDropAndRestore($sWikiArticlesFile);
 			echo "...done\n";
 		}
 		else
@@ -229,7 +259,7 @@
 		if (file_exists($sWikiRedirectsFile))
 		{
 			echo "Importing wikipedia redirects...";
-			pgsqlRunRestoreData($sWikiRedirectsFile);
+			pgsqlRunDropAndRestore($sWikiRedirectsFile);
 			echo "...done\n";
 		}
 		else
@@ -241,7 +271,7 @@
 
 	if ($aCMDResult['load-data'] || $aCMDResult['all'])
 	{
-		echo "Load Data\n";
+		echo "Drop old Data\n";
 		$bDidSomething = true;
 
 		$oDB =& getDB();
@@ -264,14 +294,31 @@
 		if (!pg_query($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($oDB->connection));
+		echo ".\n";
+
 		// pre-create the word list
-		if (!pg_query($oDB->connection, 'select count(make_keywords(v)) from (select distinct svals(name) as v from place) as w where v is not null;')) fail(pg_last_error($oDB->connection));
-		echo '.';
-		if (!pg_query($oDB->connection, 'select count(make_keywords(v)) from (select distinct postcode as v from place) as w where v is not null;')) fail(pg_last_error($oDB->connection));
-		echo '.';
-		if (!pg_query($oDB->connection, 'select count(getorcreate_housenumber_id(v)) from (select distinct housenumber as v from place where housenumber is not null) as w;')) fail(pg_last_error($oDB->connection));
-		echo '.';
+		if (!$aCMDResult['disable-token-precalc'])
+		{
+			echo "Loading word list\n";
+			pgsqlRunScriptFile(CONST_BasePath.'/data/words.sql');
+		}
 
+		echo "Load Data\n";
 		$aDBInstances = array();
 		for($i = 0; $i < $iInstances; $i++)
 		{
@@ -331,6 +378,8 @@
 	{
 		$bDidSomething = true;
 
+		pgsqlRunScriptFile(CONST_BasePath.'/sql/tiger_import_start.sql');
+
 		$aDBInstances = array();
 		for($i = 0; $i < $iInstances; $i++)
 		{
@@ -379,6 +428,9 @@
 			}
 			echo "\n";
 		}
+
+		echo "Creating indexes\n";
+		pgsqlRunScriptFile(CONST_BasePath.'/sql/tiger_import_finish.sql');
 	}
 
 	if ($aCMDResult['calculate-postcodes'] || $aCMDResult['all'])
@@ -406,7 +458,12 @@
 
 		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');
+		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');
+		}
 
 		$sDate = $aCMDResult['osmosis-init-date'];
 		$aDate = date_parse_from_format("Y-m-d\TH-i", $sDate);
@@ -431,11 +488,11 @@
 		$sOutputFile = '';
 		if (isset($aCMDResult['index-output'])) $sOutputFile = ' -F '.$aCMDResult['index-output'];
 		$sBaseCmd = CONST_BasePath.'/nominatim/nominatim -i -d '.$aDSNInfo['database'].' -t '.$iInstances.$sOutputFile;
-		passthru($sBaseCmd.' -R 4');
-		pgsqlRunScript('ANALYSE');
-		passthru($sBaseCmd.' -r 5 -R 25');
-		pgsqlRunScript('ANALYSE');
-		passthru($sBaseCmd.' -r 26');
+		passthruCheckReturn($sBaseCmd.' -R 4');
+		if (!$aCMDResult['index-noanalyse']) pgsqlRunScript('ANALYSE');
+		passthruCheckReturn($sBaseCmd.' -r 5 -R 25');
+		if (!$aCMDResult['index-noanalyse']) pgsqlRunScript('ANALYSE');
+		passthruCheckReturn($sBaseCmd.' -r 26');
 	}
 
 	if ($aCMDResult['create-search-indices'] || $aCMDResult['all'])
@@ -470,7 +527,12 @@
 	{
 		$bDidSomething = true;
 		$sTargetDir = $aCMDResult['create-website'];
-		if (!is_dir($sTargetDir)) fail('please specify a directory to setup');
+		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', $sTargetDir.'/search.php');
@@ -489,7 +551,7 @@
 	{
 		if (!file_exists($sFilename)) fail('unable to find '.$sFilename);
 
-		// Convert database DSN to psql paramaters
+		// Convert database DSN to psql parameters
 		$aDSNInfo = DB::parseDSN(CONST_Database_DSN);
 		if (!isset($aDSNInfo['port']) || !$aDSNInfo['port']) $aDSNInfo['port'] = 5432;
 		$sCMD = 'psql -p '.$aDSNInfo['port'].' -d '.$aDSNInfo['database'].' -f '.$sFilename;
@@ -517,7 +579,7 @@
 
 	function pgsqlRunScript($sScript)
 	{
-		// Convert database DSN to psql paramaters
+		// Convert database DSN to psql parameters
 		$aDSNInfo = DB::parseDSN(CONST_Database_DSN);
 		if (!isset($aDSNInfo['port']) || !$aDSNInfo['port']) $aDSNInfo['port'] = 5432;
 		$sCMD = 'psql -p '.$aDSNInfo['port'].' -d '.$aDSNInfo['database'];
@@ -541,7 +603,7 @@
 
 	function pgsqlRunRestoreData($sDumpFile)
 	{
-		// Convert database DSN to psql paramaters
+		// 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 '.$aDSNInfo['database'].' -Fc -a '.$sDumpFile;
@@ -566,3 +628,38 @@
 
 		proc_close($hProcess);
 	}
+
+	function pgsqlRunDropAndRestore($sDumpFile)
+	{
+		// 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);
+	}