X-Git-Url: https://git.openstreetmap.org/nominatim.git/blobdiff_plain/b11718699ede4fc152dae5f2dd14a27d739b4cf3..7dbbc43a85dd6c46110493e63256149cbc89491c:/utils/setup.php

diff --git a/utils/setup.php b/utils/setup.php
index c842ba33..a203db7b 100755
--- a/utils/setup.php
+++ b/utils/setup.php
@@ -32,7 +32,6 @@
 		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'),
@@ -43,6 +42,26 @@
 
 	$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');
+		}
+	}
+
+
 	// 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)
@@ -64,16 +83,6 @@
 		echo "WARNING: resetting cache memory to $iCacheMemory\n";
 	}
 
-	if (isset($aCMDResult['osm-file']) && !isset($aCMDResult['osmosis-init-date']))
-	{
-		$sBaseFile = basename($aCMDResult['osm-file']);
-		if (preg_match('#^planet-([0-9]{2})([0-9]{2})([0-9]{2})[.]#', $sBaseFile, $aMatch))
-		{
-			$iTime = mktime(0, 0, 0, $aMatch[2], $aMatch[3], '20'.$aMatch[1]);
-			$iTime -= (60*60*24);
-			$aCMDResult['osmosis-init-date'] = date('Y-m-d', $iTime).'T22:00:00Z';
-		}
-	}
 	$aDSNInfo = DB::parseDSN(CONST_Database_DSN);
 	if (!isset($aDSNInfo['port']) || !$aDSNInfo['port']) $aDSNInfo['port'] = 5432;
 
@@ -89,9 +98,9 @@
 		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.
 
@@ -251,7 +260,7 @@
 
 	if ($aCMDResult['load-data'] || $aCMDResult['all'])
 	{
-		echo "Load Data\n";
+		echo "Drop old Data\n";
 		$bDidSomething = true;
 
 		$oDB =& getDB();
@@ -287,17 +296,18 @@
 			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($oDB->connection));
+		echo ".\n";
+
 		// pre-create the word list
 		if (!$aCMDResult['disable-token-precalc'])
 		{
-			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 '.';
+			echo "Loading word list\n";
+			pgsqlRunScriptFile(CONST_BasePath.'/data/words.sql');
 		}
 
+		echo "Load Data\n";
 		$aDBInstances = array();
 		for($i = 0; $i < $iInstances; $i++)
 		{
@@ -334,7 +344,7 @@
 		{
 			$aDBInstances[$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, country_code, geometry) from ';
+			$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));
@@ -357,6 +367,8 @@
 	{
 		$bDidSomething = true;
 
+		pgsqlRunScriptFile(CONST_BasePath.'/sql/tiger_import_start.sql');
+
 		$aDBInstances = array();
 		for($i = 0; $i < $iInstances; $i++)
 		{
@@ -405,6 +417,9 @@
 			}
 			echo "\n";
 		}
+
+		echo "Creating indexes\n";
+		pgsqlRunScriptFile(CONST_BasePath.'/sql/tiger_import_finish.sql');
 	}
 
 	if ($aCMDResult['calculate-postcodes'] || $aCMDResult['all'])
@@ -412,43 +427,90 @@
 		$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,country_code,geometry) ";
-		$sSQL .= "select 'P',nextval('seq_postcodes'),'place','postcode',postcode,country_code,";
-		$sSQL .= "ST_SetSRID(ST_Point(x,y),4326) as geometry from (select country_code,postcode,";
+		$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 country_code,postcode) as x";
+		$sSQL .= "from placex where postcode is not null group by calculated_country_code,postcode) as x";
 		if (!pg_query($oDB->connection, $sSQL)) fail(pg_last_error($oDB->connection));
 
-		$sSQL = "insert into placex (osm_type,osm_id,class,type,postcode,country_code,geometry) ";
+		$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'] || $aCMDResult['all']) && isset($aCMDResult['osmosis-init-date']))
+	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');
-
-		$sDate = $aCMDResult['osmosis-init-date'];
-		$aDate = date_parse_from_format("Y-m-d\TH-i", $sDate);
-		$sURL = 'http://toolserver.org/~mazder/replicate-sequences/?';
-		$sURL .= 'Y='.$aDate['year'].'&m='.$aDate['month'].'&d='.$aDate['day'];
-		$sURL .= '&H='.$aDate['hour'].'&i='.$aDate['minute'].'&s=0';
-		$sURL .= '&stream=minute';
-		echo "Getting state file: $sURL\n";
-		$sStateFile = file_get_contents($sURL);
+		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');
-		$sSQL = "INSERT INTO import_status VALUES('".$sDate."')";
+		$sSQL = "INSERT INTO import_status VALUES('".$aRepMatch[2]."')";
 		pg_query($oDB->connection, $sSQL);
-
+		exit;
 	}
 
 	if ($aCMDResult['index'] || $aCMDResult['all'])