X-Git-Url: https://git.openstreetmap.org./nominatim.git/blobdiff_plain/e1be3d9f48b51a927c26d1ac0035b50b82423916..115c909fdf65966774940ef957cf0bd63ba844f2:/utils/setup.php
diff --git a/utils/setup.php b/utils/setup.php
index dcce1ead..67c7fd7f 100755
--- a/utils/setup.php
+++ b/utils/setup.php
@@ -38,6 +38,7 @@ $aCMDOptions
array('index', '', 0, 1, 0, 0, 'bool', 'Index the data'),
array('index-noanalyse', '', 0, 1, 0, 0, 'bool', 'Do not perform analyse operations during index (EXPERT)'),
array('create-search-indices', '', 0, 1, 0, 0, 'bool', 'Create additional indices required for search and update'),
+ array('create-country-names', '', 0, 1, 0, 0, 'bool', 'Create default list of searchable country names'),
array('drop', '', 0, 1, 0, 0, 'bool', 'Drop tables needed for updates, making the database readonly (EXPERIMENTAL)'),
);
getCmdOpt($_SERVER['argv'], $aCMDOptions, $aCMDResult, true, true);
@@ -123,13 +124,23 @@ if ($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');
+ }
+
+ if (!file_exists(CONST_ExtraDataPath.'/country_osm_grid.sql.gz')) {
+ echo "Error: you need to download the country_osm_grid first:";
+ echo "\n wget -O ".CONST_ExtraDataPath."/country_osm_grid.sql.gz http://www.nominatim.org/data/country_grid.sql.gz\n";
+ exit(1);
}
pgsqlRunScriptFile(CONST_BasePath.'/data/country_name.sql');
pgsqlRunScriptFile(CONST_BasePath.'/data/country_naturalearthdata.sql');
- pgsqlRunScriptFile(CONST_BasePath.'/data/country_osm_grid.sql');
+ pgsqlRunScriptFile(CONST_BasePath.'/data/country_osm_grid.sql.gz');
pgsqlRunScriptFile(CONST_BasePath.'/data/gb_postcode_table.sql');
if (file_exists(CONST_BasePath.'/data/gb_postcode_data.sql.gz')) {
pgsqlRunScriptFile(CONST_BasePath.'/data/gb_postcode_data.sql.gz');
@@ -158,7 +169,8 @@ 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 '$osm2pgsql'");
}
@@ -180,7 +192,7 @@ if ($aCMDResult['import-data'] || $aCMDResult['all']) {
passthruCheckReturn($osm2pgsql);
$oDB =& getDB();
- if (!chksql($oDB->getRow('select * from place limit 1'))) {
+ if (!$aCMDResult['ignore-errors'] && !chksql($oDB->getRow('select * from place limit 1'))) {
fail('No Data');
}
}
@@ -286,8 +298,8 @@ if ($aCMDResult['create-partition-functions'] || $aCMDResult['all']) {
if ($aCMDResult['import-wikipedia-articles'] || $aCMDResult['all']) {
$bDidSomething = true;
- $sWikiArticlesFile = CONST_BasePath.'/data/wikipedia_article.sql.bin';
- $sWikiRedirectsFile = CONST_BasePath.'/data/wikipedia_redirect.sql.bin';
+ $sWikiArticlesFile = CONST_Wikipedia_Data_Path.'/wikipedia_article.sql.bin';
+ $sWikiRedirectsFile = CONST_Wikipedia_Data_Path.'/wikipedia_redirect.sql.bin';
if (file_exists($sWikiArticlesFile)) {
echo "Importing wikipedia articles...";
pgsqlRunDropAndRestore($sWikiArticlesFile);
@@ -340,7 +352,11 @@ if ($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
@@ -350,24 +366,31 @@ if ($aCMDResult['load-data'] || $aCMDResult['all']) {
}
echo "Load Data\n";
+ $sColumns = 'osm_type, osm_id, class, type, name, admin_level, address, extratags, geometry';
+
$aDBInstances = array();
$iLoadThreads = max(1, $iInstances - 1);
for ($i = 0; $i < $iLoadThreads; $i++) {
$aDBInstances[$i] =& getDB(true);
- $sSQL = 'insert into placex (osm_type, osm_id, class, type, name, admin_level, ';
- $sSQL .= 'housenumber, street, addr_place, isin, postcode, country_code, extratags, ';
- $sSQL .= 'geometry) select * from place where osm_id % '.$iLoadThreads.' = '.$i;
- $sSQL .= " and not (class='place' and type='houses' and osm_type='W' and ST_GeometryType(geometry) = 'ST_LineString')";
+ $sSQL = "INSERT INTO placex ($sColumns) SELECT $sColumns FROM place WHERE osm_id % $iLoadThreads = $i";
+ $sSQL .= " and not (class='place' and type='houses' and osm_type='W'";
+ $sSQL .= " and ST_GeometryType(geometry) = 'ST_LineString')";
+ $sSQL .= " and ST_IsValid(geometry)";
if ($aCMDResult['verbose']) echo "$sSQL\n";
- if (!pg_send_query($aDBInstances[$i]->connection, $sSQL)) fail(pg_last_error($oDB->connection));
+ if (!pg_send_query($aDBInstances[$i]->connection, $sSQL)) {
+ fail(pg_last_error($aDBInstances[$i]->connection));
+ }
}
// last thread for interpolation lines
$aDBInstances[$iLoadThreads] =& getDB(true);
- $sSQL = 'select insert_osmline (osm_id, housenumber, street, addr_place, postcode, country_code, ';
- $sSQL .= 'geometry) from place where ';
+ $sSQL = 'insert into location_property_osmline';
+ $sSQL .= ' (osm_id, address, linegeo)';
+ $sSQL .= ' SELECT osm_id, address, geometry from place where ';
$sSQL .= "class='place' and type='houses' and osm_type='W' and ST_GeometryType(geometry) = 'ST_LineString'";
if ($aCMDResult['verbose']) echo "$sSQL\n";
- if (!pg_send_query($aDBInstances[$i]->connection, $sSQL)) fail(pg_last_error($oDB->connection));
+ if (!pg_send_query($aDBInstances[$iLoadThreads]->connection, $sSQL)) {
+ fail(pg_last_error($aDBInstances[$iLoadThreads]->connection));
+ }
$bAnyBusy = true;
while ($bAnyBusy) {
@@ -381,6 +404,16 @@ if ($aCMDResult['load-data'] || $aCMDResult['all']) {
echo "\n";
echo "Reanalysing database...\n";
pgsqlRunScript('ANALYSE');
+
+ $sDatabaseDate = getDatabaseDate($oDB);
+ pg_query($oDB->connection, 'TRUNCATE import_status');
+ if ($sDatabaseDate === false) {
+ echo "WARNING: could not determine database date.\n";
+ } else {
+ $sSQL = "INSERT INTO import_status (lastimportdate) VALUES('".$sDatabaseDate."')";
+ pg_query($oDB->connection, $sSQL);
+ echo "Latest data imported from $sDatabaseDate.\n";
+ }
}
if ($aCMDResult['import-tiger-data']) {
@@ -460,101 +493,66 @@ if ($aCMDResult['import-tiger-data']) {
if ($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));
+ if (!pg_query($oDB->connection, 'TRUNCATE location_postcode')) {
+ fail(pg_last_error($oDB->connection));
+ }
+
+ $sSQL = "INSERT INTO location_postcode";
+ $sSQL .= " (place_id, indexed_status, country_code, postcode, geometry) ";
+ $sSQL .= "SELECT nextval('seq_place'), 1, country_code,";
+ $sSQL .= " upper(trim (both ' ' from address->'postcode')) as pc,";
+ $sSQL .= " ST_Centroid(ST_Collect(ST_Centroid(geometry)))";
+ $sSQL .= " FROM placex";
+ $sSQL .= " WHERE address ? 'postcode' AND address->'postcode' NOT SIMILAR TO '%(,|;)%'";
+ $sSQL .= " AND geometry IS NOT null";
+ $sSQL .= " GROUP BY country_code, pc";
+
+ if (!pg_query($oDB->connection, $sSQL)) {
+ fail(pg_last_error($oDB->connection));
+ }
if (CONST_Use_Extra_US_Postcodes) {
- $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";
+ // only add postcodes that are not yet available in OSM
+ $sSQL = "INSERT INTO location_postcode";
+ $sSQL .= " (place_id, indexed_status, country_code, postcode, geometry) ";
+ $sSQL .= "SELECT nextval('seq_place'), 1, 'us', postcode,";
+ $sSQL .= " ST_SetSRID(ST_Point(x,y),4326)";
+ $sSQL .= " FROM us_postcode WHERE postcode NOT IN";
+ $sSQL .= " (SELECT postcode FROM location_postcode";
+ $sSQL .= " WHERE country_code = 'us')";
+
if (!pg_query($oDB->connection, $sSQL)) fail(pg_last_error($oDB->connection));
}
-}
-if ($aCMDResult['osmosis-init'] || ($aCMDResult['all'] && !$aCMDResult['drop'])) { // no use doing osmosis-init when dropping update tables
- $bDidSomething = true;
- $oDB =& getDB();
+ // add missing postcodes for GB (if available)
+ $sSQL = "INSERT INTO location_postcode";
+ $sSQL .= " (place_id, indexed_status, country_code, postcode, geometry) ";
+ $sSQL .= "SELECT nextval('seq_place'), 1, 'gb', postcode, geometry";
+ $sSQL .= " FROM gb_postcode WHERE postcode NOT IN";
+ $sSQL .= " (SELECT postcode FROM location_postcode";
+ $sSQL .= " WHERE country_code = 'gb')";
+ if (!pg_query($oDB->connection, $sSQL)) fail(pg_last_error($oDB->connection));
- if (!file_exists(CONST_Osmosis_Binary)) {
- echo "Please download osmosis.\nIf it is already installed, check the path in your local settings (settings/local.php) file.\n";
- if (!$aCMDResult['all']) {
- fail("osmosis not found in '".CONST_Osmosis_Binary."'");
- }
- } else {
- if (file_exists(CONST_InstallPath.'/settings/configuration.txt')) {
- echo "settings/configuration.txt already exists\n";
- } else {
- passthru(CONST_Osmosis_Binary.' --read-replication-interval-init '.CONST_InstallPath.'/settings');
- // update osmosis configuration.txt with our settings
- passthru("sed -i 's!baseUrl=.*!baseUrl=".CONST_Replication_Url."!' ".CONST_InstallPath.'/settings/configuration.txt');
- passthru("sed -i 's:maxInterval = .*:maxInterval = ".CONST_Replication_MaxInterval.":' ".CONST_InstallPath.'/settings/configuration.txt');
+ if (!$aCMDResult['all']) {
+ $sSQL = "DELETE FROM word WHERE class='place' and type='postcode'";
+ $sSQL .= "and word NOT IN (SELECT postcode FROM location_postcode)";
+ if (!pg_query($oDB->connection, $sSQL)) {
+ fail(pg_last_error($oDB->connection));
}
+ }
+ $sSQL = "SELECT count(getorcreate_postcode_id(v)) FROM ";
+ $sSQL .= "(SELECT distinct(postcode) as v FROM location_postcode) p";
- // Find the last node in the DB
- $iLastOSMID = $oDB->getOne("select max(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."/1";
- $sLastNodeXML = file_get_contents($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 so need to sort by date descending
- $sRepURL = CONST_Replication_Url."/";
- $sRep = file_get_contents($sRepURL."?C=M;O=D;F=1");
- // download.geofabrik.de: 000/| 26-Feb-2013 11:53 |
- // planet.openstreetmap.org: 273/ 2013-03-11 07:41 -
- preg_match_all('#([0-9]{3}/)\s*([-0-9a-zA-Z]+ [0-9]{2}:[0-9]{2})#', $sRep, $aRepMatches, PREG_SET_ORDER);
- if ($aRepMatches) {
- $aPrevRepMatch = false;
- foreach ($aRepMatches as $aRepMatch) {
- if (strtotime($aRepMatch[2]) < $iLastNodeTimestamp) break;
- $aPrevRepMatch = $aRepMatch;
- }
- if ($aPrevRepMatch) $aRepMatch = $aPrevRepMatch;
-
- $sRepURL .= $aRepMatch[1];
- $sRep = file_get_contents($sRepURL."?C=M;O=D;F=1");
- preg_match_all('#([0-9]{3}/)\s*([-0-9a-zA-Z]+ [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."?C=M;O=D;F=1");
- preg_match_all('#([0-9]{3}).state.txt\s*([-0-9a-zA-Z]+ [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_InstallPath.'/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, $sSQL);
- } else {
- if (!$aCMDResult['all']) {
- fail("Cannot read state file directory.");
- }
- }
+ if (!pg_query($oDB->connection, $sSQL)) {
+ fail(pg_last_error($oDB->connection));
}
}
+if ($aCMDResult['osmosis-init']) {
+ $bDidSomething = true;
+ echo "Command 'osmosis-init' no longer available, please use utils/update.php --init-updates.\n";
+}
+
if ($aCMDResult['index'] || $aCMDResult['all']) {
$bDidSomething = true;
$sOutputFile = '';
@@ -564,6 +562,11 @@ if ($aCMDResult['index'] || $aCMDResult['all']) {
passthruCheckReturn($sBaseCmd.' -r 5 -R 25');
if (!$aCMDResult['index-noanalyse']) pgsqlRunScript('ANALYSE');
passthruCheckReturn($sBaseCmd.' -r 26');
+
+ echo "Indexing postcodes....\n";
+ $oDB =& getDB();
+ $sSQL = 'UPDATE location_postcode SET indexed_status = 0';
+ if (!pg_query($oDB->connection, $sSQL)) fail(pg_last_error($oDB->connection));
}
if ($aCMDResult['create-search-indices'] || $aCMDResult['all']) {
@@ -590,6 +593,32 @@ if ($aCMDResult['create-search-indices'] || $aCMDResult['all']) {
pgsqlRunScript($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.