X-Git-Url: https://git.openstreetmap.org./nominatim.git/blobdiff_plain/6c1977b448e8b195bf96b6144674ffe0527e79de..b3f1f45f97ebbe378be8e9e3b09f90aa5e22c0e1:/utils/update.php

diff --git a/utils/update.php b/utils/update.php
old mode 100755
new mode 100644
index 5a5d41ff..f71c4f43
--- a/utils/update.php
+++ b/utils/update.php
@@ -1,10 +1,15 @@
-#!/usr/bin/php -Cq
 <?php
 
-require_once(dirname(dirname(__FILE__)).'/settings/settings.php');
 require_once(CONST_BasePath.'/lib/init-cmd.php');
+require_once(CONST_BasePath.'/lib/setup_functions.php');
+require_once(CONST_BasePath.'/lib/setup/SetupClass.php');
+require_once(CONST_BasePath.'/lib/setup/AddressLevelParser.php');
+
 ini_set('memory_limit', '800M');
 
+use Nominatim\Setup\SetupFunctions as SetupFunctions;
+
+// (long-opt, short-opt, min-occurs, max-occurs, num-arguments, num-arguments, type, help)
 $aCMDOptions
 = array(
    'Import / update / index osm data',
@@ -13,6 +18,8 @@ $aCMDOptions
    array('verbose', 'v', 0, 1, 0, 0, 'bool', 'Verbose output'),
 
    array('init-updates', '', 0, 1, 0, 0, 'bool', 'Set up database for updating'),
+   array('check-for-updates', '', 0, 1, 0, 0, 'bool', 'Check if new updates are available'),
+   array('no-update-functions', '', 0, 1, 0, 0, 'bool', 'Do not update trigger functions to support differential updates (assuming the diff update logic is already present)'),
    array('import-osmosis', '', 0, 1, 0, 0, 'bool', 'Import updates once'),
    array('import-osmosis-all', '', 0, 1, 0, 0, 'bool', 'Import updates forever'),
    array('no-index', '', 0, 1, 0, 0, 'bool', 'Do not index the new data'),
@@ -33,18 +40,24 @@ $aCMDOptions
    array('index-instances', '', 0, 1, 1, 1, 'int', 'Number of indexing instances (threads)'),
 
    array('deduplicate', '', 0, 1, 0, 0, 'bool', 'Deduplicate tokens'),
+   array('recompute-word-counts', '', 0, 1, 0, 0, 'bool', 'Compute frequency of full-word search terms'),
+   array('update-address-levels', '', 0, 1, 0, 0, 'bool', 'Reimport address level configuration (EXPERT)'),
+   array('recompute-importance', '', 0, 1, 0, 0, 'bool', 'Recompute place importances'),
    array('no-npi', '', 0, 1, 0, 0, 'bool', '(obsolete)'),
   );
+
 getCmdOpt($_SERVER['argv'], $aCMDOptions, $aResult, true, true);
 
 if (!isset($aResult['index-instances'])) $aResult['index-instances'] = 1;
+
 if (!isset($aResult['index-rank'])) $aResult['index-rank'] = 0;
 
 date_default_timezone_set('Etc/UTC');
 
-$oDB =& getDB();
+$oDB = new Nominatim\DB();
+$oDB->connect();
 
-$aDSNInfo = DB::parseDSN(CONST_Database_DSN);
+$aDSNInfo = Nominatim\DB::parseDSN(CONST_Database_DSN);
 if (!isset($aDSNInfo['port']) || !$aDSNInfo['port']) $aDSNInfo['port'] = 5432;
 
 // cache memory to be used by osm2pgsql, should not be more than the available memory
@@ -53,8 +66,19 @@ if ($iCacheMemory + 500 > getTotalMemoryMB()) {
     $iCacheMemory = getCacheMemoryMB();
     echo "WARNING: resetting cache memory to $iCacheMemory\n";
 }
-$sOsm2pgsqlCmd = CONST_Osm2pgsql_Binary.' -klas --number-processes 1 -C '.$iCacheMemory.' -O gazetteer -d '.$aDSNInfo['database'].' -P '.$aDSNInfo['port'];
-if (!is_null(CONST_Osm2pgsql_Flatnode_File)) {
+$sOsm2pgsqlCmd = CONST_Osm2pgsql_Binary.' -klas --number-processes 1 -C '.$iCacheMemory.' -O gazetteer -S '.CONST_Import_Style.' -d '.$aDSNInfo['database'].' -P '.$aDSNInfo['port'];
+if (isset($aDSNInfo['username']) && $aDSNInfo['username']) {
+    $sOsm2pgsqlCmd .= ' -U ' . $aDSNInfo['username'];
+}
+if (isset($aDSNInfo['hostspec']) && $aDSNInfo['hostspec']) {
+    $sOsm2pgsqlCmd .= ' -H ' . $aDSNInfo['hostspec'];
+}
+$aProcEnv = null;
+if (isset($aDSNInfo['password']) && $aDSNInfo['password']) {
+    $aProcEnv = array_merge(array('PGPASSWORD' => $aDSNInfo['password']), $_ENV);
+}
+
+if (!is_null(CONST_Osm2pgsql_Flatnode_File) && CONST_Osm2pgsql_Flatnode_File) {
     $sOsm2pgsqlCmd .= ' --flat-nodes '.CONST_Osm2pgsql_Flatnode_File;
 }
 
@@ -66,15 +90,35 @@ if ($aResult['init-updates']) {
         echo "Does the URL point to a directory containing OSM update data?\n\n";
         fail('replication URL not reachable.');
     }
-    $sSetup = CONST_InstallPath.'/utils/setup.php';
-    $iRet = -1;
-    passthru($sSetup.' --create-functions --enable-diff-updates', $iRet);
+    // sanity check for pyosmium-get-changes
+    if (!CONST_Pyosmium_Binary) {
+        echo "\nCONST_Pyosmium_Binary not configured.\n";
+        echo "You need to install pyosmium and set up the path to pyosmium-get-changes\n";
+        echo "in your local settings file.\n\n";
+        fail('CONST_Pyosmium_Binary not configured');
+    }
+    $aOutput = 0;
+    $sCmd = CONST_Pyosmium_Binary.' --help';
+    exec($sCmd, $aOutput, $iRet);
     if ($iRet != 0) {
-        fail('Error running setup script');
+        echo "Cannot execute pyosmium-get-changes.\n";
+        echo "Make sure you have pyosmium installed correctly\n";
+        echo "and have set up CONST_Pyosmium_Binary to point to pyosmium-get-changes.\n";
+        fail('pyosmium-get-changes not found or not usable');
+    }
+
+    if (!$aResult['no-update-functions']) {
+        // instantiate setupClass to use the function therein
+        $cSetup = new SetupFunctions(array(
+                                      'enable-diff-updates' => true,
+                                      'verbose' => $aResult['verbose']
+                                     ));
+        $cSetup->connect();
+        $cSetup->createFunctions();
     }
 
     $sDatabaseDate = getDatabaseDate($oDB);
-    if ($sDatabaseDate === false) {
+    if (!$sDatabaseDate) {
         fail('Cannot determine date of database.');
     }
     $sWindBack = strftime('%Y-%m-%dT%H:%M:%SZ', strtotime($sDatabaseDate) - (3*60*60));
@@ -87,16 +131,30 @@ if ($aResult['init-updates']) {
         fail('Error running pyosmium tools');
     }
 
-    pg_query($oDB->connection, 'TRUNCATE import_status');
+    $oDB->exec('TRUNCATE import_status');
     $sSQL = "INSERT INTO import_status (lastimportdate, sequence_id, indexed) VALUES('";
     $sSQL .= $sDatabaseDate."',".$aOutput[0].', true)';
-    if (!pg_query($oDB->connection, $sSQL)) {
+
+    try {
+        $oDB->exec($sSQL);
+    } catch (\Nominatim\DatabaseError $e) {
         fail('Could not enter sequence into database.');
     }
 
     echo "Done. Database updates will start at sequence $aOutput[0] ($sWindBack)\n";
 }
 
+if ($aResult['check-for-updates']) {
+    $aLastState = $oDB->getRow('SELECT sequence_id FROM import_status');
+
+    if (!$aLastState['sequence_id']) {
+        fail('Updates not set up. Please run ./utils/update.php --init-updates.');
+    }
+
+    system(CONST_BasePath.'/utils/check_server_for_updates.py '.CONST_Replication_Url.' '.$aLastState['sequence_id'], $iRet);
+    exit($iRet);
+}
+
 if (isset($aResult['import-diff']) || isset($aResult['import-file'])) {
     // import diffs and files directly (e.g. from osmosis --rri)
     $sNextFile = isset($aResult['import-diff']) ? $aResult['import-diff'] : $aResult['import-file'];
@@ -108,7 +166,7 @@ if (isset($aResult['import-diff']) || isset($aResult['import-file'])) {
     // Import the file
     $sCMD = $sOsm2pgsqlCmd.' '.$sNextFile;
     echo $sCMD."\n";
-    exec($sCMD, $sJunk, $iErrorLevel);
+    $iErrorLevel = runWithEnv($sCMD, $aProcEnv);
 
     if ($iErrorLevel) {
         fail("Error from osm2pgsql, $iErrorLevel\n");
@@ -129,25 +187,25 @@ $bUseOSMApi = isset($aResult['import-from-main-api']) && $aResult['import-from-m
 $sContentURL = '';
 if (isset($aResult['import-node']) && $aResult['import-node']) {
     if ($bUseOSMApi) {
-        $sContentURL = 'http://www.openstreetmap.org/api/0.6/node/'.$aResult['import-node'];
+        $sContentURL = 'https://www.openstreetmap.org/api/0.6/node/'.$aResult['import-node'];
     } else {
-        $sContentURL = 'http://overpass-api.de/api/interpreter?data=node('.$aResult['import-node'].');out%20meta;';
+        $sContentURL = 'https://overpass-api.de/api/interpreter?data=node('.$aResult['import-node'].');out%20meta;';
     }
 }
 
 if (isset($aResult['import-way']) && $aResult['import-way']) {
     if ($bUseOSMApi) {
-        $sContentURL = 'http://www.openstreetmap.org/api/0.6/way/'.$aResult['import-way'].'/full';
+        $sContentURL = 'https://www.openstreetmap.org/api/0.6/way/'.$aResult['import-way'].'/full';
     } else {
-        $sContentURL = 'http://overpass-api.de/api/interpreter?data=(way('.$aResult['import-way'].');node(w););out%20meta;';
+        $sContentURL = 'https://overpass-api.de/api/interpreter?data=(way('.$aResult['import-way'].');node(w););out%20meta;';
     }
 }
 
 if (isset($aResult['import-relation']) && $aResult['import-relation']) {
     if ($bUseOSMApi) {
-        $sContentURLsModifyXMLstr = 'http://www.openstreetmap.org/api/0.6/relation/'.$aResult['import-relation'].'/full';
+        $sContentURLsModifyXMLstr = 'https://www.openstreetmap.org/api/0.6/relation/'.$aResult['import-relation'].'/full';
     } else {
-        $sContentURL = 'http://overpass-api.de/api/interpreter?data=((rel('.$aResult['import-relation'].');way(r);node(w));node(r));out%20meta;';
+        $sContentURL = 'https://overpass-api.de/api/interpreter?data=((rel('.$aResult['import-relation'].');way(r);node(w));node(r));out%20meta;';
     }
 }
 
@@ -160,27 +218,28 @@ if ($bHaveDiff) {
     // import generated change file
     $sCMD = $sOsm2pgsqlCmd.' '.$sTemporaryFile;
     echo $sCMD."\n";
-    exec($sCMD, $sJunk, $iErrorLevel);
+    $iErrorLevel = runWithEnv($sCMD, $aProcEnv);
     if ($iErrorLevel) {
         fail("osm2pgsql exited with error level $iErrorLevel\n");
     }
 }
 
 if ($aResult['deduplicate']) {
-    $oDB =& getDB();
+    $oDB = new Nominatim\DB();
+    $oDB->connect();
 
-    if (getPostgresVersion($oDB) < 9.3) {
+    if ($oDB->getPostgresVersion() < 9.3) {
         fail('ERROR: deduplicate is only currently supported in postgresql 9.3');
     }
 
     $sSQL = 'select partition from country_name order by country_code';
-    $aPartitions = chksql($oDB->getCol($sSQL));
+    $aPartitions = $oDB->getCol($sSQL);
     $aPartitions[] = 0;
 
     // we don't care about empty search_name_* partitions, they can't contain mentions of duplicates
     foreach ($aPartitions as $i => $sPartition) {
         $sSQL = 'select count(*) from search_name_'.$sPartition;
-        $nEntries = chksql($oDB->getOne($sSQL));
+        $nEntries = $oDB->getOne($sSQL);
         if ($nEntries == 0) {
             unset($aPartitions[$i]);
         }
@@ -189,7 +248,7 @@ if ($aResult['deduplicate']) {
     $sSQL = "select word_token,count(*) from word where substr(word_token, 1, 1) = ' '";
     $sSQL .= ' and class is null and type is null and country_code is null';
     $sSQL .= ' group by word_token having count(*) > 1 order by word_token';
-    $aDuplicateTokens = chksql($oDB->getAll($sSQL));
+    $aDuplicateTokens = $oDB->getAll($sSQL);
     foreach ($aDuplicateTokens as $aToken) {
         if (trim($aToken['word_token']) == '' || trim($aToken['word_token']) == '-') continue;
         echo 'Deduping '.$aToken['word_token']."\n";
@@ -197,7 +256,7 @@ if ($aResult['deduplicate']) {
         $sSQL .= ' (select count(*) from search_name where nameaddress_vector @> ARRAY[word_id]) as num';
         $sSQL .= " from word where word_token = '".$aToken['word_token'];
         $sSQL .= "' and class is null and type is null and country_code is null order by num desc";
-        $aTokenSet = chksql($oDB->getAll($sSQL));
+        $aTokenSet = $oDB->getAll($sSQL);
 
         $aKeep = array_shift($aTokenSet);
         $iKeepID = $aKeep['word_id'];
@@ -207,54 +266,100 @@ if ($aResult['deduplicate']) {
             $sSQL .= ' name_vector = array_replace(name_vector,'.$aRemove['word_id'].','.$iKeepID.'),';
             $sSQL .= ' nameaddress_vector = array_replace(nameaddress_vector,'.$aRemove['word_id'].','.$iKeepID.')';
             $sSQL .= ' where name_vector @> ARRAY['.$aRemove['word_id'].']';
-            chksql($oDB->query($sSQL));
+            $oDB->exec($sSQL);
 
             $sSQL = 'update search_name set';
             $sSQL .= ' nameaddress_vector = array_replace(nameaddress_vector,'.$aRemove['word_id'].','.$iKeepID.')';
             $sSQL .= ' where nameaddress_vector @> ARRAY['.$aRemove['word_id'].']';
-            chksql($oDB->query($sSQL));
+            $oDB->exec($sSQL);
 
             $sSQL = 'update location_area_country set';
             $sSQL .= ' keywords = array_replace(keywords,'.$aRemove['word_id'].','.$iKeepID.')';
             $sSQL .= ' where keywords @> ARRAY['.$aRemove['word_id'].']';
-            chksql($oDB->query($sSQL));
+            $oDB->exec($sSQL);
 
             foreach ($aPartitions as $sPartition) {
                 $sSQL = 'update search_name_'.$sPartition.' set';
                 $sSQL .= ' name_vector = array_replace(name_vector,'.$aRemove['word_id'].','.$iKeepID.')';
                 $sSQL .= ' where name_vector @> ARRAY['.$aRemove['word_id'].']';
-                chksql($oDB->query($sSQL));
+                $oDB->exec($sSQL);
 
                 $sSQL = 'update location_area_country set';
                 $sSQL .= ' keywords = array_replace(keywords,'.$aRemove['word_id'].','.$iKeepID.')';
                 $sSQL .= ' where keywords @> ARRAY['.$aRemove['word_id'].']';
-                chksql($oDB->query($sSQL));
+                $oDB->exec($sSQL);
             }
 
             $sSQL = 'delete from word where word_id = '.$aRemove['word_id'];
-            chksql($oDB->query($sSQL));
+            $oDB->exec($sSQL);
         }
     }
 }
 
+if ($aResult['recompute-word-counts']) {
+    info('Recompute frequency of full-word search terms');
+    $sTemplate = file_get_contents(CONST_BasePath.'/sql/words_from_search_name.sql');
+    runSQLScript($sTemplate, true, true);
+}
+
 if ($aResult['index']) {
-    passthru(CONST_InstallPath.'/nominatim/nominatim -i -d '.$aDSNInfo['database'].' -P '.$aDSNInfo['port'].' -t '.$aResult['index-instances'].' -r '.$aResult['index-rank']);
+    $sCmd = CONST_InstallPath.'/nominatim/nominatim -i -d '.$aDSNInfo['database'].' -P '.$aDSNInfo['port'].' -t '.$aResult['index-instances'].' -r '.$aResult['index-rank'];
+    if (isset($aDSNInfo['hostspec']) && $aDSNInfo['hostspec']) {
+        $sCmd .= ' -H ' . $aDSNInfo['hostspec'];
+    }
+    if (isset($aDSNInfo['username']) && $aDSNInfo['username']) {
+        $sCmd .= ' -U ' . $aDSNInfo['username'];
+    }
+
+    runWithEnv($sCmd, $aProcEnv);
+
+    $oDB->exec('update import_status set indexed = true');
+}
+
+if ($aResult['update-address-levels']) {
+    echo 'Updating address levels from '.CONST_Address_Level_Config.".\n";
+    $oAlParser = new \Nominatim\Setup\AddressLevelParser(CONST_Address_Level_Config);
+    $oAlParser->createTable($oDB, 'address_levels');
+}
+
+if ($aResult['recompute-importance']) {
+    echo "Updating importance values for database.\n";
+    $oDB = new Nominatim\DB();
+    $oDB->connect();
+
+    $sSQL = 'ALTER TABLE placex DISABLE TRIGGER ALL;';
+    $sSQL .= 'UPDATE placex SET (wikipedia, importance) =';
+    $sSQL .= '   (SELECT wikipedia, importance';
+    $sSQL .= '    FROM compute_importance(extratags, country_code, osm_type, osm_id));';
+    $sSQL .= 'UPDATE placex s SET wikipedia = d.wikipedia, importance = d.importance';
+    $sSQL .= ' FROM placex d';
+    $sSQL .= ' WHERE s.place_id = d.linked_place_id and d.wikipedia is not null';
+    $sSQL .= '       and (s.wikipedia is null or s.importance < d.importance);';
+    $sSQL .= 'ALTER TABLE placex ENABLE TRIGGER ALL;';
+    $oDB->exec($sSQL);
 }
 
 if ($aResult['import-osmosis'] || $aResult['import-osmosis-all']) {
     //
     if (strpos(CONST_Replication_Url, 'download.geofabrik.de') !== false && CONST_Replication_Update_Interval < 86400) {
-        fail("Error: Update interval too low for download.geofabrik.de.  Please check install documentation (http://wiki.openstreetmap.org/wiki/Nominatim/Installation#Updates)\n");
+        fail('Error: Update interval too low for download.geofabrik.de. ' .
+             "Please check install documentation (http://nominatim.org/release-docs/latest/Import-and-Update#setting-up-the-update-process)\n");
     }
 
     $sImportFile = CONST_InstallPath.'/osmosischange.osc';
     $sCMDDownload = CONST_Pyosmium_Binary.' --server '.CONST_Replication_Url.' -o '.$sImportFile.' -s '.CONST_Replication_Max_Diff_size;
     $sCMDImport = $sOsm2pgsqlCmd.' '.$sImportFile;
     $sCMDIndex = CONST_InstallPath.'/nominatim/nominatim -i -d '.$aDSNInfo['database'].' -P '.$aDSNInfo['port'].' -t '.$aResult['index-instances'];
+    if (isset($aDSNInfo['hostspec']) && $aDSNInfo['hostspec']) {
+        $sCMDIndex .= ' -H ' . $aDSNInfo['hostspec'];
+    }
+    if (isset($aDSNInfo['username']) && $aDSNInfo['username']) {
+        $sCMDIndex .= ' -U ' . $aDSNInfo['username'];
+    }
 
     while (true) {
         $fStartTime = time();
-        $aLastState = chksql($oDB->getRow('SELECT *, EXTRACT (EPOCH FROM lastimportdate) as unix_ts FROM import_status'));
+        $aLastState = $oDB->getRow('SELECT *, EXTRACT (EPOCH FROM lastimportdate) as unix_ts FROM import_status');
 
         if (!$aLastState['sequence_id']) {
             echo "Updates not set up. Please run ./utils/update.php --init-updates.\n";
@@ -266,7 +371,7 @@ if ($aResult['import-osmosis'] || $aResult['import-osmosis-all']) {
         $sBatchEnd = $aLastState['lastimportdate'];
         $iEndSequence = $aLastState['sequence_id'];
 
-        if ($aLastState['indexed'] == 't') {
+        if ($aLastState['indexed']) {
             // Sleep if the update interval has not yet been reached.
             $fNextUpdate = $aLastState['unix_ts'] + CONST_Replication_Update_Interval;
             if ($fNextUpdate > $fStartTime) {
@@ -281,7 +386,9 @@ if ($aResult['import-osmosis'] || $aResult['import-osmosis-all']) {
                 $iNextSeq = (int) $aLastState['sequence_id'];
                 unset($aOutput);
                 echo "$sCMDDownload -I $iNextSeq\n";
-                unlink($sImportFile);
+                if (file_exists($sImportFile)) {
+                    unlink($sImportFile);
+                }
                 exec($sCMDDownload.' -I '.$iNextSeq, $aOutput, $iResult);
 
                 if ($iResult == 3) {
@@ -316,7 +423,7 @@ if ($aResult['import-osmosis'] || $aResult['import-osmosis-all']) {
             $fCMDStartTime = time();
             echo $sCMDImport."\n";
             unset($sJunk);
-            exec($sCMDImport, $sJunk, $iErrorLevel);
+            $iErrorLevel = runWithEnv($sCMDImport, $aProcEnv);
             if ($iErrorLevel) {
                 echo "Error executing osm2pgsql: $iErrorLevel\n";
                 exit($iErrorLevel);
@@ -330,12 +437,12 @@ if ($aResult['import-osmosis'] || $aResult['import-osmosis-all']) {
             $sSQL .= date('Y-m-d H:i:s', $fCMDStartTime)."','";
             $sSQL .= date('Y-m-d H:i:s')."','import')";
             var_Dump($sSQL);
-            chksql($oDB->query($sSQL));
+            $oDB->exec($sSQL);
 
             // update the status
             $sSQL = "UPDATE import_status SET lastimportdate = '$sBatchEnd', indexed=false, sequence_id = $iEndSequence";
             var_Dump($sSQL);
-            chksql($oDB->query($sSQL));
+            $oDB->exec($sSQL);
             echo date('Y-m-d H:i:s')." Completed download step for $sBatchEnd in ".round((time()-$fCMDStartTime)/60, 2)." minutes\n";
         }
 
@@ -345,7 +452,7 @@ if ($aResult['import-osmosis'] || $aResult['import-osmosis-all']) {
             $fCMDStartTime = time();
 
             echo "$sThisIndexCmd\n";
-            exec($sThisIndexCmd, $sJunk, $iErrorLevel);
+            $iErrorLevel = runWithEnv($sThisIndexCmd, $aProcEnv);
             if ($iErrorLevel) {
                 echo "Error: $iErrorLevel\n";
                 exit($iErrorLevel);
@@ -353,15 +460,20 @@ if ($aResult['import-osmosis'] || $aResult['import-osmosis-all']) {
 
             $sSQL = 'INSERT INTO import_osmosis_log';
             $sSQL .= '(batchend, batchseq, batchsize, starttime, endtime, event)';
-            $sSQL .= " values ('$sBatchEnd',$iEndSequence,$iFileSize,'";
+            $sSQL .= " values ('$sBatchEnd',$iEndSequence,NULL,'";
             $sSQL .= date('Y-m-d H:i:s', $fCMDStartTime)."','";
             $sSQL .= date('Y-m-d H:i:s')."','index')";
             var_Dump($sSQL);
-            $oDB->query($sSQL);
+            $oDB->exec($sSQL);
             echo date('Y-m-d H:i:s')." Completed index step for $sBatchEnd in ".round((time()-$fCMDStartTime)/60, 2)." minutes\n";
 
             $sSQL = 'update import_status set indexed = true';
-            $oDB->query($sSQL);
+            $oDB->exec($sSQL);
+        } else {
+            if ($aResult['import-osmosis-all']) {
+                echo "Error: --no-index cannot be used with continuous imports (--import-osmosis-all).\n";
+                exit(1);
+            }
         }
 
         $fDuration = time() - $fStartTime;