X-Git-Url: https://git.openstreetmap.org/nominatim.git/blobdiff_plain/8e0ffde3e0a403d289b91ad720444fd94f3ae210..e286536959144de9ab18abf7dce7a58e630b3bbc:/utils/update.php

diff --git a/utils/update.php b/utils/update.php
index d729519a..46f70062 100755
--- a/utils/update.php
+++ b/utils/update.php
@@ -1,18 +1,22 @@
-#!/usr/bin/php -Cq
+#!@PHP_BIN@ -Cq
 <?php
 
 require_once(dirname(dirname(__FILE__)).'/settings/settings.php');
 require_once(CONST_BasePath.'/lib/init-cmd.php');
 ini_set('memory_limit', '800M');
 
+# (long-opt, short-opt, min-occurs, max-occurs, num-arguments, num-arguments, type, help)
+
 $aCMDOptions
 = array(
-   "Import / update / index osm data",
+   'Import / update / index osm data',
    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('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'),
@@ -33,6 +37,7 @@ $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('no-npi', '', 0, 1, 0, 0, 'bool', '(obsolete)'),
   );
 getCmdOpt($_SERVER['argv'], $aCMDOptions, $aResult, true, true);
@@ -54,7 +59,18 @@ if ($iCacheMemory + 500 > getTotalMemoryMB()) {
     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)) {
+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;
 }
 
@@ -64,18 +80,36 @@ if ($aResult['init-updates']) {
     if ($sBaseState === false) {
         echo "\nCannot find state.txt file at the configured replication URL.\n";
         echo "Does the URL point to a directory containing OSM update data?\n\n";
-        fail("replication URL not reachable.");
+        fail('replication URL not reachable.');
+    }
+    // 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');
     }
-    $sSetup = CONST_InstallPath.'/utils/setup.php';
-    $iRet = -1;
-    passthru($sSetup.' --create-functions --enable-diff-updates', $iRet);
+    $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']) {
+        $sSetup = CONST_InstallPath.'/utils/setup.php';
+        $iRet = -1;
+        passthru($sSetup.' --create-functions --enable-diff-updates', $iRet);
+        if ($iRet != 0) {
+            fail('Error running setup script');
+        }
     }
 
     $sDatabaseDate = getDatabaseDate($oDB);
     if ($sDatabaseDate === false) {
-        fail("Cannot determine date of database.");
+        fail('Cannot determine date of database.');
     }
     $sWindBack = strftime('%Y-%m-%dT%H:%M:%SZ', strtotime($sDatabaseDate) - (3*60*60));
 
@@ -89,14 +123,25 @@ if ($aResult['init-updates']) {
 
     pg_query($oDB->connection, 'TRUNCATE import_status');
     $sSQL = "INSERT INTO import_status (lastimportdate, sequence_id, indexed) VALUES('";
-    $sSQL .= $sDatabaseDate."',".$aOutput[0].", true)";
+    $sSQL .= $sDatabaseDate."',".$aOutput[0].', true)';
     if (!pg_query($oDB->connection, $sSQL)) {
-        fail("Could not enter sequence into database.");
+        fail('Could not enter sequence into database.');
     }
 
     echo "Done. Database updates will start at sequence $aOutput[0] ($sWindBack)\n";
 }
 
+if ($aResult['check-for-updates']) {
+    $aLastState = chksql($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'] : $aResult['import-file'];
@@ -108,7 +153,7 @@ if (isset($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, $iErrorLevel\n");
@@ -118,7 +163,7 @@ if (isset($aResult['import-diff']) || isset($aResult['import-file'])) {
 }
 
 if ($aResult['calculate-postcodes']) {
-    info("Update postcodes centroids");
+    info('Update postcodes centroids');
     $sTemplate = file_get_contents(CONST_BasePath.'/sql/update-postcodes.sql');
     runSQLScript($sTemplate, true, true);
 }
@@ -129,25 +174,25 @@ $bUseOSMApi = isset($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;';
     }
 }
 
@@ -160,7 +205,7 @@ 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 $iErrorLevel\n");
     }
@@ -170,7 +215,7 @@ if ($aResult['deduplicate']) {
     $oDB =& getDB();
 
     if (getPostgresVersion($oDB) < 9.3) {
-        fail("ERROR: deduplicate is only currently supported in postgresql 9.3");
+        fail('ERROR: deduplicate is only currently supported in postgresql 9.3');
     }
 
     $sSQL = 'select partition from country_name order by country_code';
@@ -179,7 +224,7 @@ if ($aResult['deduplicate']) {
 
     // 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;
+        $sSQL = 'select count(*) from search_name_'.$sPartition;
         $nEntries = chksql($oDB->getOne($sSQL));
         if ($nEntries == 0) {
             unset($aPartitions[$i]);
@@ -187,14 +232,14 @@ 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";
+    $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($sSQL));
     foreach ($aDuplicateTokens as $aToken) {
         if (trim($aToken['word_token']) == '' || trim($aToken['word_token']) == '-') continue;
-        echo "Deduping ".$aToken['word_token']."\n";
-        $sSQL = "select word_id,";
-        $sSQL .= " (select count(*) from search_name where nameaddress_vector @> ARRAY[word_id]) as num";
+        echo 'Deduping '.$aToken['word_token']."\n";
+        $sSQL = 'select word_id,';
+        $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($sSQL));
@@ -203,54 +248,75 @@ if ($aResult['deduplicate']) {
         $iKeepID = $aKeep['word_id'];
 
         foreach ($aTokenSet as $aRemove) {
-            $sSQL = "update search_name set";
-            $sSQL .= " name_vector = array_replace(name_vector,".$aRemove['word_id'].",".$iKeepID."),";
-            $sSQL .= " nameaddress_vector = array_replace(nameaddress_vector,".$aRemove['word_id'].",".$iKeepID.")";
-            $sSQL .= " where name_vector @> ARRAY[".$aRemove['word_id']."]";
+            $sSQL = 'update search_name set';
+            $sSQL .= ' name_vector = array_replace(name_vector,'.$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));
 
-            $sSQL = "update search_name set";
-            $sSQL .= " nameaddress_vector = array_replace(nameaddress_vector,".$aRemove['word_id'].",".$iKeepID.")";
-            $sSQL .= " where nameaddress_vector @> ARRAY[".$aRemove['word_id']."]";
+            $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));
 
-            $sSQL = "update location_area_country set";
-            $sSQL .= " keywords = array_replace(keywords,".$aRemove['word_id'].",".$iKeepID.")";
-            $sSQL .= " where keywords @> ARRAY[".$aRemove['word_id']."]";
+            $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));
 
             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']."]";
+                $sSQL = 'update search_name_'.$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));
 
-                $sSQL = "update location_area_country set";
-                $sSQL .= " keywords = array_replace(keywords,".$aRemove['word_id'].",".$iKeepID.")";
-                $sSQL .= " where keywords @> ARRAY[".$aRemove['word_id']."]";
+                $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));
             }
 
-            $sSQL = "delete from word where word_id = ".$aRemove['word_id'];
+            $sSQL = 'delete from word where word_id = '.$aRemove['word_id'];
             chksql($oDB->query($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);
 }
 
 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) {
         $fStartTime = time();
@@ -281,7 +347,9 @@ if ($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 ($iResult == 3) {
@@ -316,7 +384,7 @@ if ($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);
@@ -345,7 +413,7 @@ 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);
@@ -360,7 +428,7 @@ if ($aResult['import-osmosis'] || $aResult['import-osmosis-all']) {
             $oDB->query($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";
+            $sSQL = 'update import_status set indexed = true';
             $oDB->query($sSQL);
         }