ignore frequent partial search terms

[nominatim.git] / utils / setup.php

diff --git a/utils/setup.php b/utils/setup.php
index db15aa446b640c32f9e481c7d4396fd59aaf80b6..bfc7c7205e47866b00511e763c5f5490ddeab6a6 100755 (executable)
--- a/utils/setup.php
+++ b/utils/setup.php
@@ -20,16 +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('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('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('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('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'),
                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'),


@@ -38,7 +43,27 @@
 
        $bDidSomething = false;
 
 
        $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)
        {
        $iInstances = isset($aCMDResult['threads'])?$aCMDResult['threads']:(getProcessorCount()-1);
        if ($iInstances < 1)
        {


@@ -84,9 +109,9 @@
                passthru('createdb -E UTF-8 '.$aDSNInfo['database']);
        }
 
                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.
 
                $bDidSomething = true;
                // TODO: path detection, detection memory, etc.
 


@@ -116,11 +141,15 @@
                $bDidSomething = true;
 
                $osm2pgsql = CONST_Osm2pgsql_Binary;
                $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'];
                $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');
 
                $oDB =& getDB();
                $x = $oDB->getRow('select * from place limit 1');


@@ -136,7 +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');
                $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);
        }
 
                pgsqlRunScript($sTemplate);
        }
 


@@ -210,9 +241,37 @@
                pgsqlRunScript($sTemplate);
        }
 
                pgsqlRunScript($sTemplate);
        }
 

+       if ($aCMDResult['import-wikipedia-articles'] || $aCMDResult['all'])
+       {
+               $bDidSomething = true;
+               $sWikiArticlesFile = CONST_BasePath.'/data/wikipedia_article.sql.bin';
+               $sWikiRedirectsFile = CONST_BasePath.'/data/wikipedia_redirect.sql.bin';
+               if (file_exists($sWikiArticlesFile))
+               {
+                       echo "Importing wikipedia articles...";
+                       pgsqlRunDropAndRestore($sWikiArticlesFile);
+                       echo "...done\n";
+               }
+               else
+               {
+                       echo "WARNING: wikipedia article dump file not found - places will have default importance\n";
+               }
+               if (file_exists($sWikiRedirectsFile))
+               {
+                       echo "Importing wikipedia redirects...";
+                       pgsqlRunDropAndRestore($sWikiRedirectsFile);
+                       echo "...done\n";
+               }
+               else
+               {
+                       echo "WARNING: wikipedia redirect dump file not found - some place importance values may be missing\n";
+               }
+       }
+
+

        if ($aCMDResult['load-data'] || $aCMDResult['all'])
        {
        if ($aCMDResult['load-data'] || $aCMDResult['all'])
        {

-               echo "Load Data\n";
+               echo "Drop old Data\n";

                $bDidSomething = true;
 
                $oDB =& getDB();
                $bDidSomething = true;
 
                $oDB =& getDB();


@@ -235,6 +294,31 @@
                if (!pg_query($oDB->connection, 'CREATE SEQUENCE seq_place start 100000')) fail(pg_last_error($oDB->connection));
                echo '.';
 
                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($oDB->connection));
+               echo ".\n";
+
+               // pre-create the word list
+               if (!$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++)
                {
                $aDBInstances = array();
                for($i = 0; $i < $iInstances; $i++)
                {


@@ -294,6 +378,8 @@
        {
                $bDidSomething = true;
 
        {
                $bDidSomething = true;
 

+               pgsqlRunScriptFile(CONST_BasePath.'/sql/tiger_import_start.sql');
+

                $aDBInstances = array();
                for($i = 0; $i < $iInstances; $i++)
                {
                $aDBInstances = array();
                for($i = 0; $i < $iInstances; $i++)
                {


@@ -329,7 +415,7 @@
                        }
 
                        fclose($hFile);
                        }
 
                        fclose($hFile);

-       
+

                        $bAnyBusy = true;
                        while($bAnyBusy)
                        {
                        $bAnyBusy = true;
                        while($bAnyBusy)
                        {


@@ -342,6 +428,9 @@
                        }
                        echo "\n";
                }
                        }
                        echo "\n";
                }

+
+               echo "Creating indexes\n";
+               pgsqlRunScriptFile(CONST_BasePath.'/sql/tiger_import_finish.sql');

        }
 
        if ($aCMDResult['calculate-postcodes'] || $aCMDResult['all'])
        }
 
        if ($aCMDResult['calculate-postcodes'] || $aCMDResult['all'])


@@ -369,7 +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";
 
                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);
 
                $sDate = $aCMDResult['osmosis-init-date'];
                $aDate = date_parse_from_format("Y-m-d\TH-i", $sDate);


@@ -394,11 +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;
                $sOutputFile = '';
                if (isset($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'])
        }
 
        if ($aCMDResult['create-search-indices'] || $aCMDResult['all'])


@@ -433,7 +527,12 @@
        {
                $bDidSomething = true;
                $sTargetDir = $aCMDResult['create-website'];
        {
                $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');
                @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');


@@ -452,9 +551,10 @@
        {
                if (!file_exists($sFilename)) fail('unable to find '.$sFilename);
 
        {
                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);
                $aDSNInfo = DB::parseDSN(CONST_Database_DSN);

-               $sCMD = 'psql -f '.$sFilename.' '.$aDSNInfo['database'];
+               if (!isset($aDSNInfo['port']) || !$aDSNInfo['port']) $aDSNInfo['port'] = 5432;
+               $sCMD = 'psql -p '.$aDSNInfo['port'].' -d '.$aDSNInfo['database'].' -f '.$sFilename;

 
                $aDescriptors = array(
                        0 => array('pipe', 'r'),
 
                $aDescriptors = array(
                        0 => array('pipe', 'r'),


@@ -479,10 +579,10 @@
 
        function pgsqlRunScript($sScript)
        {
 
        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;
                $aDSNInfo = DB::parseDSN(CONST_Database_DSN);
                if (!isset($aDSNInfo['port']) || !$aDSNInfo['port']) $aDSNInfo['port'] = 5432;

-               $sCMD = 'psql -p '.$aDSNInfo['port'].' '.$aDSNInfo['database'];
+               $sCMD = 'psql -p '.$aDSNInfo['port'].' -d '.$aDSNInfo['database'];

                $aDescriptors = array(
                        0 => array('pipe', 'r'),
                        1 => STDOUT, 
                $aDescriptors = array(
                        0 => array('pipe', 'r'),
                        1 => STDOUT, 


@@ -500,3 +600,66 @@
                fclose($ahPipes[0]);
                proc_close($hProcess);
        }
                fclose($ahPipes[0]);
                proc_close($hProcess);
        }

+
+       function pgsqlRunRestoreData($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 '.$aDSNInfo['database'].' -Fc -a '.$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($hProcess);
+       }
+
+       function pgsqlRunDropAndRestore($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);
+       }