index admin boundaries before everything else

Avoids irregularities that might happen because the address
rank of a boundary is changed through linking.

diff --git a/lib/setup/SetupClass.php b/lib/setup/SetupClass.php
index ba700a2cde9e305f11880e9a28e7795cac24e687..2815f8c4b11a120bad4746776a7b6006ffc392db 100755 (executable)
--- a/lib/setup/SetupClass.php
+++ b/lib/setup/SetupClass.php
@@ -566,19 +566,27 @@ class SetupFunctions
         info('Index ranks 0 - 4');
         $oCmd = (clone $oBaseCmd)->addParams('--maxrank', 4);
         echo $oCmd->escapedCmd();
-        
+
         $iStatus = $oCmd->run();
         if ($iStatus != 0) {
             fail('error status ' . $iStatus . ' running nominatim!');
         }
         if (!$bIndexNoanalyse) $this->pgsqlRunScript('ANALYSE');
 
+        info('Index administrative boundaries');
+        $oCmd = (clone $oBaseCmd)->addParams('-b');
+        $iStatus = $oCmd->run();
+        if ($iStatus != 0) {
+            fail('error status ' . $iStatus . ' running nominatim!');
+        }
+
         info('Index ranks 5 - 25');
         $oCmd = (clone $oBaseCmd)->addParams('--minrank', 5, '--maxrank', 25);
         $iStatus = $oCmd->run();
         if ($iStatus != 0) {
             fail('error status ' . $iStatus . ' running nominatim!');
         }
+
         if (!$bIndexNoanalyse) $this->pgsqlRunScript('ANALYSE');
 
         info('Index ranks 26 - 30');

diff --git a/nominatim/nominatim.py b/nominatim/nominatim.py
index 67cd42ee9bde66bbfe199245be6abc5989b3ef99..a4f4a62e7cdf262353a9917e2909db50ece46cad 100755 (executable)
--- a/nominatim/nominatim.py
+++ b/nominatim/nominatim.py
@@ -82,6 +82,29 @@ class InterpolationRunner(object):
                   SET indexed_status = 0 WHERE place_id IN ({})"""\
                .format(','.join((str(i) for i in ids)))
 
+class BoundaryRunner(object):
+    """ Returns SQL commands for indexing the administrative boundaries
+        by partition.
+    """
+
+    def name(self):
+        return "boundaries"
+
+    def sql_count_objects(self):
+        return """SELECT count(*) FROM placex
+                  WHERE indexed_status > 0
+                    AND rank_search < 26
+                    AND class = 'boundary' and type = 'administrative'"""
+
+    def sql_get_objects(self):
+        return """SELECT place_id FROM placex
+                  WHERE indexed_status > 0 and rank_search < 26
+                        and class = 'boundary' and type = 'administrative'
+                  ORDER BY partition, admin_level"""
+
+    def sql_index_place(self, ids):
+        return "UPDATE placex SET indexed_status = 0 WHERE place_id IN ({})"\
+               .format(','.join((str(i) for i in ids)))
 
 class Indexer(object):
     """ Main indexing routine.
@@ -93,8 +116,14 @@ class Indexer(object):
         self.conn = make_connection(options)
         self.threads = [DBConnection(options) for i in range(options.threads)]
 
-    def run(self):
-        """ Run indexing over the entire database.
+    def index_boundaries(self):
+        log.warning("Starting indexing boundaries using {} threads".format(
+                      len(self.threads)))
+
+        self.index(BoundaryRunner())
+
+    def index_by_rank(self):
+        """ Run classic indexing by rank.
         """
         log.warning("Starting indexing rank ({} to {}) using {} threads".format(
                  self.minrank, self.maxrank, len(self.threads)))
@@ -198,6 +227,9 @@ def nominatim_arg_parser():
     p.add_argument('-P', '--port',
                    dest='port', action='store',
                    help='PostgreSQL server port')
+    p.add_argument('-b', '--boundary-only',
+                   dest='boundary_only', action='store_true',
+                   help='Only index administrative boundaries (ignores min/maxrank).')
     p.add_argument('-r', '--minrank',
                    dest='minrank', type=int, metavar='RANK', default=0,
                    help='Minimum/starting rank.')
@@ -225,4 +257,7 @@ if __name__ == '__main__':
         password = getpass.getpass("Database password: ")
         options.password = password
 
-    Indexer(options).run()
+    if options.boundary_only:
+        Indexer(options).index_boundaries()
+    else:
+        Indexer(options).index_by_rank()

diff --git a/utils/update.php b/utils/update.php
index 9e74b4babd8bdb6bf2ef36c5e7671cf42ce830c8..f0b45b428cf4b866c624ea5717272337bc3e1b7a 100644 (file)
--- a/utils/update.php
+++ b/utils/update.php
@@ -278,9 +278,11 @@ if ($aResult['recompute-word-counts']) {
 
 if ($aResult['index']) {
     $oCmd = (clone $oIndexCmd)
-            ->addParams('--minrank', $aResult['index-rank']);
+            ->addParams('--minrank', $aResult['index-rank'], '-b');
+    $oCmd->run();
 
-    // echo $oCmd->escapedCmd()."\n";
+    $oCmd = (clone $oIndexCmd)
+            ->addParams('--minrank', $aResult['index-rank']);
     $oCmd->run();
 
     $oDB->exec('update import_status set indexed = true');
@@ -421,9 +423,18 @@ if ($aResult['import-osmosis'] || $aResult['import-osmosis-all']) {
 
         // Index file
         if (!$aResult['no-index']) {
-            $oThisIndexCmd = clone($oIndexCmd);
             $fCMDStartTime = time();
 
+            $oThisIndexCmd = clone($oIndexCmd);
+            $oThisIndexCmd->addParams('-b');
+            echo $oThisIndexCmd->escapedCmd()."\n";
+            $iErrorLevel = $oThisIndexCmd->run();
+            if ($iErrorLevel) {
+                echo "Error: $iErrorLevel\n";
+                exit($iErrorLevel);
+            }
+
+            $oThisIndexCmd = clone($oIndexCmd);
             echo $oThisIndexCmd->escapedCmd()."\n";
             $iErrorLevel = $oThisIndexCmd->run();
             if ($iErrorLevel) {