Merge remote-tracking branch 'upstream/master'

[nominatim.git] / nominatim / tools / database_import.py

diff --git a/nominatim/tools/database_import.py b/nominatim/tools/database_import.py
index 069157000568fea2ecc59452f86c605ab9459041..25efedb9a9a6a5ffb2c165df73e2c8a51f4431b4 100644 (file)
--- a/nominatim/tools/database_import.py
+++ b/nominatim/tools/database_import.py
@@ -11,13 +11,13 @@ from pathlib import Path
 import psutil
 import psycopg2
 
-from ..db.connection import connect, get_pg_env
-from ..db import utils as db_utils
-from ..db.async_connection import DBConnection
-from ..db.sql_preprocessor import SQLPreprocessor
-from .exec_utils import run_osm2pgsql
-from ..errors import UsageError
-from ..version import POSTGRESQL_REQUIRED_VERSION, POSTGIS_REQUIRED_VERSION
+from nominatim.db.connection import connect, get_pg_env
+from nominatim.db import utils as db_utils
+from nominatim.db.async_connection import DBConnection
+from nominatim.db.sql_preprocessor import SQLPreprocessor
+from nominatim.tools.exec_utils import run_osm2pgsql
+from nominatim.errors import UsageError
+from nominatim.version import POSTGRESQL_REQUIRED_VERSION, POSTGIS_REQUIRED_VERSION
 
 LOG = logging.getLogger()
 
@@ -179,29 +179,29 @@ def import_osm_data(osm_file, options, drop=False, ignore_errors=False):
             Path(options['flatnode_file']).unlink()
 
 
-def create_tables(conn, config, sqllib_dir, reverse_only=False):
+def create_tables(conn, config, reverse_only=False):
     """ Create the set of basic tables.
         When `reverse_only` is True, then the main table for searching will
         be skipped and only reverse search is possible.
     """
-    sql = SQLPreprocessor(conn, config, sqllib_dir)
+    sql = SQLPreprocessor(conn, config)
     sql.env.globals['db']['reverse_only'] = reverse_only
 
     sql.run_sql_file(conn, 'tables.sql')
 
 
-def create_table_triggers(conn, config, sqllib_dir):
+def create_table_triggers(conn, config):
     """ Create the triggers for the tables. The trigger functions must already
         have been imported with refresh.create_functions().
     """
-    sql = SQLPreprocessor(conn, config, sqllib_dir)
+    sql = SQLPreprocessor(conn, config)
     sql.run_sql_file(conn, 'table-triggers.sql')
 
 
-def create_partition_tables(conn, config, sqllib_dir):
-    """ Create tables that have explicit partioning.
+def create_partition_tables(conn, config):
+    """ Create tables that have explicit partitioning.
     """
-    sql = SQLPreprocessor(conn, config, sqllib_dir)
+    sql = SQLPreprocessor(conn, config)
     sql.run_sql_file(conn, 'partition-tables.src.sql')
 
 
@@ -285,3 +285,56 @@ def load_data(dsn, data_dir, threads):
     with connect(dsn) as conn:
         with conn.cursor() as cur:
             cur.execute('ANALYSE')
+
+
+def create_search_indices(conn, config, drop=False):
+    """ Create tables that have explicit partitioning.
+    """
+
+    # If index creation failed and left an index invalid, they need to be
+    # cleaned out first, so that the script recreates them.
+    with conn.cursor() as cur:
+        cur.execute("""SELECT relname FROM pg_class, pg_index
+                       WHERE pg_index.indisvalid = false
+                             AND pg_index.indexrelid = pg_class.oid""")
+        bad_indices = [row[0] for row in list(cur)]
+        for idx in bad_indices:
+            LOG.info("Drop invalid index %s.", idx)
+            cur.execute('DROP INDEX "{}"'.format(idx))
+    conn.commit()
+
+    sql = SQLPreprocessor(conn, config)
+
+    sql.run_sql_file(conn, 'indices.sql', drop=drop)
+
+def create_country_names(conn, config):
+    """ Create search index for default country names.
+    """
+
+    with conn.cursor() as cur:
+        cur.execute("""SELECT getorcreate_country(make_standard_name('uk'), 'gb')""")
+        cur.execute("""SELECT getorcreate_country(make_standard_name('united states'), 'us')""")
+        cur.execute("""SELECT COUNT(*) FROM
+                       (SELECT getorcreate_country(make_standard_name(country_code),
+                       country_code) FROM country_name WHERE country_code is not null) AS x""")
+        cur.execute("""SELECT COUNT(*) FROM
+                       (SELECT getorcreate_country(make_standard_name(name->'name'), country_code) 
+                       FROM country_name WHERE name ? 'name') AS x""")
+        sql_statement = """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"""
+
+        languages = config.LANGUAGES
+
+        if languages:
+            sql_statement = "{} IN (".format(sql_statement)
+            delim = ''
+            for language in languages.split(','):
+                sql_statement = "{}{}'name:{}'".format(sql_statement, delim, language)
+                delim = ', '
+            sql_statement = '{})'.format(sql_statement)
+        else:
+            sql_statement = "{} LIKE 'name:%'".format(sql_statement)
+        sql_statement = "{}) v".format(sql_statement)
+        cur.execute(sql_statement)
+    conn.commit()