X-Git-Url: https://git.openstreetmap.org./nominatim.git/blobdiff_plain/f6e894a53af83a69f553555cb4a6248d57a58391..b88b952f56d6ac1eec46202ba05a04fcd12575d9:/test/python/test_tools_database_import.py diff --git a/test/python/test_tools_database_import.py b/test/python/test_tools_database_import.py index a4ab16f9..1311ef5d 100644 --- a/test/python/test_tools_database_import.py +++ b/test/python/test_tools_database_import.py @@ -4,6 +4,7 @@ Tests for functions to import a new database. import pytest import psycopg2 import sys +from pathlib import Path from nominatim.tools import database_import from nominatim.errors import UsageError @@ -23,6 +24,24 @@ def nonexistant_db(): with conn.cursor() as cur: cur.execute('DROP DATABASE IF EXISTS {}'.format(dbname)) +@pytest.mark.parametrize("no_partitions", (True, False)) +def test_setup_skeleton(src_dir, nonexistant_db, no_partitions): + database_import.setup_database_skeleton('dbname=' + nonexistant_db, + src_dir / 'data', no_partitions) + + conn = psycopg2.connect(database=nonexistant_db) + + try: + with conn.cursor() as cur: + cur.execute("SELECT distinct partition FROM country_name") + partitions = set([r[0] for r in list(cur)]) + if no_partitions: + assert partitions == set([0]) + else: + assert len(partitions) > 10 + finally: + conn.close() + def test_create_db_success(nonexistant_db): database_import.create_db('dbname=' + nonexistant_db, rouser='www-data') @@ -78,6 +97,22 @@ def test_install_module(tmp_path): assert outfile.stat().st_mode == 33261 +def test_install_module_custom(tmp_path): + (tmp_path / 'nominatim.so').write_text('TEST nomiantim.so') + + database_import.install_module(tmp_path, tmp_path, str(tmp_path.resolve())) + + assert not (tmp_path / 'module').exists() + + +def test_install_module_fail_access(temp_db_conn, tmp_path): + (tmp_path / 'nominatim.so').write_text('TEST nomiantim.so') + + with pytest.raises(UsageError, match='.*module cannot be accessed.*'): + database_import.install_module(tmp_path, tmp_path, '', + conn=temp_db_conn) + + def test_import_base_data(src_dir, temp_db, temp_db_cursor): temp_db_cursor.execute('CREATE EXTENSION hstore') temp_db_cursor.execute('CREATE EXTENSION postgis') @@ -94,3 +129,103 @@ def test_import_base_data_ignore_partitions(src_dir, temp_db, temp_db_cursor): assert temp_db_cursor.scalar('SELECT count(*) FROM country_name') > 0 assert temp_db_cursor.scalar('SELECT count(*) FROM country_name WHERE partition != 0') == 0 + + +def test_import_osm_data_simple(temp_db_cursor,osm2pgsql_options): + temp_db_cursor.execute('CREATE TABLE place (id INT)') + temp_db_cursor.execute('INSERT INTO place values (1)') + + database_import.import_osm_data('file.pdf', osm2pgsql_options) + + +def test_import_osm_data_simple_no_data(temp_db_cursor,osm2pgsql_options): + temp_db_cursor.execute('CREATE TABLE place (id INT)') + + with pytest.raises(UsageError, match='No data.*'): + database_import.import_osm_data('file.pdf', osm2pgsql_options) + + +def test_import_osm_data_drop(temp_db_conn, temp_db_cursor, tmp_path, osm2pgsql_options): + temp_db_cursor.execute('CREATE TABLE place (id INT)') + temp_db_cursor.execute('CREATE TABLE planet_osm_nodes (id INT)') + temp_db_cursor.execute('INSERT INTO place values (1)') + + flatfile = tmp_path / 'flatfile' + flatfile.write_text('touch') + + osm2pgsql_options['flatnode_file'] = str(flatfile.resolve()) + + database_import.import_osm_data('file.pdf', osm2pgsql_options, drop=True) + + assert not flatfile.exists() + assert not temp_db_conn.table_exists('planet_osm_nodes') + + +def test_import_osm_data_default_cache(temp_db_cursor,osm2pgsql_options): + temp_db_cursor.execute('CREATE TABLE place (id INT)') + temp_db_cursor.execute('INSERT INTO place values (1)') + + osm2pgsql_options['osm2pgsql_cache'] = 0 + + database_import.import_osm_data(Path(__file__), osm2pgsql_options) + + +def test_truncate_database_tables(temp_db_conn, temp_db_cursor, table_factory): + tables = ('word', 'placex', 'place_addressline', 'location_area', + 'location_area_country', 'location_property', + 'location_property_tiger', 'location_property_osmline', + 'location_postcode', 'search_name', 'location_road_23') + for table in tables: + table_factory(table, content=(1, 2, 3)) + + database_import.truncate_data_tables(temp_db_conn, max_word_frequency=23) + + for table in tables: + assert temp_db_cursor.table_rows(table) == 0 + + +@pytest.mark.parametrize("threads", (1, 5)) +def test_load_data(dsn, src_dir, place_row, placex_table, osmline_table, word_table, + temp_db_cursor, threads): + for func in ('precompute_words', 'getorcreate_housenumber_id', 'make_standard_name'): + temp_db_cursor.execute("""CREATE FUNCTION {} (src TEXT) + RETURNS TEXT AS $$ SELECT 'a'::TEXT $$ LANGUAGE SQL + """.format(func)) + for oid in range(100, 130): + place_row(osm_id=oid) + place_row(osm_type='W', osm_id=342, cls='place', typ='houses', + geom='SRID=4326;LINESTRING(0 0, 10 10)') + + database_import.load_data(dsn, src_dir / 'data', threads) + + assert temp_db_cursor.table_rows('placex') == 30 + assert temp_db_cursor.table_rows('location_property_osmline') == 1 + +@pytest.mark.parametrize("languages", (False, True)) +def test_create_country_names(temp_db_conn, temp_db_cursor, def_config, + temp_db_with_extensions, monkeypatch, languages): + if languages: + monkeypatch.setenv('NOMINATIM_LANGUAGES', 'fr,en') + temp_db_cursor.execute("""CREATE FUNCTION make_standard_name (name TEXT) + RETURNS TEXT AS $$ SELECT 'a'::TEXT $$ LANGUAGE SQL + """) + temp_db_cursor.execute('CREATE TABLE country_name (country_code varchar(2), name hstore)') + temp_db_cursor.execute('CREATE TABLE word (code varchar(2))') + temp_db_cursor.execute("""INSERT INTO country_name VALUES ('us', + '"name"=>"us","name:af"=>"us"')""") + temp_db_cursor.execute("""CREATE OR REPLACE FUNCTION getorcreate_country(lookup_word TEXT, + lookup_country_code varchar(2)) + RETURNS INTEGER + AS $$ + BEGIN + INSERT INTO word VALUES (lookup_country_code); + RETURN 5; + END; + $$ + LANGUAGE plpgsql; + """) + database_import.create_country_names(temp_db_conn, def_config) + if languages: + assert temp_db_cursor.table_rows('word') == 4 + else: + assert temp_db_cursor.table_rows('word') == 5