X-Git-Url: https://git.openstreetmap.org./nominatim.git/blobdiff_plain/2c909c1f0c7c6e6a176eb5cc15e8d6a45449ebe8..d8ed1bfc601c6eb19b2db50df466654e843f0777:/test/python/test_indexing.py diff --git a/test/python/test_indexing.py b/test/python/test_indexing.py index 2fe21e80..d6876906 100644 --- a/test/python/test_indexing.py +++ b/test/python/test_indexing.py @@ -5,17 +5,20 @@ import itertools import psycopg2 import pytest -from nominatim.indexer.indexer import Indexer +from nominatim.indexer import indexer +from nominatim.tokenizer import factory class IndexerTestDB: def __init__(self, conn): self.placex_id = itertools.count(100000) self.osmline_id = itertools.count(500000) + self.postcode_id = itertools.count(700000) self.conn = conn self.conn.set_isolation_level(0) with self.conn.cursor() as cur: + cur.execute('CREATE EXTENSION hstore') cur.execute("""CREATE TABLE placex (place_id BIGINT, class TEXT, type TEXT, @@ -25,12 +28,20 @@ class IndexerTestDB: indexed_date TIMESTAMP, partition SMALLINT, admin_level SMALLINT, + address HSTORE, + token_info JSONB, geometry_sector INTEGER)""") cur.execute("""CREATE TABLE location_property_osmline ( place_id BIGINT, indexed_status SMALLINT, indexed_date TIMESTAMP, geometry_sector INTEGER)""") + cur.execute("""CREATE TABLE location_postcode ( + place_id BIGINT, + indexed_status SMALLINT, + indexed_date TIMESTAMP, + country_code varchar(2), + postcode TEXT)""") cur.execute("""CREATE OR REPLACE FUNCTION date_update() RETURNS TRIGGER AS $$ BEGIN @@ -39,10 +50,21 @@ class IndexerTestDB: END IF; RETURN NEW; END; $$ LANGUAGE plpgsql;""") - cur.execute("""CREATE TRIGGER placex_update BEFORE UPDATE ON placex - FOR EACH ROW EXECUTE PROCEDURE date_update()""") - cur.execute("""CREATE TRIGGER osmline_update BEFORE UPDATE ON location_property_osmline - FOR EACH ROW EXECUTE PROCEDURE date_update()""") + cur.execute("""CREATE OR REPLACE FUNCTION placex_prepare_update(p placex, + OUT name HSTORE, + OUT address HSTORE, + OUT country_feature VARCHAR) + AS $$ + BEGIN + address := p.address; + name := p.address; + END; + $$ LANGUAGE plpgsql STABLE; + """) + for table in ('placex', 'location_property_osmline', 'location_postcode'): + cur.execute("""CREATE TRIGGER {0}_update BEFORE UPDATE ON {0} + FOR EACH ROW EXECUTE PROCEDURE date_update() + """.format(table)) def scalar(self, query): with self.conn.cursor() as cur: @@ -74,6 +96,15 @@ class IndexerTestDB: (next_id, sector)) return next_id + def add_postcode(self, country, postcode): + next_id = next(self.postcode_id) + with self.conn.cursor() as cur: + cur.execute("""INSERT INTO location_postcode + (place_id, indexed_status, country_code, postcode) + VALUES (%s, 1, %s, %s)""", + (next_id, country, postcode)) + return next_id + def placex_unindexed(self): return self.scalar('SELECT count(*) from placex where indexed_status > 0') @@ -82,14 +113,18 @@ class IndexerTestDB: @pytest.fixture -def test_db(temp_db): - conn = psycopg2.connect(database=temp_db) - yield IndexerTestDB(conn) - conn.close() +def test_db(temp_db_conn): + yield IndexerTestDB(temp_db_conn) + + +@pytest.fixture +def test_tokenizer(tokenizer_mock, def_config, tmp_path): + def_config.project_dir = tmp_path + return factory.create_tokenizer(def_config) @pytest.mark.parametrize("threads", [1, 15]) -def test_index_full(test_db, threads): +def test_index_all_by_rank(test_db, threads, test_tokenizer): for rank in range(31): test_db.add_place(rank_address=rank, rank_search=rank) test_db.add_osmline() @@ -97,7 +132,7 @@ def test_index_full(test_db, threads): assert 31 == test_db.placex_unindexed() assert 1 == test_db.osmline_unindexed() - idx = Indexer('dbname=test_nominatim_python_unittest', threads) + idx = indexer.Indexer('dbname=test_nominatim_python_unittest', test_tokenizer, threads) idx.index_by_rank(0, 30) assert 0 == test_db.placex_unindexed() @@ -128,7 +163,7 @@ def test_index_full(test_db, threads): @pytest.mark.parametrize("threads", [1, 15]) -def test_index_partial_without_30(test_db, threads): +def test_index_partial_without_30(test_db, threads, test_tokenizer): for rank in range(31): test_db.add_place(rank_address=rank, rank_search=rank) test_db.add_osmline() @@ -136,7 +171,8 @@ def test_index_partial_without_30(test_db, threads): assert 31 == test_db.placex_unindexed() assert 1 == test_db.osmline_unindexed() - idx = Indexer('dbname=test_nominatim_python_unittest', threads) + idx = indexer.Indexer('dbname=test_nominatim_python_unittest', + test_tokenizer, threads) idx.index_by_rank(4, 15) assert 19 == test_db.placex_unindexed() @@ -148,7 +184,7 @@ def test_index_partial_without_30(test_db, threads): @pytest.mark.parametrize("threads", [1, 15]) -def test_index_partial_with_30(test_db, threads): +def test_index_partial_with_30(test_db, threads, test_tokenizer): for rank in range(31): test_db.add_place(rank_address=rank, rank_search=rank) test_db.add_osmline() @@ -156,7 +192,7 @@ def test_index_partial_with_30(test_db, threads): assert 31 == test_db.placex_unindexed() assert 1 == test_db.osmline_unindexed() - idx = Indexer('dbname=test_nominatim_python_unittest', threads) + idx = indexer.Indexer('dbname=test_nominatim_python_unittest', test_tokenizer, threads) idx.index_by_rank(28, 30) assert 27 == test_db.placex_unindexed() @@ -167,7 +203,7 @@ def test_index_partial_with_30(test_db, threads): WHERE indexed_status = 0 AND rank_address between 1 and 27""") @pytest.mark.parametrize("threads", [1, 15]) -def test_index_boundaries(test_db, threads): +def test_index_boundaries(test_db, threads, test_tokenizer): for rank in range(4, 10): test_db.add_admin(rank_address=rank, rank_search=rank) for rank in range(31): @@ -177,7 +213,7 @@ def test_index_boundaries(test_db, threads): assert 37 == test_db.placex_unindexed() assert 1 == test_db.osmline_unindexed() - idx = Indexer('dbname=test_nominatim_python_unittest', threads) + idx = indexer.Indexer('dbname=test_nominatim_python_unittest', test_tokenizer, threads) idx.index_boundaries(0, 30) assert 31 == test_db.placex_unindexed() @@ -186,3 +222,49 @@ def test_index_boundaries(test_db, threads): assert 0 == test_db.scalar(""" SELECT count(*) FROM placex WHERE indexed_status = 0 AND class != 'boundary'""") + + +@pytest.mark.parametrize("threads", [1, 15]) +def test_index_postcodes(test_db, threads, test_tokenizer): + for postcode in range(1000): + test_db.add_postcode('de', postcode) + for postcode in range(32000, 33000): + test_db.add_postcode('us', postcode) + + idx = indexer.Indexer('dbname=test_nominatim_python_unittest', test_tokenizer, threads) + idx.index_postcodes() + + assert 0 == test_db.scalar("""SELECT count(*) FROM location_postcode + WHERE indexed_status != 0""") + + +@pytest.mark.parametrize("analyse", [True, False]) +def test_index_full(test_db, analyse, test_tokenizer): + for rank in range(4, 10): + test_db.add_admin(rank_address=rank, rank_search=rank) + for rank in range(31): + test_db.add_place(rank_address=rank, rank_search=rank) + test_db.add_osmline() + for postcode in range(1000): + test_db.add_postcode('de', postcode) + + idx = indexer.Indexer('dbname=test_nominatim_python_unittest', test_tokenizer, 4) + idx.index_full(analyse=analyse) + + assert 0 == test_db.placex_unindexed() + assert 0 == test_db.osmline_unindexed() + assert 0 == test_db.scalar("""SELECT count(*) FROM location_postcode + WHERE indexed_status != 0""") + + +@pytest.mark.parametrize("threads", [1, 15]) +def test_index_reopen_connection(test_db, threads, monkeypatch, test_tokenizer): + monkeypatch.setattr(indexer.WorkerPool, "REOPEN_CONNECTIONS_AFTER", 15) + + for _ in range(1000): + test_db.add_place(rank_address=30, rank_search=30) + + idx = indexer.Indexer('dbname=test_nominatim_python_unittest', test_tokenizer, threads) + idx.index_by_rank(28, 30) + + assert 0 == test_db.placex_unindexed()