Nominatim/test/python/test_indexing.py
2021-04-30 11:30:51 +02:00

282 lines
12 KiB
Python

"""
Tests for running the indexing.
"""
import itertools
import psycopg2
import pytest
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,
rank_address SMALLINT,
rank_search SMALLINT,
indexed_status SMALLINT,
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,
osm_id BIGINT,
address HSTORE,
token_info JSONB,
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 $$
BEGIN
IF NEW.indexed_status = 0 and OLD.indexed_status != 0 THEN
NEW.indexed_date = now();
END IF;
RETURN NEW;
END; $$ LANGUAGE plpgsql;""")
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;
""")
cur.execute("""CREATE OR REPLACE FUNCTION get_interpolation_address(in_address HSTORE, wayid BIGINT)
RETURNS HSTORE AS $$
BEGIN
RETURN in_address;
END;
$$ 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:
cur.execute(query)
return cur.fetchone()[0]
def add_place(self, cls='place', typ='locality',
rank_search=30, rank_address=30, sector=20):
next_id = next(self.placex_id)
with self.conn.cursor() as cur:
cur.execute("""INSERT INTO placex
(place_id, class, type, rank_search, rank_address,
indexed_status, geometry_sector)
VALUES (%s, %s, %s, %s, %s, 1, %s)""",
(next_id, cls, typ, rank_search, rank_address, sector))
return next_id
def add_admin(self, **kwargs):
kwargs['cls'] = 'boundary'
kwargs['typ'] = 'administrative'
return self.add_place(**kwargs)
def add_osmline(self, sector=20):
next_id = next(self.osmline_id)
with self.conn.cursor() as cur:
cur.execute("""INSERT INTO location_property_osmline
(place_id, osm_id, indexed_status, geometry_sector)
VALUES (%s, %s, 1, %s)""",
(next_id, 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')
def osmline_unindexed(self):
return self.scalar('SELECT count(*) from location_property_osmline where indexed_status > 0')
@pytest.fixture
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_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()
assert 31 == test_db.placex_unindexed()
assert 1 == test_db.osmline_unindexed()
idx = indexer.Indexer('dbname=test_nominatim_python_unittest', test_tokenizer, threads)
idx.index_by_rank(0, 30)
assert 0 == test_db.placex_unindexed()
assert 0 == test_db.osmline_unindexed()
assert 0 == test_db.scalar("""SELECT count(*) from placex
WHERE indexed_status = 0 and indexed_date is null""")
# ranks come in order of rank address
assert 0 == test_db.scalar("""
SELECT count(*) FROM placex p WHERE rank_address > 0
AND indexed_date >= (SELECT min(indexed_date) FROM placex o
WHERE p.rank_address < o.rank_address)""")
# placex rank < 30 objects come before interpolations
assert 0 == test_db.scalar(
"""SELECT count(*) FROM placex WHERE rank_address < 30
AND indexed_date > (SELECT min(indexed_date) FROM location_property_osmline)""")
# placex rank = 30 objects come after interpolations
assert 0 == test_db.scalar(
"""SELECT count(*) FROM placex WHERE rank_address = 30
AND indexed_date < (SELECT max(indexed_date) FROM location_property_osmline)""")
# rank 0 comes after rank 29 and before rank 30
assert 0 == test_db.scalar(
"""SELECT count(*) FROM placex WHERE rank_address < 30
AND indexed_date > (SELECT min(indexed_date) FROM placex WHERE rank_address = 0)""")
assert 0 == test_db.scalar(
"""SELECT count(*) FROM placex WHERE rank_address = 30
AND indexed_date < (SELECT max(indexed_date) FROM placex WHERE rank_address = 0)""")
@pytest.mark.parametrize("threads", [1, 15])
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()
assert 31 == test_db.placex_unindexed()
assert 1 == test_db.osmline_unindexed()
idx = indexer.Indexer('dbname=test_nominatim_python_unittest',
test_tokenizer, threads)
idx.index_by_rank(4, 15)
assert 19 == test_db.placex_unindexed()
assert 1 == test_db.osmline_unindexed()
assert 0 == test_db.scalar("""
SELECT count(*) FROM placex
WHERE indexed_status = 0 AND not rank_address between 4 and 15""")
@pytest.mark.parametrize("threads", [1, 15])
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()
assert 31 == test_db.placex_unindexed()
assert 1 == test_db.osmline_unindexed()
idx = indexer.Indexer('dbname=test_nominatim_python_unittest', test_tokenizer, threads)
idx.index_by_rank(28, 30)
assert 27 == test_db.placex_unindexed()
assert 0 == test_db.osmline_unindexed()
assert 0 == test_db.scalar("""
SELECT count(*) FROM placex
WHERE indexed_status = 0 AND rank_address between 1 and 27""")
@pytest.mark.parametrize("threads", [1, 15])
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):
test_db.add_place(rank_address=rank, rank_search=rank)
test_db.add_osmline()
assert 37 == test_db.placex_unindexed()
assert 1 == test_db.osmline_unindexed()
idx = indexer.Indexer('dbname=test_nominatim_python_unittest', test_tokenizer, threads)
idx.index_boundaries(0, 30)
assert 31 == test_db.placex_unindexed()
assert 1 == test_db.osmline_unindexed()
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()