move default country name creation to tokenizer

The new function is also used, when a country us updated. All SQL
function related to country names have been removed.
This commit is contained in:
Sarah Hoffmann
2021-04-27 11:37:18 +02:00
parent dc700c25b6
commit bef300305e
8 changed files with 105 additions and 134 deletions

View File

@@ -121,9 +121,8 @@ def table_factory(temp_db_cursor):
def mk_table(name, definition='id INT', content=None):
temp_db_cursor.execute('CREATE TABLE {} ({})'.format(name, definition))
if content is not None:
if not isinstance(content, str):
content = '),('.join([str(x) for x in content])
temp_db_cursor.execute("INSERT INTO {} VALUES ({})".format(name, content))
psycopg2.extras.execute_values(
temp_db_cursor, "INSERT INTO {} VALUES %s".format(name), content)
return mk_table
@@ -290,7 +289,7 @@ def osm2pgsql_options(temp_db):
@pytest.fixture
def sql_preprocessor(temp_db_conn, tmp_path, monkeypatch, table_factory):
table_factory('country_name', 'partition INT', (0, 1, 2))
table_factory('country_name', 'partition INT', ((0, ), (1, ), (2, )))
cfg = Configuration(None, SRC_DIR.resolve() / 'settings')
cfg.set_libdirs(module='.', osm2pgsql='.', php=SRC_DIR / 'lib-php',
sql=tmp_path, data=SRC_DIR / 'data')
@@ -299,9 +298,10 @@ def sql_preprocessor(temp_db_conn, tmp_path, monkeypatch, table_factory):
@pytest.fixture
def tokenizer_mock(monkeypatch, property_table, temp_db_conn, dsn):
def tokenizer_mock(monkeypatch, property_table, temp_db_conn, tmp_path):
""" Sets up the configuration so that the test dummy tokenizer will be
loaded.
loaded when the tokenizer factory is used. Also returns a factory
with which a new dummy tokenizer may be created.
"""
monkeypatch.setenv('NOMINATIM_TOKENIZER', 'dummy')
@@ -310,3 +310,8 @@ def tokenizer_mock(monkeypatch, property_table, temp_db_conn, dsn):
monkeypatch.setattr(importlib, "import_module", _import_dummy)
properties.set_property(temp_db_conn, 'tokenizer', 'dummy')
def _create_tokenizer():
return dummy_tokenizer.DummyTokenizer(None, None)
return _create_tokenizer

View File

@@ -13,6 +13,7 @@ class DummyTokenizer:
self.dsn = dsn
self.data_dir = data_dir
self.init_state = None
self.analyser_cache = {}
def init_new_db(self, config):
@@ -26,7 +27,7 @@ class DummyTokenizer:
def name_analyzer(self):
return DummyNameAnalyzer()
return DummyNameAnalyzer(self.analyser_cache)
class DummyNameAnalyzer:
@@ -38,18 +39,20 @@ class DummyNameAnalyzer:
self.close()
def __init__(self, cache):
self.analyser_cache = cache
cache['countries'] = []
def close(self):
""" Free all resources used by the analyzer.
"""
pass
def add_postcodes_from_db(self):
pass
def process_place(self, place):
""" Determine tokenizer information about the given place.
Returns a JSON-serialisable structure that will be handed into
the database via the token_info field.
"""
def add_country_names(self, code, names):
self.analyser_cache['countries'].append((code, names))
def process_place(self, place):
return {}

View File

@@ -143,7 +143,8 @@ def test_truncate_database_tables(temp_db_conn, temp_db_cursor, table_factory):
'location_property_tiger', 'location_property_osmline',
'location_postcode', 'search_name', 'location_road_23')
for table in tables:
table_factory(table, content=(1, 2, 3))
table_factory(table, content=((1, ), (2, ), (3, )))
assert temp_db_cursor.table_rows(table) == 3
database_import.truncate_data_tables(temp_db_conn)
@@ -168,31 +169,28 @@ def test_load_data(dsn, src_dir, place_row, placex_table, osmline_table, word_ta
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):
@pytest.mark.parametrize("languages", (None, ' fr,en'))
def test_create_country_names(temp_db_with_extensions, temp_db_conn, temp_db_cursor,
table_factory, tokenizer_mock, languages):
table_factory('country_name', 'country_code varchar(2), name hstore',
content=(('us', '"name"=>"us1","name:af"=>"us2"'),
('fr', '"name"=>"Fra", "name:en"=>"Fren"')))
assert temp_db_cursor.scalar("SELECT count(*) FROM country_name") == 2
tokenizer = tokenizer_mock()
database_import.create_country_names(temp_db_conn, tokenizer, languages)
assert len(tokenizer.analyser_cache['countries']) == 2
result_set = {k: set(v) for k, v in tokenizer.analyser_cache['countries']}
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
assert result_set == {'us' : set(('us', 'us1', 'United States')),
'fr' : set(('fr', 'Fra', 'Fren'))}
else:
assert temp_db_cursor.table_rows('word') == 5
assert result_set == {'us' : set(('us', 'us1', 'us2', 'United States')),
'fr' : set(('fr', 'Fra', 'Fren'))}

View File

@@ -11,9 +11,7 @@ def sql_tmp_path(tmp_path, def_config):
return tmp_path
@pytest.fixture
def conn(temp_db_conn, table_factory, monkeypatch):
monkeypatch.setenv('NOMINATIM_DATABASE_MODULE_PATH', '.')
table_factory('country_name', 'partition INT', (0, 1, 2))
def conn(sql_preprocessor, temp_db_conn):
return temp_db_conn