safety-screen/Nominatim
1
0
Fork 0
forked from hans/Nominatim
Code Pull Requests Activity

move postcode normalization into tokenizer

Browse Source
This commit is contained in:
Sarah Hoffmann
2021-04-25 18:26:36 +02:00
parent d8ed1bfc60
commit ffc2d82b0e
12 changed files with 181 additions and 74 deletions
Download Patch File Download Diff File Expand all files Collapse all files

17
test/python/test_tools_postcodes.py
View File

@@ -5,6 +5,11 @@ Tests for functions to maintain the artificial postcode table.
import pytest
from nominatim.tools import postcodes
import dummy_tokenizer
@pytest.fixture
def tokenizer():
return dummy_tokenizer.DummyTokenizer(None, None)
@pytest.fixture
def postcode_table(temp_db_with_extensions, temp_db_cursor, table_factory,
@@ -20,26 +25,26 @@ def postcode_table(temp_db_with_extensions, temp_db_cursor, table_factory,
postcode TEXT,
geometry GEOMETRY(Geometry, 4326)""")
temp_db_cursor.execute('CREATE SEQUENCE seq_place')
temp_db_cursor.execute("""CREATE OR REPLACE FUNCTION getorcreate_postcode_id(postcode TEXT)
RETURNS INTEGER AS $$ BEGIN RETURN 1; END; $$ LANGUAGE plpgsql;
temp_db_cursor.execute("""CREATE OR REPLACE FUNCTION token_normalized_postcode(postcode TEXT)
RETURNS TEXT AS $$ BEGIN RETURN postcode; END; $$ LANGUAGE plpgsql;
""")
def test_import_postcodes_empty(dsn, temp_db_cursor, postcode_table, tmp_path):
postcodes.import_postcodes(dsn, tmp_path)
def test_import_postcodes_empty(dsn, temp_db_cursor, postcode_table, tmp_path, tokenizer):
postcodes.import_postcodes(dsn, tmp_path, tokenizer)
assert temp_db_cursor.table_exists('gb_postcode')
assert temp_db_cursor.table_exists('us_postcode')
assert temp_db_cursor.table_rows('location_postcode') == 0
def test_import_postcodes_from_placex(dsn, temp_db_cursor, postcode_table, tmp_path):
def test_import_postcodes_from_placex(dsn, temp_db_cursor, postcode_table, tmp_path, tokenizer):
temp_db_cursor.execute("""
INSERT INTO placex (place_id, country_code, address, geometry)
VALUES (1, 'xx', '"postcode"=>"9486"', 'SRID=4326;POINT(10 12)')
""")
postcodes.import_postcodes(dsn, tmp_path)
postcodes.import_postcodes(dsn, tmp_path, tokenizer)
rows = temp_db_cursor.row_set(""" SELECT postcode, country_code,
ST_X(geometry), ST_Y(geometry)