remove country restriction from tokenizer

[nominatim.git] / 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 453248340d9b876132ee21cdee709eef7c860f98..aa90f8dba02ed3aaf3380362122dcb6029ce1f47 100644 (file)
--- a/test/python/test_tools_database_import.py
+++ b/test/python/test_tools_database_import.py
@@ -1,10 +1,10 @@
 """
 Tests for functions to import a new database.
 """
 """
 Tests for functions to import a new database.
 """

+from pathlib import Path
+

 import pytest
 import psycopg2
 import pytest
 import psycopg2

-import sys
-from pathlib import Path

 
 from nominatim.tools import database_import
 from nominatim.errors import UsageError
 
 from nominatim.tools import database_import
 from nominatim.errors import UsageError


@@ -34,9 +34,9 @@ def test_setup_skeleton(src_dir, nonexistant_db, no_partitions):
     try:
         with conn.cursor() as cur:
             cur.execute("SELECT distinct partition FROM country_name")
     try:
         with conn.cursor() as cur:
             cur.execute("SELECT distinct partition FROM country_name")

-            partitions = set([r[0] for r in list(cur)])
+            partitions = set((r[0] for r in list(cur)))

             if no_partitions:
             if no_partitions:

-                assert partitions == set([0])
+                assert partitions == set((0, ))

             else:
                 assert len(partitions) > 10
     finally:
             else:
                 assert len(partitions) > 10
     finally:


@@ -67,10 +67,11 @@ def test_create_db_missing_ro_user(nonexistant_db):
         database_import.create_db('dbname=' + nonexistant_db, rouser='sdfwkjkjgdugu2;jgsafkljas;')
 
 
         database_import.create_db('dbname=' + nonexistant_db, rouser='sdfwkjkjgdugu2;jgsafkljas;')
 
 

-def test_setup_extensions(temp_db_conn, temp_db_cursor):
+def test_setup_extensions(temp_db_conn, table_factory):

     database_import.setup_extensions(temp_db_conn)
 
     database_import.setup_extensions(temp_db_conn)
 

-    temp_db_cursor.execute('CREATE TABLE t (h HSTORE, geom GEOMETRY(Geometry, 4326))')
+    # Use table creation to check that hstore and geometry types are available.
+    table_factory('t', 'h HSTORE, geom GEOMETRY(Geometry, 4326)')

 
 
 def test_setup_extensions_old_postgis(temp_db_conn, monkeypatch):
 
 
 def test_setup_extensions_old_postgis(temp_db_conn, monkeypatch):


@@ -80,90 +81,61 @@ def test_setup_extensions_old_postgis(temp_db_conn, monkeypatch):
         database_import.setup_extensions(temp_db_conn)
 
 
         database_import.setup_extensions(temp_db_conn)
 
 

-def test_install_module(tmp_path):
-    src_dir = tmp_path / 'source'
-    src_dir.mkdir()
-    (src_dir / 'nominatim.so').write_text('TEST nomiantim.so')
-
-    project_dir = tmp_path / 'project'
-    project_dir.mkdir()
-
-    database_import.install_module(src_dir, project_dir, '')
-
-    outfile = project_dir / 'module' / 'nominatim.so'
-
-    assert outfile.exists()
-    assert outfile.read_text() == 'TEST nomiantim.so'
-    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_import_base_data(dsn, src_dir, temp_db_with_extensions, temp_db_cursor):
+    database_import.import_base_data(dsn, src_dir / 'data')

 
 

+    assert temp_db_cursor.table_rows('country_name') > 0

 
 

-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_ignore_partitions(dsn, src_dir, temp_db_with_extensions,
+                                            temp_db_cursor):
+    database_import.import_base_data(dsn, src_dir / 'data', ignore_partitions=True)

 
 

+    assert temp_db_cursor.table_rows('country_name') > 0
+    assert temp_db_cursor.table_rows('country_name', where='partition != 0') == 0

 
 

-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')
-    database_import.import_base_data('dbname=' + temp_db, src_dir / 'data')

 
 

-    assert temp_db_cursor.scalar('SELECT count(*) FROM country_name') > 0
+def test_import_osm_data_simple(table_factory, osm2pgsql_options):
+    table_factory('place', content=((1, ), ))

 
 

+    database_import.import_osm_data(Path('file.pbf'), osm2pgsql_options)

 
 

-def test_import_base_data_ignore_partitions(src_dir, temp_db, temp_db_cursor):
-    temp_db_cursor.execute('CREATE EXTENSION hstore')
-    temp_db_cursor.execute('CREATE EXTENSION postgis')
-    database_import.import_base_data('dbname=' + temp_db, src_dir / 'data',
-                                     ignore_partitions=True)
-
-    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_multifile(table_factory, tmp_path, osm2pgsql_options):
+    table_factory('place', content=((1, ), ))
+    osm2pgsql_options['osm2pgsql_cache'] = 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)')
+    files = [tmp_path / 'file1.osm', tmp_path / 'file2.osm']
+    for f in files:
+        f.write_text('test')

 
 

-    database_import.import_osm_data('file.pdf', osm2pgsql_options)
+    database_import.import_osm_data(files, osm2pgsql_options)

 
 
 
 

-def test_import_osm_data_simple_no_data(temp_db_cursor,osm2pgsql_options):
-    temp_db_cursor.execute('CREATE TABLE place (id INT)')
+def test_import_osm_data_simple_no_data(table_factory, osm2pgsql_options):
+    table_factory('place')

 
     with pytest.raises(UsageError, match='No data.*'):
 
     with pytest.raises(UsageError, match='No data.*'):

-        database_import.import_osm_data('file.pdf', osm2pgsql_options)
+        database_import.import_osm_data(Path('file.pbf'), 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)')
+def test_import_osm_data_drop(table_factory, temp_db_conn, tmp_path, osm2pgsql_options):
+    table_factory('place', content=((1, ), ))
+    table_factory('planet_osm_nodes')

 
     flatfile = tmp_path / 'flatfile'
     flatfile.write_text('touch')
 
     osm2pgsql_options['flatnode_file'] = str(flatfile.resolve())
 
 
     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)
+    database_import.import_osm_data(Path('file.pbf'), osm2pgsql_options, drop=True)

 
     assert not flatfile.exists()
     assert not temp_db_conn.table_exists('planet_osm_nodes')
 
 
 
     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)')
+def test_import_osm_data_default_cache(table_factory, osm2pgsql_options):
+    table_factory('place', content=((1, ), ))

 
     osm2pgsql_options['osm2pgsql_cache'] = 0
 
 
     osm2pgsql_options['osm2pgsql_cache'] = 0
 


@@ -171,23 +143,24 @@ def test_import_osm_data_default_cache(temp_db_cursor,osm2pgsql_options):
 
 
 def test_truncate_database_tables(temp_db_conn, temp_db_cursor, table_factory):
 
 
 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',
+    tables = ('placex', 'place_addressline', 'location_area',
+              'location_area_country',

               'location_property_tiger', 'location_property_osmline',
               'location_postcode', 'search_name', 'location_road_23')
     for table in tables:
               '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, max_word_frequency=23)
+    database_import.truncate_data_tables(temp_db_conn)

 
     for table in tables:
         assert temp_db_cursor.table_rows(table) == 0
 
 
 @pytest.mark.parametrize("threads", (1, 5))
 
     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 ('make_keywords', 'getorcreate_housenumber_id', 'make_standard_name'):
+def test_load_data(dsn, 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))
         temp_db_cursor.execute("""CREATE FUNCTION {} (src TEXT)
                                   RETURNS TEXT AS $$ SELECT 'a'::TEXT $$ LANGUAGE SQL
                                """.format(func))


@@ -196,7 +169,33 @@ def test_load_data(dsn, src_dir, place_row, placex_table, osmline_table, word_ta
     place_row(osm_type='W', osm_id=342, cls='place', typ='houses',
               geom='SRID=4326;LINESTRING(0 0, 10 10)')
 
     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)
+    database_import.load_data(dsn, threads)

 
     assert temp_db_cursor.table_rows('placex') == 30
     assert temp_db_cursor.table_rows('location_property_osmline') == 1
 
     assert temp_db_cursor.table_rows('placex') == 30
     assert temp_db_cursor.table_rows('location_property_osmline') == 1

+
+
+@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.values()) for k, v in tokenizer.analyser_cache['countries']}
+
+    if languages:
+        assert result_set == {'us' : set(('us', 'us1', 'United States')),
+                              'fr' : set(('fr', 'Fra', 'Fren'))}
+    else:
+        assert result_set == {'us' : set(('us', 'us1', 'us2', 'United States')),
+                              'fr' : set(('fr', 'Fra', 'Fren'))}