]> git.openstreetmap.org Git - nominatim.git/commitdiff
Merge pull request #2276 from lonvia/port-country-code-creation-to-python
authorSarah Hoffmann <lonvia@denofr.de>
Fri, 16 Apr 2021 14:57:04 +0000 (16:57 +0200)
committerGitHub <noreply@github.com>
Fri, 16 Apr 2021 14:57:04 +0000 (16:57 +0200)
Port country code creation to python

nominatim/clicmd/setup.py
nominatim/tools/postcodes.py [new file with mode: 0644]
test/python/conftest.py
test/python/test_cli.py
test/python/test_tools_postcodes.py [new file with mode: 0644]

index 100fdf6747cbe814b401a45d6531c79d161ff81a..fe7c8dc18de4ee650e7bbb42824c813f4e2d9753 100644 (file)
@@ -6,7 +6,6 @@ from pathlib import Path
 
 import psutil
 
-from nominatim.tools.exec_utils import run_legacy_script
 from nominatim.db.connection import connect
 from nominatim.db import status, properties
 from nominatim.version import NOMINATIM_VERSION
@@ -56,6 +55,7 @@ class SetupAll:
         from ..tools import database_import
         from ..tools import refresh
         from ..indexer.indexer import Indexer
+        from ..tools import postcodes
 
         if args.osm_file and not Path(args.osm_file).is_file():
             LOG.fatal("OSM file '%s' does not exist.", args.osm_file)
@@ -116,8 +116,7 @@ class SetupAll:
                                       args.threads or psutil.cpu_count() or 1)
 
             LOG.warning('Calculate postcodes')
-            run_legacy_script('setup.php', '--calculate-postcodes',
-                              nominatim_env=args, throw_on_fail=not args.ignore_errors)
+            postcodes.import_postcodes(args.config.get_libpq_dsn(), args.project_dir)
 
         if args.continue_at is None or args.continue_at in ('load-data', 'indexing'):
             LOG.warning('Indexing places')
diff --git a/nominatim/tools/postcodes.py b/nominatim/tools/postcodes.py
new file mode 100644 (file)
index 0000000..0a568cb
--- /dev/null
@@ -0,0 +1,80 @@
+"""
+Functions for importing, updating and otherwise maintaining the table
+of artificial postcode centroids.
+"""
+
+from nominatim.db.utils import execute_file
+from nominatim.db.connection import connect
+
+def import_postcodes(dsn, project_dir):
+    """ Set up the initial list of postcodes.
+    """
+
+    with connect(dsn) as conn:
+        conn.drop_table('gb_postcode')
+        conn.drop_table('us_postcode')
+
+        with conn.cursor() as cur:
+            cur.execute("""CREATE TABLE gb_postcode (
+                            id integer,
+                            postcode character varying(9),
+                            geometry GEOMETRY(Point, 4326))""")
+
+        with conn.cursor() as cur:
+            cur.execute("""CREATE TABLE us_postcode (
+                            postcode text,
+                            x double precision,
+                            y double precision)""")
+        conn.commit()
+
+        gb_postcodes = project_dir / 'gb_postcode_data.sql.gz'
+        if gb_postcodes.is_file():
+            execute_file(dsn, gb_postcodes)
+
+        us_postcodes = project_dir / 'us_postcode_data.sql.gz'
+        if us_postcodes.is_file():
+            execute_file(dsn, us_postcodes)
+
+        with conn.cursor() as cur:
+            cur.execute("TRUNCATE location_postcode")
+            cur.execute("""
+                INSERT INTO location_postcode
+                 (place_id, indexed_status, country_code, postcode, geometry)
+                SELECT nextval('seq_place'), 1, country_code,
+                       upper(trim (both ' ' from address->'postcode')) as pc,
+                       ST_Centroid(ST_Collect(ST_Centroid(geometry)))
+                  FROM placex
+                 WHERE address ? 'postcode' AND address->'postcode' NOT SIMILAR TO '%(,|;)%'
+                       AND geometry IS NOT null
+                 GROUP BY country_code, pc
+            """)
+
+            cur.execute("""
+                INSERT INTO location_postcode
+                 (place_id, indexed_status, country_code, postcode, geometry)
+                SELECT nextval('seq_place'), 1, 'us', postcode,
+                       ST_SetSRID(ST_Point(x,y),4326)
+                  FROM us_postcode WHERE postcode NOT IN
+                        (SELECT postcode FROM location_postcode
+                          WHERE country_code = 'us')
+            """)
+
+            cur.execute("""
+                INSERT INTO location_postcode
+                 (place_id, indexed_status, country_code, postcode, geometry)
+                SELECT nextval('seq_place'), 1, 'gb', postcode, geometry
+                  FROM gb_postcode WHERE postcode NOT IN
+                           (SELECT postcode FROM location_postcode
+                             WHERE country_code = 'gb')
+            """)
+
+            cur.execute("""
+                    DELETE FROM word WHERE class='place' and type='postcode'
+                    and word NOT IN (SELECT postcode FROM location_postcode)
+            """)
+
+            cur.execute("""
+                SELECT count(getorcreate_postcode_id(v)) FROM
+                (SELECT distinct(postcode) as v FROM location_postcode) p
+            """)
+        conn.commit()
index 871365d90214f0515365179644d6ffe69e8a0100..0d1cd2f37168f8446182d030d2e26e4427922ead 100644 (file)
@@ -33,8 +33,6 @@ class _TestingCursor(psycopg2.extras.DictCursor):
         """ Execute a query and return the result as a set of tuples.
         """
         self.execute(sql, params)
-        if self.rowcount == 1:
-            return set(tuple(self.fetchone()))
 
         return set((tuple(row) for row in self))
 
index eb0ee58487b5917112824ad4b0ce2940d5d21f3a..38bbaefee82688bbad8194ee5623884222af6b03 100644 (file)
@@ -21,6 +21,7 @@ import nominatim.tools.check_database
 import nominatim.tools.database_import
 import nominatim.tools.freeze
 import nominatim.tools.refresh
+import nominatim.tools.postcodes
 
 from mocks import MockParamCapture
 
@@ -96,13 +97,13 @@ def test_import_full(temp_db, mock_func_factory):
         mock_func_factory(nominatim.tools.database_import, 'create_search_indices'),
         mock_func_factory(nominatim.tools.database_import, 'create_country_names'),
         mock_func_factory(nominatim.tools.refresh, 'load_address_levels_from_file'),
+        mock_func_factory(nominatim.tools.postcodes, 'import_postcodes'),
         mock_func_factory(nominatim.indexer.indexer.Indexer, 'index_full'),
         mock_func_factory(nominatim.tools.refresh, 'setup_website'),
         mock_func_factory(nominatim.db.properties, 'set_property')
     ]
 
     cf_mock = mock_func_factory(nominatim.tools.refresh, 'create_functions')
-    mock_func_factory(nominatim.clicmd.setup, 'run_legacy_script')
 
     assert 0 == call_nominatim('import', '--osm-file', __file__)
 
diff --git a/test/python/test_tools_postcodes.py b/test/python/test_tools_postcodes.py
new file mode 100644 (file)
index 0000000..1fc060b
--- /dev/null
@@ -0,0 +1,50 @@
+"""
+Tests for functions to maintain the artificial postcode table.
+"""
+
+import pytest
+
+from nominatim.tools import postcodes
+
+@pytest.fixture
+def postcode_table(temp_db_with_extensions, temp_db_cursor, table_factory,
+                   placex_table, word_table):
+    table_factory('location_postcode',
+                  """ place_id BIGINT,
+                      parent_place_id BIGINT,
+                      rank_search SMALLINT,
+                      rank_address SMALLINT,
+                      indexed_status SMALLINT,
+                      indexed_date TIMESTAMP,
+                      country_code varchar(2),
+                      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;
+                           """)
+
+
+def test_import_postcodes_empty(dsn, temp_db_cursor, postcode_table, tmp_path):
+    postcodes.import_postcodes(dsn, tmp_path)
+
+    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):
+    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)
+
+    rows = temp_db_cursor.row_set(""" SELECT postcode, country_code,
+                                      ST_X(geometry), ST_Y(geometry)
+                                      FROM location_postcode""")
+    print(rows)
+    assert len(rows) == 1
+    assert rows == set((('9486', 'xx', 10, 12), ))
+