US TIGER data 2021 released

[nominatim.git] / test / python / test_indexing.py

diff --git a/test/python/test_indexing.py b/test/python/test_indexing.py
index 6692eba61d349b8908415d793078fba17b0db3af..60ad0bc4cbd3c5705891c60c847809ba74d7b985 100644 (file)
--- a/test/python/test_indexing.py
+++ b/test/python/test_indexing.py
@@ -2,10 +2,10 @@
 Tests for running the indexing.
 """
 import itertools
 Tests for running the indexing.
 """
 import itertools

-import psycopg2

 import pytest
 
 from nominatim.indexer import indexer
 import pytest
 
 from nominatim.indexer import indexer

+from nominatim.tokenizer import factory

 
 class IndexerTestDB:
 
 
 class IndexerTestDB:
 


@@ -17,18 +17,26 @@ class IndexerTestDB:
         self.conn = conn
         self.conn.set_isolation_level(0)
         with self.conn.cursor() as cur:
         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,
             cur.execute("""CREATE TABLE placex (place_id BIGINT,

+                                                name HSTORE,

                                                 class TEXT,
                                                 type TEXT,
                                                 class TEXT,
                                                 type TEXT,

+                                                linked_place_id BIGINT,

                                                 rank_address SMALLINT,
                                                 rank_search SMALLINT,
                                                 indexed_status SMALLINT,
                                                 indexed_date TIMESTAMP,
                                                 partition SMALLINT,
                                                 admin_level SMALLINT,
                                                 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,
                                                 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)""")
                                indexed_status SMALLINT,
                                indexed_date TIMESTAMP,
                                geometry_sector INTEGER)""")


@@ -46,6 +54,27 @@ class IndexerTestDB:
                              END IF;
                              RETURN NEW;
                            END; $$ LANGUAGE plpgsql;""")
                              END IF;
                              RETURN NEW;
                            END; $$ LANGUAGE plpgsql;""")

+            cur.execute("""CREATE OR REPLACE FUNCTION placex_prepare_update(p placex,
+                                                      OUT name HSTORE,
+                                                      OUT address HSTORE,
+                                                      OUT country_feature VARCHAR,
+                                                      OUT linked_place_id BIGINT)
+                           AS $$
+                           BEGIN
+                            address := p.address;
+                            name := p.name;
+                           END;
+                           $$ 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()
             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()


@@ -76,9 +105,9 @@ class IndexerTestDB:
         next_id = next(self.osmline_id)
         with self.conn.cursor() as cur:
             cur.execute("""INSERT INTO location_property_osmline
         next_id = next(self.osmline_id)
         with self.conn.cursor() as cur:
             cur.execute("""INSERT INTO location_property_osmline

-                              (place_id, indexed_status, geometry_sector)
-                              VALUES (%s, 1, %s)""",
-                        (next_id, sector))
+                              (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):
         return next_id
 
     def add_postcode(self, country, postcode):


@@ -94,7 +123,8 @@ class IndexerTestDB:
         return self.scalar('SELECT count(*) from placex where indexed_status > 0')
 
     def osmline_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')
+        return self.scalar("""SELECT count(*) from location_property_osmline
+                              WHERE indexed_status > 0""")

 
 
 @pytest.fixture
 
 
 @pytest.fixture


@@ -102,122 +132,133 @@ def test_db(temp_db_conn):
     yield IndexerTestDB(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])
 @pytest.mark.parametrize("threads", [1, 15])

-def test_index_all_by_rank(test_db, threads):
+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()
 
     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()
+    assert test_db.placex_unindexed() == 31
+    assert test_db.osmline_unindexed() == 1

 
 

-    idx = indexer.Indexer('dbname=test_nominatim_python_unittest', threads)
+    idx = indexer.Indexer('dbname=test_nominatim_python_unittest', test_tokenizer, threads)

     idx.index_by_rank(0, 30)
 
     idx.index_by_rank(0, 30)
 

-    assert 0 == test_db.placex_unindexed()
-    assert 0 == test_db.osmline_unindexed()
+    assert test_db.placex_unindexed() == 0
+    assert test_db.osmline_unindexed() == 0

 
 

-    assert 0 == test_db.scalar("""SELECT count(*) from placex
-                               WHERE indexed_status = 0 and indexed_date is null""")
+    assert test_db.scalar("""SELECT count(*) from placex
+                             WHERE indexed_status = 0 and indexed_date is null""") == 0

     # ranks come in order of rank address
     # ranks come in order of rank address

-    assert 0 == test_db.scalar("""
+    assert test_db.scalar("""

         SELECT count(*) FROM placex p WHERE rank_address > 0
           AND indexed_date >= (SELECT min(indexed_date) FROM placex o
         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)""")
+                               WHERE p.rank_address < o.rank_address)""") == 0

     # placex rank < 30 objects come before interpolations
     # placex rank < 30 objects come before interpolations

-    assert 0 == test_db.scalar(
+    assert test_db.scalar(

         """SELECT count(*) FROM placex WHERE rank_address < 30
         """SELECT count(*) FROM placex WHERE rank_address < 30

-             AND indexed_date > (SELECT min(indexed_date) FROM location_property_osmline)""")
+             AND indexed_date >
+                   (SELECT min(indexed_date) FROM location_property_osmline)""") == 0

     # placex rank = 30 objects come after interpolations
     # placex rank = 30 objects come after interpolations

-    assert 0 == test_db.scalar(
+    assert test_db.scalar(

         """SELECT count(*) FROM placex WHERE rank_address = 30
         """SELECT count(*) FROM placex WHERE rank_address = 30

-             AND indexed_date < (SELECT max(indexed_date) FROM location_property_osmline)""")
+             AND indexed_date <
+                   (SELECT max(indexed_date) FROM location_property_osmline)""") == 0

     # rank 0 comes after rank 29 and before rank 30
     # rank 0 comes after rank 29 and before rank 30

-    assert 0 == test_db.scalar(
+    assert test_db.scalar(

         """SELECT count(*) FROM placex WHERE rank_address < 30
         """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(
+             AND indexed_date >
+                   (SELECT min(indexed_date) FROM placex WHERE rank_address = 0)""") == 0
+    assert test_db.scalar(

         """SELECT count(*) FROM placex WHERE rank_address = 30
         """SELECT count(*) FROM placex WHERE rank_address = 30

-             AND indexed_date < (SELECT max(indexed_date) FROM placex WHERE rank_address = 0)""")
+             AND indexed_date <
+                   (SELECT max(indexed_date) FROM placex WHERE rank_address = 0)""") == 0

 
 
 @pytest.mark.parametrize("threads", [1, 15])
 
 
 @pytest.mark.parametrize("threads", [1, 15])

-def test_index_partial_without_30(test_db, threads):
+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()
 
     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()
+    assert test_db.placex_unindexed() == 31
+    assert test_db.osmline_unindexed() == 1

 
 

-    idx = indexer.Indexer('dbname=test_nominatim_python_unittest', threads)
+    idx = indexer.Indexer('dbname=test_nominatim_python_unittest',
+                          test_tokenizer, threads)

     idx.index_by_rank(4, 15)
 
     idx.index_by_rank(4, 15)
 

-    assert 19 == test_db.placex_unindexed()
-    assert 1 == test_db.osmline_unindexed()
+    assert test_db.placex_unindexed() == 19
+    assert test_db.osmline_unindexed() == 1

 
 

-    assert 0 == test_db.scalar("""
+    assert test_db.scalar("""

                     SELECT count(*) FROM placex
                     SELECT count(*) FROM placex

-                      WHERE indexed_status = 0 AND not rank_address between 4 and 15""")
+                      WHERE indexed_status = 0 AND not rank_address between 4 and 15""") == 0

 
 
 @pytest.mark.parametrize("threads", [1, 15])
 
 
 @pytest.mark.parametrize("threads", [1, 15])

-def test_index_partial_with_30(test_db, threads):
+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()
 
     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()
+    assert test_db.placex_unindexed() == 31
+    assert test_db.osmline_unindexed() == 1

 
 

-    idx = indexer.Indexer('dbname=test_nominatim_python_unittest', threads)
+    idx = indexer.Indexer('dbname=test_nominatim_python_unittest', test_tokenizer, threads)

     idx.index_by_rank(28, 30)
 
     idx.index_by_rank(28, 30)
 

-    assert 27 == test_db.placex_unindexed()
-    assert 0 == test_db.osmline_unindexed()
+    assert test_db.placex_unindexed() == 27
+    assert test_db.osmline_unindexed() == 0

 
 

-    assert 0 == test_db.scalar("""
+    assert test_db.scalar("""

                     SELECT count(*) FROM placex
                     SELECT count(*) FROM placex

-                      WHERE indexed_status = 0 AND rank_address between 1 and 27""")
+                      WHERE indexed_status = 0 AND rank_address between 1 and 27""") == 0

 
 @pytest.mark.parametrize("threads", [1, 15])
 
 @pytest.mark.parametrize("threads", [1, 15])

-def test_index_boundaries(test_db, threads):
+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()
 
     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()
+    assert test_db.placex_unindexed() == 37
+    assert test_db.osmline_unindexed() == 1

 
 

-    idx = indexer.Indexer('dbname=test_nominatim_python_unittest', threads)
+    idx = indexer.Indexer('dbname=test_nominatim_python_unittest', test_tokenizer, threads)

     idx.index_boundaries(0, 30)
 
     idx.index_boundaries(0, 30)
 

-    assert 31 == test_db.placex_unindexed()
-    assert 1 == test_db.osmline_unindexed()
+    assert test_db.placex_unindexed() == 31
+    assert test_db.osmline_unindexed() == 1

 
 

-    assert 0 == test_db.scalar("""
+    assert test_db.scalar("""

                     SELECT count(*) FROM placex
                     SELECT count(*) FROM placex

-                      WHERE indexed_status = 0 AND class != 'boundary'""")
+                      WHERE indexed_status = 0 AND class != 'boundary'""") == 0

 
 
 @pytest.mark.parametrize("threads", [1, 15])
 
 
 @pytest.mark.parametrize("threads", [1, 15])

-def test_index_postcodes(test_db, threads):
+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)
 
     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', threads)
+    idx = indexer.Indexer('dbname=test_nominatim_python_unittest', test_tokenizer, threads)

     idx.index_postcodes()
 
     idx.index_postcodes()
 

-    assert 0 == test_db.scalar("""SELECT count(*) FROM location_postcode
-                                  WHERE indexed_status != 0""")
+    assert test_db.scalar("""SELECT count(*) FROM location_postcode
+                                  WHERE indexed_status != 0""") == 0

 
 
 @pytest.mark.parametrize("analyse", [True, False])
 
 
 @pytest.mark.parametrize("analyse", [True, False])

-def test_index_full(test_db, analyse):
+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):
     for rank in range(4, 10):
         test_db.add_admin(rank_address=rank, rank_search=rank)
     for rank in range(31):


@@ -226,23 +267,23 @@ def test_index_full(test_db, analyse):
     for postcode in range(1000):
         test_db.add_postcode('de', postcode)
 
     for postcode in range(1000):
         test_db.add_postcode('de', postcode)
 

-    idx = indexer.Indexer('dbname=test_nominatim_python_unittest', 4)
+    idx = indexer.Indexer('dbname=test_nominatim_python_unittest', test_tokenizer, 4)

     idx.index_full(analyse=analyse)
 
     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""")
+    assert test_db.placex_unindexed() == 0
+    assert test_db.osmline_unindexed() == 0
+    assert test_db.scalar("""SELECT count(*) FROM location_postcode
+                             WHERE indexed_status != 0""") == 0

 
 
 @pytest.mark.parametrize("threads", [1, 15])
 
 
 @pytest.mark.parametrize("threads", [1, 15])

-def test_index_reopen_connection(test_db, threads, monkeypatch):
+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)
 
     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', threads)
+    idx = indexer.Indexer('dbname=test_nominatim_python_unittest', test_tokenizer, threads)

     idx.index_by_rank(28, 30)
 
     idx.index_by_rank(28, 30)
 

-    assert 0 == test_db.placex_unindexed()
+    assert test_db.placex_unindexed() == 0