Merge remote-tracking branch 'upstream/master'

[nominatim.git] / test / bdd / steps / steps_db_ops.py

diff --git a/test/bdd/steps/steps_db_ops.py b/test/bdd/steps/steps_db_ops.py
index b4f0d8532ac39984746aae2911e66ab8564232fd..8df5d6170e741e4807b8ddfac541c72d6437b89e 100644 (file)
--- a/test/bdd/steps/steps_db_ops.py
+++ b/test/bdd/steps/steps_db_ops.py
@@ -1,3 +1,9 @@
+# SPDX-License-Identifier: GPL-2.0-only
+#
+# This file is part of Nominatim. (https://nominatim.org)
+#
+# Copyright (C) 2022 by the Nominatim developer community.
+# For a full list of authors see the git log.
 import logging
 from itertools import chain
 
@@ -87,29 +93,8 @@ def add_data_to_planet_ways(context):
 def import_and_index_data_from_place_table(context):
     """ Import data previously set up in the place table.
     """
-    nctx = context.nominatim
-
-    tokenizer = tokenizer_factory.create_tokenizer(nctx.get_test_config())
-    context.nominatim.copy_from_place(context.db)
-
-    # XXX use tool function as soon as it is ported
-    with context.db.cursor() as cur:
-        with (context.nominatim.src_dir / 'lib-sql' / 'postcode_tables.sql').open('r') as fd:
-            cur.execute(fd.read())
-        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""")
-
-    # Call directly as the refresh function does not include postcodes.
-    indexer.LOG.setLevel(logging.ERROR)
-    indexer.Indexer(context.nominatim.get_libpq_dsn(), tokenizer, 1).index_full(analyse=False)
+    context.nominatim.run_nominatim('import', '--continue', 'load-data',
+                                              '--index-noanalyse', '-q')
 
     check_database_integrity(context)
 
@@ -262,24 +247,40 @@ def check_location_postcode(context):
         for row in context.table:
             db_row = results.get((row['country'],row['postcode']))
             assert db_row is not None, \
-                "Missing row for country '{r['country']}' postcode '{r['postcode']}'.".format(r=row)
+                f"Missing row for country '{row['country']}' postcode '{row['postcode']}'."
 
             db_row.assert_row(row, ('country', 'postcode'))
 
-@then("word contains(?P<exclude> not)?")
-def check_word_table(context, exclude):
-    """ Check the contents of the word table. Each row represents a table row
-        and all data must match. Data not present in the expected table, may
-        be arbitry. The rows are identified via all given columns.
+@then("there are(?P<exclude> no)? word tokens for postcodes (?P<postcodes>.*)")
+def check_word_table_for_postcodes(context, exclude, postcodes):
+    """ Check that the tokenizer produces postcode tokens for the given
+        postcodes. The postcodes are a comma-separated list of postcodes.
+        Whitespace matters.
     """
+    nctx = context.nominatim
+    tokenizer = tokenizer_factory.get_tokenizer_for_db(nctx.get_test_config())
+    with tokenizer.name_analyzer() as ana:
+        plist = [ana.normalize_postcode(p) for p in postcodes.split(',')]
+
+    plist.sort()
+
     with context.db.cursor(cursor_factory=psycopg2.extras.DictCursor) as cur:
-        for row in context.table:
-            wheres = ' AND '.join(["{} = %s".format(h) for h in row.headings])
-            cur.execute("SELECT * from word WHERE " + wheres, list(row.cells))
-            if exclude:
-                assert cur.rowcount == 0, "Row still in word table: %s" % '/'.join(values)
-            else:
-                assert cur.rowcount > 0, "Row not in word table: %s" % '/'.join(values)
+        if nctx.tokenizer == 'icu':
+            cur.execute("SELECT word FROM word WHERE type = 'P' and word = any(%s)",
+                        (plist,))
+        else:
+            cur.execute("""SELECT word FROM word WHERE word = any(%s)
+                             and class = 'place' and type = 'postcode'""",
+                        (plist,))
+
+        found = [row[0] for row in cur]
+        assert len(found) == len(set(found)), f"Duplicate rows for postcodes: {found}"
+
+    if exclude:
+        assert len(found) == 0, f"Unexpected postcodes: {found}"
+    else:
+        assert set(found) == set(plist), \
+        f"Missing postcodes {set(plist) - set(found)}. Found: {found}"
 
 @then("place_addressline contains")
 def check_place_addressline(context):
@@ -311,12 +312,13 @@ def check_place_addressline_exclude(context):
     with context.db.cursor(cursor_factory=psycopg2.extras.DictCursor) as cur:
         for row in context.table:
             pid = NominatimID(row['object']).get_place_id(cur)
-            apid = NominatimID(row['address']).get_place_id(cur)
-            cur.execute(""" SELECT * FROM place_addressline
-                            WHERE place_id = %s AND address_place_id = %s""",
-                        (pid, apid))
-            assert cur.rowcount == 0, \
-                "Row found for place %s and address %s" % (row['object'], row['address'])
+            apid = NominatimID(row['address']).get_place_id(cur, allow_empty=True)
+            if apid is not None:
+                cur.execute(""" SELECT * FROM place_addressline
+                                WHERE place_id = %s AND address_place_id = %s""",
+                            (pid, apid))
+                assert cur.rowcount == 0, \
+                    "Row found for place %s and address %s" % (row['object'], row['address'])
 
 @then("W(?P<oid>\d+) expands to(?P<neg> no)? interpolation")
 def check_location_property_osmline(context, oid, neg):