X-Git-Url: https://git.openstreetmap.org./nominatim.git/blobdiff_plain/810056349f0dde9e0ddeb6655cbf1aa48399b3a5..fdd20a76bdba7b1115361e337aff661ed830b11f:/nominatim/tools/migration.py

diff --git a/nominatim/tools/migration.py b/nominatim/tools/migration.py
index 87febacc..76726e8c 100644
--- a/nominatim/tools/migration.py
+++ b/nominatim/tools/migration.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.
 """
 Functions for database migration to newer software versions.
 """
@@ -26,7 +32,7 @@ def migrate(config, paths):
 
         if db_version_str is not None:
             parts = db_version_str.split('.')
-            db_version = tuple([int(x) for x in parts[:2] + parts[2].split('-')])
+            db_version = tuple(int(x) for x in parts[:2] + parts[2].split('-'))
 
             if db_version == NOMINATIM_VERSION:
                 LOG.warning("Database already at latest version (%s)", db_version_str)
@@ -96,6 +102,7 @@ def _migration(major, minor, patch=0, dbpatch=0):
     """
     def decorator(func):
         _MIGRATION_FUNCTIONS.append(((major, minor, patch, dbpatch), func))
+        return func
 
     return decorator
 
@@ -195,7 +202,7 @@ def install_legacy_tokenizer(conn, config, **_):
 
 
 @_migration(4, 0, 99, 0)
-def create_tiger_housenumber_index(conn, _, **_):
+def create_tiger_housenumber_index(conn, **_):
     """ Create idx_location_property_tiger_parent_place_id with included
         house number.
 
@@ -209,3 +216,94 @@ def create_tiger_housenumber_index(conn, _, **_):
                             ON location_property_tiger
                             USING btree(parent_place_id)
                             INCLUDE (startnumber, endnumber) """)
+
+
+@_migration(4, 0, 99, 1)
+def create_interpolation_index_on_place(conn, **_):
+    """ Create idx_place_interpolations for lookup of interpolation lines
+        on updates.
+    """
+    with conn.cursor() as cur:
+        cur.execute("""CREATE INDEX IF NOT EXISTS idx_place_interpolations
+                       ON place USING gist(geometry)
+                       WHERE osm_type = 'W' and address ? 'interpolation'""")
+
+
+@_migration(4, 0, 99, 2)
+def add_step_column_for_interpolation(conn, **_):
+    """ Add a new column 'step' to the interpolations table.
+
+        Also convers the data into the stricter format which requires that
+        startnumbers comply with the odd/even requirements.
+    """
+    if conn.table_has_column('location_property_osmline', 'step'):
+        return
+
+    with conn.cursor() as cur:
+        # Mark invalid all interpolations with no intermediate numbers.
+        cur.execute("""UPDATE location_property_osmline SET startnumber = null
+                       WHERE endnumber - startnumber <= 1 """)
+        # Align the start numbers where odd/even does not match.
+        cur.execute("""UPDATE location_property_osmline
+                       SET startnumber = startnumber + 1,
+                           linegeo = ST_LineSubString(linegeo,
+                                                      1.0 / (endnumber - startnumber)::float,
+                                                      1)
+                       WHERE (interpolationtype = 'odd' and startnumber % 2 = 0)
+                              or (interpolationtype = 'even' and startnumber % 2 = 1)
+                    """)
+        # Mark invalid odd/even interpolations with no intermediate numbers.
+        cur.execute("""UPDATE location_property_osmline SET startnumber = null
+                       WHERE interpolationtype in ('odd', 'even')
+                             and endnumber - startnumber = 2""")
+        # Finally add the new column and populate it.
+        cur.execute("ALTER TABLE location_property_osmline ADD COLUMN step SMALLINT")
+        cur.execute("""UPDATE location_property_osmline
+                         SET step = CASE WHEN interpolationtype = 'all'
+                                         THEN 1 ELSE 2 END
+                    """)
+
+
+@_migration(4, 0, 99, 3)
+def add_step_column_for_tiger(conn, **_):
+    """ Add a new column 'step' to the tiger data table.
+    """
+    if conn.table_has_column('location_property_tiger', 'step'):
+        return
+
+    with conn.cursor() as cur:
+        cur.execute("ALTER TABLE location_property_tiger ADD COLUMN step SMALLINT")
+        cur.execute("""UPDATE location_property_tiger
+                         SET step = CASE WHEN interpolationtype = 'all'
+                                         THEN 1 ELSE 2 END
+                    """)
+
+
+@_migration(4, 0, 99, 4)
+def add_derived_name_column_for_country_names(conn, **_):
+    """ Add a new column 'derived_name' which in the future takes the
+        country names as imported from OSM data.
+    """
+    if not conn.table_has_column('country_name', 'derived_name'):
+        with conn.cursor() as cur:
+            cur.execute("ALTER TABLE country_name ADD COLUMN derived_name public.HSTORE")
+
+
+@_migration(4, 0, 99, 5)
+def mark_internal_country_names(conn, config, **_):
+    """ Names from the country table should be marked as internal to prevent
+        them from being deleted. Only necessary for ICU tokenizer.
+    """
+    import psycopg2.extras # pylint: disable=import-outside-toplevel
+
+    tokenizer = tokenizer_factory.get_tokenizer_for_db(config)
+    with tokenizer.name_analyzer() as analyzer:
+        with conn.cursor() as cur:
+            psycopg2.extras.register_hstore(cur)
+            cur.execute("SELECT country_code, name FROM country_name")
+
+            for country_code, names in cur:
+                if not names:
+                    names = {}
+                names['countrycode'] = country_code
+                analyzer.add_country_names(country_code, names)