X-Git-Url: https://git.openstreetmap.org./nominatim.git/blobdiff_plain/8b1a509442a3fa051146f82b8293126916ad8617..27f7c7fd88e588c48c006b459d9912c23b4e13c0:/nominatim/indexer/runners.py diff --git a/nominatim/indexer/runners.py b/nominatim/indexer/runners.py index aa607faa..ac7a0015 100644 --- a/nominatim/indexer/runners.py +++ b/nominatim/indexer/runners.py @@ -1,17 +1,30 @@ +# 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. """ Mix-ins that provide the actual commands for the indexer for various indexing tasks. """ import functools -import psycopg2.extras +from psycopg2 import sql as pysql + +from nominatim.indexer.place_info import PlaceInfo # pylint: disable=C0111 +def _mk_valuelist(template, num): + return pysql.SQL(',').join([pysql.SQL(template)] * num) + + class AbstractPlacexRunner: """ Returns SQL commands for indexing of the placex table. """ - SELECT_SQL = 'SELECT place_id FROM placex' + SELECT_SQL = pysql.SQL('SELECT place_id FROM placex ') + UPDATE_LINE = "(%s, %s::hstore, %s::hstore, %s::int, %s::jsonb)" def __init__(self, rank, analyzer): self.rank = rank @@ -21,25 +34,29 @@ class AbstractPlacexRunner: @staticmethod @functools.lru_cache(maxsize=1) def _index_sql(num_places): - return """ UPDATE placex - SET indexed_status = 0, address = v.addr, token_info = v.ti - FROM (VALUES {}) as v(id, addr, ti) - WHERE place_id = v.id - """.format(','.join(["(%s, %s::hstore, %s::jsonb)"] * num_places)) + return pysql.SQL( + """ UPDATE placex + SET indexed_status = 0, address = v.addr, token_info = v.ti, + name = v.name, linked_place_id = v.linked_place_id + FROM (VALUES {}) as v(id, name, addr, linked_place_id, ti) + WHERE place_id = v.id + """).format(_mk_valuelist(AbstractPlacexRunner.UPDATE_LINE, num_places)) @staticmethod def get_place_details(worker, ids): - worker.perform("""SELECT place_id, (placex_prepare_update(placex)).* - FROM placex WHERE place_id IN %s""", + worker.perform("""SELECT place_id, extra.* + FROM placex, LATERAL placex_indexing_prepare(placex) as extra + WHERE place_id IN %s""", (tuple((p[0] for p in ids)), )) def index_places(self, worker, places): values = [] for place in places: - values.extend((place[x] for x in ('place_id', 'address'))) - values.append(psycopg2.extras.Json(self.analyzer.process_place(place))) + for field in ('place_id', 'name', 'address', 'linked_place_id'): + values.append(place[field]) + values.append(PlaceInfo(place).analyze(self.analyzer)) worker.perform(self._index_sql(len(places)), values) @@ -52,14 +69,15 @@ class RankRunner(AbstractPlacexRunner): return "rank {}".format(self.rank) def sql_count_objects(self): - return """SELECT count(*) FROM placex - WHERE rank_address = {} and indexed_status > 0 - """.format(self.rank) + return pysql.SQL("""SELECT count(*) FROM placex + WHERE rank_address = {} and indexed_status > 0 + """).format(pysql.Literal(self.rank)) def sql_get_objects(self): - return """{} WHERE indexed_status > 0 and rank_address = {} - ORDER BY geometry_sector - """.format(self.SELECT_SQL, self.rank) + return self.SELECT_SQL + pysql.SQL( + """WHERE indexed_status > 0 and rank_address = {} + ORDER BY geometry_sector + """).format(pysql.Literal(self.rank)) class BoundaryRunner(AbstractPlacexRunner): @@ -71,17 +89,18 @@ class BoundaryRunner(AbstractPlacexRunner): return "boundaries rank {}".format(self.rank) def sql_count_objects(self): - return """SELECT count(*) FROM placex - WHERE indexed_status > 0 - AND rank_search = {} - AND class = 'boundary' and type = 'administrative' - """.format(self.rank) + return pysql.SQL("""SELECT count(*) FROM placex + WHERE indexed_status > 0 + AND rank_search = {} + AND class = 'boundary' and type = 'administrative' + """).format(pysql.Literal(self.rank)) def sql_get_objects(self): - return """{} WHERE indexed_status > 0 and rank_search = {} - and class = 'boundary' and type = 'administrative' - ORDER BY partition, admin_level - """.format(self.SELECT_SQL, self.rank) + return self.SELECT_SQL + pysql.SQL( + """WHERE indexed_status > 0 and rank_search = {} + and class = 'boundary' and type = 'administrative' + ORDER BY partition, admin_level + """).format(pysql.Literal(self.rank)) class InterpolationRunner: @@ -120,18 +139,18 @@ class InterpolationRunner: @staticmethod @functools.lru_cache(maxsize=1) def _index_sql(num_places): - return """ UPDATE location_property_osmline - SET indexed_status = 0, address = v.addr, token_info = v.ti - FROM (VALUES {}) as v(id, addr, ti) - WHERE place_id = v.id - """.format(','.join(["(%s, %s::hstore, %s::jsonb)"] * num_places)) + return pysql.SQL("""UPDATE location_property_osmline + SET indexed_status = 0, address = v.addr, token_info = v.ti + FROM (VALUES {}) as v(id, addr, ti) + WHERE place_id = v.id + """).format(_mk_valuelist("(%s, %s::hstore, %s::jsonb)", num_places)) def index_places(self, worker, places): values = [] for place in places: values.extend((place[x] for x in ('place_id', 'address'))) - values.append(psycopg2.extras.Json(self.analyzer.process_place(place))) + values.append(PlaceInfo(place).analyze(self.analyzer)) worker.perform(self._index_sql(len(places)), values) @@ -157,6 +176,6 @@ class PostcodeRunner: @staticmethod def index_places(worker, ids): - worker.perform(""" UPDATE location_postcode SET indexed_status = 0 - WHERE place_id IN ({}) - """.format(','.join((str(i[0]) for i in ids)))) + worker.perform(pysql.SQL("""UPDATE location_postcode SET indexed_status = 0 + WHERE place_id IN ({})""") + .format(pysql.SQL(',').join((pysql.Literal(i[0]) for i in ids))))