2 Implementation of the 'index' subcommand.
6 from nominatim.db import status
7 from nominatim.db.connection import connect
9 # Do not repeat documentation of subcommand classes.
10 # pylint: disable=C0111
11 # Using non-top-level imports to avoid eventually unused imports.
12 # pylint: disable=E0012,C0415
17 Reindex all new and modified data.
19 Indexing is the process of computing the address and search terms for
20 the places in the database. Every time data is added or changed, indexing
21 needs to be run. Imports and replication updates automatically take care
22 of indexing. For other cases, this function allows to run indexing manually.
27 group = parser.add_argument_group('Filter arguments')
28 group.add_argument('--boundaries-only', action='store_true',
29 help="""Index only administrative boundaries.""")
30 group.add_argument('--no-boundaries', action='store_true',
31 help="""Index everything except administrative boundaries.""")
32 group.add_argument('--minrank', '-r', type=int, metavar='RANK', default=0,
33 help='Minimum/starting rank')
34 group.add_argument('--maxrank', '-R', type=int, metavar='RANK', default=30,
35 help='Maximum/finishing rank')
39 from ..indexer.indexer import Indexer
40 from ..tokenizer import factory as tokenizer_factory
42 tokenizer = tokenizer_factory.get_tokenizer_for_db(args.config)
44 indexer = Indexer(args.config.get_libpq_dsn(), tokenizer,
45 args.threads or psutil.cpu_count() or 1)
47 if not args.no_boundaries:
48 indexer.index_boundaries(args.minrank, args.maxrank)
49 if not args.boundaries_only:
50 indexer.index_by_rank(args.minrank, args.maxrank)
52 if not args.no_boundaries and not args.boundaries_only \
53 and args.minrank == 0 and args.maxrank == 30:
54 with connect(args.config.get_libpq_dsn()) as conn:
55 status.set_indexed(conn, True)