print nice summary after updates

[nominatim.git] / nominatim / cli.py

diff --git a/nominatim/cli.py b/nominatim/cli.py
index 65ea90bbc3534b4421558e883ecaf31796e5b0b5..4873308deb19cac7da2144fc231b9c39e43561c4 100644 (file)
--- a/nominatim/cli.py
+++ b/nominatim/cli.py
@@ -2,16 +2,21 @@
 Command-line interface to the Nominatim functions for import, update,
 database administration and querying.
 """
 Command-line interface to the Nominatim functions for import, update,
 database administration and querying.
 """

-import sys
+import datetime as dt

 import os
 import os

+import sys
+import time

 import argparse
 import logging
 from pathlib import Path
 
 from .config import Configuration
 import argparse
 import logging
 from pathlib import Path
 
 from .config import Configuration

-from .admin.exec_utils import run_legacy_script
+from .tools.exec_utils import run_legacy_script, run_api_script
+from .db.connection import connect
+from .db import status
+from .errors import UsageError

 
 

-from .indexer.indexer import Indexer
+LOG = logging.getLogger()

 
 def _num_system_cpus():
     try:
 
 def _num_system_cpus():
     try:


@@ -68,12 +73,13 @@ class CommandlineParser:
         """ Parse the command line arguments of the program and execute the
             appropriate subcommand.
         """
         """ Parse the command line arguments of the program and execute the
             appropriate subcommand.
         """

-        args = self.parser.parse_args()
+        args = self.parser.parse_args(args=kwargs.get('cli_args'))

 
         if args.subcommand is None:
 
         if args.subcommand is None:

-            return self.parser.print_help()
+            self.parser.print_help()
+            return 1

 
 

-        for arg in ('module_dir', 'osm2pgsql_path', 'phplib_dir', 'data_dir'):
+        for arg in ('module_dir', 'osm2pgsql_path', 'phplib_dir', 'data_dir', 'phpcgi_path'):

             setattr(args, arg, Path(kwargs[arg]))
         args.project_dir = Path(args.project_dir)
 
             setattr(args, arg, Path(kwargs[arg]))
         args.project_dir = Path(args.project_dir)
 


@@ -84,7 +90,27 @@ class CommandlineParser:
 
         args.config = Configuration(args.project_dir, args.data_dir / 'settings')
 
 
         args.config = Configuration(args.project_dir, args.data_dir / 'settings')
 

-        return args.command.run(args)
+        try:
+            return args.command.run(args)
+        except UsageError as exception:
+            log = logging.getLogger()
+            if log.isEnabledFor(logging.DEBUG):
+                raise # use Python's exception printing
+            log.fatal('FATAL: %s', exception)
+
+        # If we get here, then execution has failed in some way.
+        return 1
+
+
+def _osm2pgsql_options_from_args(args, default_cache, default_threads):
+    """ Set up the stanadrd osm2pgsql from the command line arguments.
+    """
+    return dict(osm2pgsql=args.osm2pgsql_path,
+                osm2pgsql_cache=args.osm2pgsql_cache or default_cache,
+                osm2pgsql_style=args.config.get_import_style_file(),
+                threads=args.threads or default_threads,
+                dsn=args.config.get_libpq_dsn(),
+                flatnode_file=args.config.FLATNODE_FILE)

 
 ##### Subcommand classes
 #
 
 ##### Subcommand classes
 #


@@ -97,6 +123,8 @@ class CommandlineParser:
 #
 # No need to document the functions each time.
 # pylint: disable=C0111
 #
 # No need to document the functions each time.
 # pylint: disable=C0111

+# Using non-top-level imports to make pyosmium optional for replication only.
+# pylint: disable=E0012,C0415

 
 
 class SetupAll:
 
 
 class SetupAll:


@@ -196,13 +224,12 @@ class SetupSpecialPhrases:
                            help='Pull special phrases from the OSM wiki.')
         group = parser.add_argument_group('Output arguments')
         group.add_argument('-o', '--output', default='-',
                            help='Pull special phrases from the OSM wiki.')
         group = parser.add_argument_group('Output arguments')
         group.add_argument('-o', '--output', default='-',

-                           type=argparse.FileType('w', encoding='UTF-8'),

                            help="""File to write the preprocessed phrases to.
                                    If omitted, it will be written to stdout.""")
 
     @staticmethod
     def run(args):
                            help="""File to write the preprocessed phrases to.
                                    If omitted, it will be written to stdout.""")
 
     @staticmethod
     def run(args):

-        if args.output.name != '<stdout>':
+        if args.output != '-':

             raise NotImplementedError('Only output to stdout is currently implemented.')
         return run_legacy_script('specialphrases.php', '--wiki-import', nominatim_env=args)
 
             raise NotImplementedError('Only output to stdout is currently implemented.')
         return run_legacy_script('specialphrases.php', '--wiki-import', nominatim_env=args)
 


@@ -230,26 +257,123 @@ class UpdateReplication:
         group.add_argument('--no-index', action='store_false', dest='do_index',
                            help="""Do not index the new data. Only applicable
                                    together with --once""")
         group.add_argument('--no-index', action='store_false', dest='do_index',
                            help="""Do not index the new data. Only applicable
                                    together with --once""")

+        group.add_argument('--osm2pgsql-cache', metavar='SIZE', type=int,
+                           help='Size of cache to be used by osm2pgsql (in MB)')
+
+    @staticmethod
+    def _init_replication(args):
+        from .tools import replication, refresh
+
+        LOG.warning("Initialising replication updates")
+        conn = connect(args.config.get_libpq_dsn())
+        replication.init_replication(conn, base_url=args.config.REPLICATION_URL)
+        if args.update_functions:
+            LOG.warning("Create functions")
+            refresh.create_functions(conn, args.config, args.data_dir,
+                                     True, False)
+        conn.close()
+        return 0
+
+
+    @staticmethod
+    def _check_for_updates(args):
+        from .tools import replication
+
+        conn = connect(args.config.get_libpq_dsn())
+        ret = replication.check_for_updates(conn, base_url=args.config.REPLICATION_URL)
+        conn.close()
+        return ret
+
+    @staticmethod
+    def _report_update(batchdate, start_import, start_index):
+        def round_time(delta):
+            return dt.timedelta(seconds=int(delta.total_seconds()))
+
+        end = dt.datetime.now(dt.timezone.utc)
+        LOG.warning("Update completed. Import: %s. %sTotal: %s. Remaining backlog: %s.",
+                    round_time((start_index or end) - start_import),
+                    "Indexing: {} ".format(round_time(end - start_index))
+                    if start_index else '',
+                    round_time(end - start_import),
+                    round_time(end - batchdate))
+
+    @staticmethod
+    def _update(args):
+        from .tools import replication
+        from .indexer.indexer import Indexer
+
+        params = _osm2pgsql_options_from_args(args, 2000, 1)
+        params.update(base_url=args.config.REPLICATION_URL,
+                      update_interval=args.config.get_int('REPLICATION_UPDATE_INTERVAL'),
+                      import_file=args.project_dir / 'osmosischange.osc',
+                      max_diff_size=args.config.get_int('REPLICATION_MAX_DIFF'),
+                      indexed_only=not args.once)
+
+        # Sanity check to not overwhelm the Geofabrik servers.
+        if 'download.geofabrik.de'in params['base_url']\
+           and params['update_interval'] < 86400:
+            LOG.fatal("Update interval too low for download.geofabrik.de.\n"
+                      "Please check install documentation "
+                      "(https://nominatim.org/release-docs/latest/admin/Import-and-Update#"
+                      "setting-up-the-update-process).")
+            raise UsageError("Invalid replication update interval setting.")
+
+        if not args.once:
+            if not args.do_index:
+                LOG.fatal("Indexing cannot be disabled when running updates continuously.")
+                raise UsageError("Bad argument '--no-index'.")
+            recheck_interval = args.config.get_int('REPLICATION_RECHECK_INTERVAL')
+
+        while True:
+            conn = connect(args.config.get_libpq_dsn())
+            start = dt.datetime.now(dt.timezone.utc)
+            state = replication.update(conn, params)
+            status.log_status(conn, start, 'import')
+            batchdate, _, _ = status.get_status(conn)
+            conn.close()
+
+            if state is not replication.UpdateState.NO_CHANGES and args.do_index:
+                index_start = dt.datetime.now(dt.timezone.utc)
+                indexer = Indexer(args.config.get_libpq_dsn(),
+                                  args.threads or 1)
+                indexer.index_boundaries(0, 30)
+                indexer.index_by_rank(0, 30)
+
+                conn = connect(args.config.get_libpq_dsn())
+                status.set_indexed(conn, True)
+                status.log_status(conn, index_start, 'index')
+                conn.close()
+            else:
+                index_start = None
+
+            if LOG.isEnabledFor(logging.WARNING):
+                UpdateReplication._report_update(batchdate, start, index_start)
+
+            if args.once:
+                break
+
+            if state is replication.UpdateState.NO_CHANGES:
+                LOG.warning("No new changes. Sleeping for %d sec.", recheck_interval)
+                time.sleep(recheck_interval)
+
+        return state.value

 
     @staticmethod
     def run(args):
 
     @staticmethod
     def run(args):

-        params = ['update.php']
+        try:
+            import osmium # pylint: disable=W0611
+        except ModuleNotFoundError:
+            LOG.fatal("pyosmium not installed. Replication functions not available.\n"
+                      "To install pyosmium via pip: pip3 install osmium")
+            return 1
+

         if args.init:
         if args.init:

-            params.append('--init-updates')
-            if not args.update_functions:
-                params.append('--no-update-functions')
-        elif args.check_for_updates:
-            params.append('--check-for-updates')
-        else:
-            if args.once:
-                params.append('--import-osmosis')
-            else:
-                params.append('--import-osmosis-all')
-            if not args.do_index:
-                params.append('--no-index')
+            return UpdateReplication._init_replication(args)

 
 

-        return run_legacy_script(*params, nominatim_env=args)
+        if args.check_for_updates:
+            return UpdateReplication._check_for_updates(args)

 
 

+        return UpdateReplication._update(args)

 
 class UpdateAddData:
     """\
 
 class UpdateAddData:
     """\


@@ -320,6 +444,8 @@ class UpdateIndex:
 
     @staticmethod
     def run(args):
 
     @staticmethod
     def run(args):

+        from .indexer.indexer import Indexer
+

         indexer = Indexer(args.config.get_libpq_dsn(),
                           args.threads or _num_system_cpus() or 1)
 
         indexer = Indexer(args.config.get_libpq_dsn(),
                           args.threads or _num_system_cpus() or 1)
 


@@ -328,8 +454,11 @@ class UpdateIndex:
         if not args.boundaries_only:
             indexer.index_by_rank(args.minrank, args.maxrank)
 
         if not args.boundaries_only:
             indexer.index_by_rank(args.minrank, args.maxrank)
 

-        if not args.no_boundaries and not args.boundaries_only:
-            indexer.update_status_table()
+        if not args.no_boundaries and not args.boundaries_only \
+           and args.minrank == 0 and args.maxrank == 30:
+            conn = connect(args.config.get_libpq_dsn())
+            status.set_indexed(conn, True)
+            conn.close()

 
         return 0
 
 
         return 0
 


@@ -366,22 +495,34 @@ class UpdateRefresh:
 
     @staticmethod
     def run(args):
 
     @staticmethod
     def run(args):

+        from .tools import refresh
+

         if args.postcodes:
         if args.postcodes:

-            run_legacy_script('update.php', '--calculate-postcodes',
-                              nominatim_env=args, throw_on_fail=True)
+            LOG.warning("Update postcodes centroid")
+            conn = connect(args.config.get_libpq_dsn())
+            refresh.update_postcodes(conn, args.data_dir)
+            conn.close()
+

         if args.word_counts:
         if args.word_counts:

-            run_legacy_script('update.php', '--recompute-word-counts',
-                              nominatim_env=args, throw_on_fail=True)
+            LOG.warning('Recompute frequency of full-word search terms')
+            conn = connect(args.config.get_libpq_dsn())
+            refresh.recompute_word_counts(conn, args.data_dir)
+            conn.close()
+

         if args.address_levels:
         if args.address_levels:

-            run_legacy_script('update.php', '--update-address-levels',
-                              nominatim_env=args, throw_on_fail=True)
+            cfg = Path(args.config.ADDRESS_LEVEL_CONFIG)
+            LOG.warning('Updating address levels from %s', cfg)
+            conn = connect(args.config.get_libpq_dsn())
+            refresh.load_address_levels_from_file(conn, cfg)
+            conn.close()
+

         if args.functions:
         if args.functions:

-            params = ['setup.php', '--create-functions', '--create-partition-functions']
-            if args.diffs:
-                params.append('--enable-diff-updates')
-            if args.enable_debug_statements:
-                params.append('--enable-debug-statements')
-            run_legacy_script(*params, nominatim_env=args, throw_on_fail=True)
+            LOG.warning('Create functions')
+            conn = connect(args.config.get_libpq_dsn())
+            refresh.create_functions(conn, args.config, args.data_dir,
+                                     args.diffs, args.enable_debug_statements)
+            conn.close()
+

         if args.wiki_data:
             run_legacy_script('setup.php', '--import-wikipedia-articles',
                               nominatim_env=args, throw_on_fail=True)
         if args.wiki_data:
             run_legacy_script('setup.php', '--import-wikipedia-articles',
                               nominatim_env=args, throw_on_fail=True)


@@ -393,6 +534,8 @@ class UpdateRefresh:
             run_legacy_script('setup.php', '--setup-website',
                               nominatim_env=args, throw_on_fail=True)
 
             run_legacy_script('setup.php', '--setup-website',
                               nominatim_env=args, throw_on_fail=True)
 

+        return 0
+

 
 class AdminCheckDatabase:
     """\
 
 class AdminCheckDatabase:
     """\


@@ -435,7 +578,7 @@ class AdminWarm:
 
 class QueryExport:
     """\
 
 class QueryExport:
     """\

-    Export addresses as CSV file from a Nominatim database.
+    Export addresses as CSV file from the database.

     """
 
     @staticmethod
     """
 
     @staticmethod


@@ -488,17 +631,245 @@ class QueryExport:
 
         return run_legacy_script(*params, nominatim_env=args)
 
 
         return run_legacy_script(*params, nominatim_env=args)
 

-class QueryTodo:
+STRUCTURED_QUERY = (
+    ('street', 'housenumber and street'),
+    ('city', 'city, town or village'),
+    ('county', 'county'),
+    ('state', 'state'),
+    ('country', 'country'),
+    ('postalcode', 'postcode')
+)
+
+EXTRADATA_PARAMS = (
+    ('addressdetails', 'Include a breakdown of the address into elements.'),
+    ('extratags', """Include additional information if available
+                     (e.g. wikipedia link, opening hours)."""),
+    ('namedetails', 'Include a list of alternative names.')
+)
+
+DETAILS_SWITCHES = (
+    ('addressdetails', 'Include a breakdown of the address into elements.'),
+    ('keywords', 'Include a list of name keywords and address keywords.'),
+    ('linkedplaces', 'Include a details of places that are linked with this one.'),
+    ('hierarchy', 'Include details of places lower in the address hierarchy.'),
+    ('group_hierarchy', 'Group the places by type.'),
+    ('polygon_geojson', 'Include geometry of result.')
+)
+
+def _add_api_output_arguments(parser):
+    group = parser.add_argument_group('Output arguments')
+    group.add_argument('--format', default='jsonv2',
+                       choices=['xml', 'json', 'jsonv2', 'geojson', 'geocodejson'],
+                       help='Format of result')
+    for name, desc in EXTRADATA_PARAMS:
+        group.add_argument('--' + name, action='store_true', help=desc)
+
+    group.add_argument('--lang', '--accept-language', metavar='LANGS',
+                       help='Preferred language order for presenting search results')
+    group.add_argument('--polygon-output',
+                       choices=['geojson', 'kml', 'svg', 'text'],
+                       help='Output geometry of results as a GeoJSON, KML, SVG or WKT.')
+    group.add_argument('--polygon-threshold', type=float, metavar='TOLERANCE',
+                       help="""Simplify output geometry.
+                               Parameter is difference tolerance in degrees.""")
+
+
+class APISearch:

     """\
     """\

-    Todo
+    Execute API search query.

     """
     """

+

     @staticmethod
     def add_args(parser):
     @staticmethod
     def add_args(parser):

-        pass
+        group = parser.add_argument_group('Query arguments')
+        group.add_argument('--query',
+                           help='Free-form query string')
+        for name, desc in STRUCTURED_QUERY:
+            group.add_argument('--' + name, help='Structured query: ' + desc)
+
+        _add_api_output_arguments(parser)
+
+        group = parser.add_argument_group('Result limitation')
+        group.add_argument('--countrycodes', metavar='CC,..',
+                           help='Limit search results to one or more countries.')
+        group.add_argument('--exclude_place_ids', metavar='ID,..',
+                           help='List of search object to be excluded')
+        group.add_argument('--limit', type=int,
+                           help='Limit the number of returned results')
+        group.add_argument('--viewbox', metavar='X1,Y1,X2,Y2',
+                           help='Preferred area to find search results')
+        group.add_argument('--bounded', action='store_true',
+                           help='Strictly restrict results to viewbox area')
+
+        group = parser.add_argument_group('Other arguments')
+        group.add_argument('--no-dedupe', action='store_false', dest='dedupe',
+                           help='Do not remove duplicates from the result list')
+

 
     @staticmethod
 
     @staticmethod

-    def run(args): # pylint: disable=W0613
-        print("TODO: searching")
+    def run(args):
+        if args.query:
+            params = dict(q=args.query)
+        else:
+            params = {k : getattr(args, k) for k, _ in STRUCTURED_QUERY if getattr(args, k)}
+
+        for param, _ in EXTRADATA_PARAMS:
+            if getattr(args, param):
+                params[param] = '1'
+        for param in ('format', 'countrycodes', 'exclude_place_ids', 'limit', 'viewbox'):
+            if getattr(args, param):
+                params[param] = getattr(args, param)
+        if args.lang:
+            params['accept-language'] = args.lang
+        if args.polygon_output:
+            params['polygon_' + args.polygon_output] = '1'
+        if args.polygon_threshold:
+            params['polygon_threshold'] = args.polygon_threshold
+        if args.bounded:
+            params['bounded'] = '1'
+        if not args.dedupe:
+            params['dedupe'] = '0'
+
+        return run_api_script('search', args.project_dir,
+                              phpcgi_bin=args.phpcgi_path, params=params)
+
+class APIReverse:
+    """\
+    Execute API reverse query.
+    """
+
+    @staticmethod
+    def add_args(parser):
+        group = parser.add_argument_group('Query arguments')
+        group.add_argument('--lat', type=float, required=True,
+                           help='Latitude of coordinate to look up (in WGS84)')
+        group.add_argument('--lon', type=float, required=True,
+                           help='Longitude of coordinate to look up (in WGS84)')
+        group.add_argument('--zoom', type=int,
+                           help='Level of detail required for the address')
+
+        _add_api_output_arguments(parser)
+
+
+    @staticmethod
+    def run(args):
+        params = dict(lat=args.lat, lon=args.lon)
+        if args.zoom is not None:
+            params['zoom'] = args.zoom
+
+        for param, _ in EXTRADATA_PARAMS:
+            if getattr(args, param):
+                params[param] = '1'
+        if args.format:
+            params['format'] = args.format
+        if args.lang:
+            params['accept-language'] = args.lang
+        if args.polygon_output:
+            params['polygon_' + args.polygon_output] = '1'
+        if args.polygon_threshold:
+            params['polygon_threshold'] = args.polygon_threshold
+
+        return run_api_script('reverse', args.project_dir,
+                              phpcgi_bin=args.phpcgi_path, params=params)
+
+
+class APILookup:
+    """\
+    Execute API reverse query.
+    """
+
+    @staticmethod
+    def add_args(parser):
+        group = parser.add_argument_group('Query arguments')
+        group.add_argument('--id', metavar='OSMID',
+                           action='append', required=True, dest='ids',
+                           help='OSM id to lookup in format <NRW><id> (may be repeated)')
+
+        _add_api_output_arguments(parser)
+
+
+    @staticmethod
+    def run(args):
+        params = dict(osm_ids=','.join(args.ids))
+
+        for param, _ in EXTRADATA_PARAMS:
+            if getattr(args, param):
+                params[param] = '1'
+        if args.format:
+            params['format'] = args.format
+        if args.lang:
+            params['accept-language'] = args.lang
+        if args.polygon_output:
+            params['polygon_' + args.polygon_output] = '1'
+        if args.polygon_threshold:
+            params['polygon_threshold'] = args.polygon_threshold
+
+        return run_api_script('lookup', args.project_dir,
+                              phpcgi_bin=args.phpcgi_path, params=params)
+
+
+class APIDetails:
+    """\
+    Execute API lookup query.
+    """
+
+    @staticmethod
+    def add_args(parser):
+        group = parser.add_argument_group('Query arguments')
+        objs = group.add_mutually_exclusive_group(required=True)
+        objs.add_argument('--node', '-n', type=int,
+                          help="Look up the OSM node with the given ID.")
+        objs.add_argument('--way', '-w', type=int,
+                          help="Look up the OSM way with the given ID.")
+        objs.add_argument('--relation', '-r', type=int,
+                          help="Look up the OSM relation with the given ID.")
+        objs.add_argument('--place_id', '-p', type=int,
+                          help='Database internal identifier of the OSM object to look up.')
+        group.add_argument('--class', dest='object_class',
+                           help="""Class type to disambiguated multiple entries
+                                   of the same object.""")
+
+        group = parser.add_argument_group('Output arguments')
+        for name, desc in DETAILS_SWITCHES:
+            group.add_argument('--' + name, action='store_true', help=desc)
+        group.add_argument('--lang', '--accept-language', metavar='LANGS',
+                           help='Preferred language order for presenting search results')
+
+    @staticmethod
+    def run(args):
+        if args.node:
+            params = dict(osmtype='N', osmid=args.node)
+        elif args.way:
+            params = dict(osmtype='W', osmid=args.node)
+        elif args.relation:
+            params = dict(osmtype='R', osmid=args.node)
+        else:
+            params = dict(place_id=args.place_id)
+        if args.object_class:
+            params['class'] = args.object_class
+        for name, _ in DETAILS_SWITCHES:
+            params[name] = '1' if getattr(args, name) else '0'
+
+        return run_api_script('details', args.project_dir,
+                              phpcgi_bin=args.phpcgi_path, params=params)
+
+
+class APIStatus:
+    """\
+    Execute API status query.
+    """
+
+    @staticmethod
+    def add_args(parser):
+        group = parser.add_argument_group('API parameters')
+        group.add_argument('--format', default='text', choices=['text', 'json'],
+                           help='Format of result')
+
+    @staticmethod
+    def run(args):
+        return run_api_script('status', args.project_dir,
+                              phpcgi_bin=args.phpcgi_path,
+                              params=dict(format=args.format))

 
 
 def nominatim(**kwargs):
 
 
 def nominatim(**kwargs):


@@ -522,10 +893,14 @@ def nominatim(**kwargs):
     parser.add_subcommand('refresh', UpdateRefresh)
 
     parser.add_subcommand('export', QueryExport)
     parser.add_subcommand('refresh', UpdateRefresh)
 
     parser.add_subcommand('export', QueryExport)

-    parser.add_subcommand('search', QueryTodo)
-    parser.add_subcommand('reverse', QueryTodo)
-    parser.add_subcommand('lookup', QueryTodo)
-    parser.add_subcommand('details', QueryTodo)
-    parser.add_subcommand('status', QueryTodo)
+
+    if kwargs.get('phpcgi_path'):
+        parser.add_subcommand('search', APISearch)
+        parser.add_subcommand('reverse', APIReverse)
+        parser.add_subcommand('lookup', APILookup)
+        parser.add_subcommand('details', APIDetails)
+        parser.add_subcommand('status', APIStatus)
+    else:
+        parser.parser.epilog = 'php-cgi not found. Query commands not available.'

 
     return parser.run(**kwargs)
 
     return parser.run(**kwargs)