Merge pull request #2284 from lonvia/cleanup-word-frequency-computation

[nominatim.git] / nominatim / cli.py

diff --git a/nominatim/cli.py b/nominatim/cli.py
index a40dc07ea07f0375b956f5ca2267448fd84216f1..55f51aac72dc7d6a2b2337b198f2966d991a30f6 100644 (file)
--- a/nominatim/cli.py
+++ b/nominatim/cli.py
@@ -2,13 +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 logging
+import os

 import sys
 import argparse
 import sys
 import argparse

-import logging

 from pathlib import Path
 
 from pathlib import Path
 

-from .config import Configuration
-from .admin.exec_utils import run_legacy_script
+from nominatim.config import Configuration
+from nominatim.tools.exec_utils import run_legacy_script, run_php_server
+from nominatim.errors import UsageError
+from nominatim import clicmd
+from nominatim.clicmd.args import NominatimArgs
+from nominatim.tools import tiger_data
+
+LOG = logging.getLogger()
+

 
 class CommandlineParser:
     """ Wraps some of the common functions for parsing the command line
 
 class CommandlineParser:
     """ Wraps some of the common functions for parsing the command line


@@ -56,140 +64,59 @@ 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 = NominatimArgs()
+        self.parser.parse_args(args=kwargs.get('cli_args'), namespace=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', 'sqllib_dir',
+                    'data_dir', 'config_dir', 'phpcgi_path'):

             setattr(args, arg, Path(kwargs[arg]))
             setattr(args, arg, Path(kwargs[arg]))

-        args.project_dir = Path(args.project_dir)
-
-        logging.basicConfig(stream=sys.stderr,
-                            format='%(asctime)s %(levelname)s: %(message)s',
-                            datefmt='%Y-%m-%d %H:%M:%S',
-                            level=max(4 - args.verbose, 1) * 10)
-
-        args.config = Configuration(args.project_dir, args.data_dir / 'settings')
-
-        args.command.run(args)
-
-
-class SetupAll:
-    """\
-    Create a new Nominatim database from an OSM file.
-    """
-
-    @staticmethod
-    def add_args(parser):
-        group_name = parser.add_argument_group('Required arguments')
-        group = group_name.add_mutually_exclusive_group(required=True)
-        group.add_argument('--osm-file',
-                           help='OSM file to be imported.')
-        group.add_argument('--continue', nargs=1,
-                           choices=['load-data', 'indexing', 'db-postprocess'],
-                           help='Continue an import that was interrupted')
-        group = parser.add_argument_group('Optional arguments')
-        group.add_argument('--osm2pgsql-cache', metavar='SIZE', type=int,
-                           help='Size of cache to be used by osm2pgsql (in MB)')
-        group.add_argument('--reverse-only', action='store_true',
-                           help='Do not create tables and indexes for searching')
-        group.add_argument('--enable-debug-statements', action='store_true',
-                           help='Include debug warning statements in SQL code')
-        group.add_argument('--no-partitions', action='store_true',
-                           help="""Do not partition search indices
-                                   (speeds up import of single country extracts)""")
-        group.add_argument('--no-updates', action='store_true',
-                           help="""Do not keep tables that are only needed for
-                                   updating the database later""")
-        group = parser.add_argument_group('Expert options')
-        group.add_argument('--ignore-errors', action='store_true',
-                           help='Continue import even when errors in SQL are present')
-        group.add_argument('--disable-token-precalc', action='store_true',
-                           help='Disable name precalculation')
-        group.add_argument('--index-noanalyse', action='store_true',
-                           help='Do not perform analyse operations during index')
-
-
-    @staticmethod
-    def run(args):
-        print("TODO: ./utils/setup.php", args)
-
-
-class SetupFreeze:
-    """\
-    Make database read-only.
-
-    About half of data in the Nominatim database is kept only to be able to
-    keep the data up-to-date with new changes made in OpenStreetMap. This
-    command drops all this data and only keeps the part needed for geocoding
-    itself.
-
-    This command has the same effect as the `--no-updates` option for imports.
-    """
-
-    @staticmethod
-    def add_args(parser):
-        pass # No options
-
-    @staticmethod
-    def run(args):
-        print("TODO: setup drop", args)
-
-
-class SetupSpecialPhrases:
-    """\
-    Maintain special phrases.
-    """
-
-    @staticmethod
-    def add_args(parser):
-        group = parser.add_argument_group('Input arguments')
-        group.add_argument('--from-wiki', action='store_true',
-                           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):
-        print("./utils/specialphrases.php --from-wiki", args)
-
-
-class UpdateReplication:
-    """\
-    Update the database using an online replication service.
-    """
-
-    @staticmethod
-    def add_args(parser):
-        group = parser.add_argument_group('Arguments for initialisation')
-        group.add_argument('--init', action='store_true',
-                           help='Initialise the update process')
-        group.add_argument('--no-update-functions', dest='update_functions',
-                           action='store_false',
-                           help="""Do not update the trigger function to
-                                   support differential updates.""")
-        group = parser.add_argument_group('Arguments for updates')
-        group.add_argument('--check-for-updates', action='store_true',
-                           help='Check if new updates are available and exit')
-        group.add_argument('--once', action='store_true',
-                           help="""Download and apply updates only once. When
-                                   not set, updates are continuously applied""")
-        group.add_argument('--no-index', action='store_false', dest='do_index',
-                           help="""Do not index the new data. Only applicable
-                                   together with --once""")
-
-    @staticmethod
-    def run(args):
-        if args.init:
-            print('./utils/update.php --init-updates', args)
-        else:
-            print('./utils/update.php --import-osmosis(-all)', args)
-
-
+        args.project_dir = Path(args.project_dir).resolve()
+
+        if 'cli_args' not in kwargs:
+            logging.basicConfig(stream=sys.stderr,
+                                format='%(asctime)s: %(message)s',
+                                datefmt='%Y-%m-%d %H:%M:%S',
+                                level=max(4 - args.verbose, 1) * 10)
+
+        args.config = Configuration(args.project_dir, args.config_dir,
+                                    environ=kwargs.get('environ', os.environ))
+        args.config.set_libdirs(module=args.module_dir,
+                                osm2pgsql=args.osm2pgsql_path,
+                                php=args.phplib_dir,
+                                sql=args.sqllib_dir,
+                                data=args.data_dir)
+
+        log = logging.getLogger()
+        log.warning('Using project directory: %s', str(args.project_dir))
+
+        try:
+            return args.command.run(args)
+        except UsageError as exception:
+            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
+
+
+##### Subcommand classes
+#
+# Each class needs to implement two functions: add_args() adds the CLI parameters
+# for the subfunction, run() executes the subcommand.
+#
+# The class documentation doubles as the help text for the command. The
+# first line is also used in the summary when calling the program without
+# a subcommand.
+#
+# 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 UpdateAddData:
     """\
     Add additional data from a file or an online source.
 class UpdateAddData:
     """\
     Add additional data from a file or an online source.


@@ -220,98 +147,29 @@ class UpdateAddData:
 
     @staticmethod
     def run(args):
 
     @staticmethod
     def run(args):

-        print('./utils/update.php --import-*', args)
-
-
-class UpdateIndex:
-    """\
-    Reindex all new and modified data.
-    """
-
-    @staticmethod
-    def add_args(parser):
-        pass
-
-    @staticmethod
-    def run(args):
-        print('./utils/update.php --index', args)
-
-
-class UpdateRefresh:
-    """\
-    Recompute auxillary data used by the indexing process.
-
-    These functions must not be run in parallel with other update commands.
-    """
-
-    @staticmethod
-    def add_args(parser):
-        group = parser.add_argument_group('Data arguments')
-        group.add_argument('--postcodes', action='store_true',
-                           help='Update postcode centroid table')
-        group.add_argument('--word-counts', action='store_true',
-                           help='Compute frequency of full-word search terms')
-        group.add_argument('--address-levels', action='store_true',
-                           help='Reimport address level configuration')
-        group.add_argument('--importance', action='store_true',
-                           help='Recompute place importances')
-        group.add_argument('--functions', action='store_true',
-                           help='Update the PL/pgSQL functions in the database')
-        group.add_argument('--wiki-data',
-                           help='Update Wikipedia/data importance numbers.')
-        group.add_argument('--website', action='store_true',
-                           help='Refresh the directory that serves the scripts for the web API')
-        group = parser.add_argument_group('Arguments for function refresh')
-        group.add_argument('--no-diff-updates', action='store_false', dest='diffs',
-                           help='Do not enable code for propagating updates')
-
-    @staticmethod
-    def run(args):
-        print('./utils/update.php', args)
-
-
-class AdminCheckDatabase:
-    """\
-    Check that the database is complete and operational.
-    """
-
-    @staticmethod
-    def add_args(parser):
-        pass # No options
-
-    @staticmethod
-    def run(args):
-        print("TODO: ./utils/check_import_finished.php", args)
-
-
-class AdminWarm:
-    """\
-    Warm database caches for search and reverse queries.
-    """
-
-    @staticmethod
-    def add_args(parser):
-        group = parser.add_argument_group('Target arguments')
-        group.add_argument('--search-only', action='store_const', dest='target',
-                           const='search',
-                           help="Only pre-warm tables for search queries")
-        group.add_argument('--reverse-only', action='store_const', dest='target',
-                           const='reverse',
-                           help="Only pre-warm tables for reverse queries")
-
-    @staticmethod
-    def run(args):
-        params = ['warm.php']
-        if args.target == 'reverse':
-            params.append('--reverse-only')
-        if args.target == 'search':
-            params.append('--search-only')
+        if args.tiger_data:
+            return tiger_data.add_tiger_data(args.tiger_data,
+                                             args.config, args.threads or 1)
+
+        params = ['update.php']
+        if args.file:
+            params.extend(('--import-file', args.file))
+        elif args.diff:
+            params.extend(('--import-diff', args.diff))
+        elif args.node:
+            params.extend(('--import-node', args.node))
+        elif args.way:
+            params.extend(('--import-way', args.way))
+        elif args.relation:
+            params.extend(('--import-relation', args.relation))
+        if args.use_main_api:
+            params.append('--use-main-api')

         return run_legacy_script(*params, nominatim_env=args)
 
 
 class QueryExport:
     """\
         return run_legacy_script(*params, nominatim_env=args)
 
 
 class QueryExport:
     """\

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

     """
 
     @staticmethod
     """
 
     @staticmethod


@@ -323,16 +181,16 @@ class QueryExport:
                            help='Type of places to output (default: street)')
         group.add_argument('--output-format',
                            default='street;suburb;city;county;state;country',
                            help='Type of places to output (default: street)')
         group.add_argument('--output-format',
                            default='street;suburb;city;county;state;country',

-                           help="""Semicolon-separated list of address types
-                                   (see --output-type). Multiple ranks can be
-                                   merged into one column by simply using a
-                                   comma-separated list.""")
+                           help=("Semicolon-separated list of address types "
+                                 "(see --output-type). Multiple ranks can be "
+                                 "merged into one column by simply using a "
+                                 "comma-separated list."))

         group.add_argument('--output-all-postcodes', action='store_true',
         group.add_argument('--output-all-postcodes', action='store_true',

-                           help="""List all postcodes for address instead of
-                                   just the most likely one""")
+                           help=("List all postcodes for address instead of "
+                                 "just the most likely one"))

         group.add_argument('--language',
         group.add_argument('--language',

-                           help="""Preferred language for output
-                                   (use local name, if omitted)""")
+                           help=("Preferred language for output "
+                                 "(use local name, if omitted)"))

         group = parser.add_argument_group('Filter arguments')
         group.add_argument('--restrict-to-country', metavar='COUNTRY_CODE',
                            help='Export only objects within country')
         group = parser.add_argument_group('Filter arguments')
         group.add_argument('--restrict-to-country', metavar='COUNTRY_CODE',
                            help='Export only objects within country')


@@ -346,46 +204,85 @@ class QueryExport:
 
     @staticmethod
     def run(args):
 
     @staticmethod
     def run(args):

-        print("TODO: ./utils/export.php", args)
+        params = ['export.php',
+                  '--output-type', args.output_type,
+                  '--output-format', args.output_format]
+        if args.output_all_postcodes:
+            params.append('--output-all-postcodes')
+        if args.language:
+            params.extend(('--language', args.language))
+        if args.restrict_to_country:
+            params.extend(('--restrict-to-country', args.restrict_to_country))
+        if args.restrict_to_osm_node:
+            params.extend(('--restrict-to-osm-node', args.restrict_to_osm_node))
+        if args.restrict_to_osm_way:
+            params.extend(('--restrict-to-osm-way', args.restrict_to_osm_way))
+        if args.restrict_to_osm_relation:
+            params.extend(('--restrict-to-osm-relation', args.restrict_to_osm_relation))
+
+        return run_legacy_script(*params, nominatim_env=args)

 
 
 
 

-class QueryTodo:
+class AdminServe:

     """\
     """\

-    Todo
+    Start a simple web server for serving the API.
+
+    This command starts the built-in PHP webserver to serve the website
+    from the current project directory. This webserver is only suitable
+    for testing and develop. Do not use it in production setups!
+
+    By the default, the webserver can be accessed at: http://127.0.0.1:8088

     """
     """

+

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

-        pass
+        group = parser.add_argument_group('Server arguments')
+        group.add_argument('--server', default='127.0.0.1:8088',
+                           help='The address the server will listen to.')

 
 

+    @staticmethod

     def run(args):
     def run(args):

-        print("TODO: searching")
-
+        run_php_server(args.server, args.project_dir / 'website')

 
 

-def nominatim(**kwargs):
+def get_set_parser(**kwargs):

     """\
     """\

-    Command-line tools for importing, updating, administrating and
-    querying the Nominatim database.
+    Initializes the parser and adds various subcommands for
+    nominatim cli.

     """
     parser = CommandlineParser('nominatim', nominatim.__doc__)
 
     """
     parser = CommandlineParser('nominatim', nominatim.__doc__)
 

-    parser.add_subcommand('import', SetupAll)
-    parser.add_subcommand('freeze', SetupFreeze)
-    parser.add_subcommand('replication', UpdateReplication)
-
-    parser.add_subcommand('check-database', AdminCheckDatabase)
-    parser.add_subcommand('warm', AdminWarm)
+    parser.add_subcommand('import', clicmd.SetupAll)
+    parser.add_subcommand('freeze', clicmd.SetupFreeze)
+    parser.add_subcommand('replication', clicmd.UpdateReplication)

 
 

-    parser.add_subcommand('special-phrases', SetupSpecialPhrases)
+    parser.add_subcommand('special-phrases', clicmd.ImportSpecialPhrases)

 
     parser.add_subcommand('add-data', UpdateAddData)
 
     parser.add_subcommand('add-data', UpdateAddData)

-    parser.add_subcommand('index', UpdateIndex)
-    parser.add_subcommand('refresh', UpdateRefresh)
+    parser.add_subcommand('index', clicmd.UpdateIndex)
+    parser.add_subcommand('refresh', clicmd.UpdateRefresh)
+
+    parser.add_subcommand('admin', clicmd.AdminFuncs)

 
     parser.add_subcommand('export', QueryExport)
 
     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)
+    parser.add_subcommand('serve', AdminServe)
+
+    if kwargs.get('phpcgi_path'):
+        parser.add_subcommand('search', clicmd.APISearch)
+        parser.add_subcommand('reverse', clicmd.APIReverse)
+        parser.add_subcommand('lookup', clicmd.APILookup)
+        parser.add_subcommand('details', clicmd.APIDetails)
+        parser.add_subcommand('status', clicmd.APIStatus)
+    else:
+        parser.parser.epilog = 'php-cgi not found. Query commands not available.'
+
+    return parser
+
+
+def nominatim(**kwargs):
+    """\
+    Command-line tools for importing, updating, administrating and
+    querying the Nominatim database.
+    """
+    parser = get_set_parser(**kwargs)

 
 

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