bdd: recreate project directory for every run

[nominatim.git] / nominatim / cli.py

diff --git a/nominatim/cli.py b/nominatim/cli.py
index 83d79a685424313b2dbd8c077d18d9e6db3542da..56ed6a078eb3a082f9430564b279b1470348c89e 100644 (file)
--- a/nominatim/cli.py
+++ b/nominatim/cli.py
@@ -1,21 +1,34 @@
+# 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.

 """
 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
+from typing import Optional, Any, List, Union
+import logging

 import os
 import os

+import sys

 import argparse
 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 import version
+from nominatim.clicmd.args import NominatimArgs, Subcommand
+
+LOG = logging.getLogger()

 
 class CommandlineParser:
     """ Wraps some of the common functions for parsing the command line
         and setting up subcommands.
     """
 
 class CommandlineParser:
     """ Wraps some of the common functions for parsing the command line
         and setting up subcommands.
     """

-    def __init__(self, prog, description):
+    def __init__(self, prog: str, description: Optional[str]):

         self.parser = argparse.ArgumentParser(
             prog=prog,
             description=description,
         self.parser = argparse.ArgumentParser(
             prog=prog,
             description=description,


@@ -24,6 +37,10 @@ class CommandlineParser:
         self.subs = self.parser.add_subparsers(title='available commands',
                                                dest='subcommand')
 
         self.subs = self.parser.add_subparsers(title='available commands',
                                                dest='subcommand')
 

+        # Global arguments that only work if no sub-command given
+        self.parser.add_argument('--version', action='store_true',
+                                 help='Print Nominatim version and exit')
+

         # Arguments added to every sub-command
         self.default_args = argparse.ArgumentParser(add_help=False)
         group = self.default_args.add_argument_group('Default arguments')
         # Arguments added to every sub-command
         self.default_args = argparse.ArgumentParser(add_help=False)
         group = self.default_args.add_argument_group('Default arguments')


@@ -40,11 +57,22 @@ class CommandlineParser:
                            help='Number of parallel threads to use')
 
 
                            help='Number of parallel threads to use')
 
 

-    def add_subcommand(self, name, cmd):
+    def nominatim_version_text(self) -> str:
+        """ Program name and version number as string
+        """
+        text = f'Nominatim version {version.version_str()}'
+        if version.GIT_COMMIT_HASH is not None:
+            text += f' ({version.GIT_COMMIT_HASH})'
+        return text
+
+
+    def add_subcommand(self, name: str, cmd: Subcommand) -> None:

         """ Add a subcommand to the parser. The subcommand must be a class
             with a function add_args() that adds the parameters for the
             subcommand and a run() function that executes the command.
         """
         """ Add a subcommand to the parser. The subcommand must be a class
             with a function add_args() that adds the parameters for the
             subcommand and a run() function that executes the command.
         """

+        assert cmd.__doc__ is not None
+

         parser = self.subs.add_parser(name, parents=[self.default_args],
                                       help=cmd.__doc__.split('\n', 1)[0],
                                       description=cmd.__doc__,
         parser = self.subs.add_parser(name, parents=[self.default_args],
                                       help=cmd.__doc__.split('\n', 1)[0],
                                       description=cmd.__doc__,


@@ -53,29 +81,54 @@ class CommandlineParser:
         parser.set_defaults(command=cmd)
         cmd.add_args(parser)
 
         parser.set_defaults(command=cmd)
         cmd.add_args(parser)
 

-    def run(self, **kwargs):
+
+    def run(self, **kwargs: Any) -> int:

         """ 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()
+        try:
+            self.parser.parse_args(args=kwargs.get('cli_args'), namespace=args)
+        except SystemExit:
+            return 1
+
+        if args.version:
+            print(self.nominatim_version_text())
+            return 0

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

-            return self.parser.print_help()
+            self.parser.print_help()
+            return 1
+
+        args.phpcgi_path = Path(kwargs['phpcgi_path'])
+        args.project_dir = Path(args.project_dir).resolve()

 
 

-        for arg in ('module_dir', 'osm2pgsql_path', 'phplib_dir', 'data_dir'):
-            setattr(args, arg, Path(kwargs[arg]))
-        args.project_dir = Path(args.project_dir)
+        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)

 
 

-        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,
+                                    environ=kwargs.get('environ', os.environ))
+        args.config.set_libdirs(module=kwargs['module_dir'],
+                                osm2pgsql=kwargs['osm2pgsql_path'])

 
 

-        args.config = Configuration(args.project_dir, args.data_dir / 'settings')
+        log = logging.getLogger()
+        log.warning('Using project directory: %s', str(args.project_dir))

 
 

-        return args.command.run(args)
+        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)

 
 

-##### Subcommand classes
+        # 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.
 #
 # Each class needs to implement two functions: add_args() adds the CLI parameters
 # for the subfunction, run() executes the subcommand.


@@ -86,330 +139,12 @@ class CommandlineParser:
 #
 # No need to document the functions each time.
 # pylint: disable=C0111
 #
 # No need to document the functions each time.
 # pylint: disable=C0111

-
-
-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', dest='continue_at',
-                           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('--index-noanalyse', action='store_true',
-                           help='Do not perform analyse operations during index')
-
-
-    @staticmethod
-    def run(args):
-        params = ['setup.php']
-        if args.osm_file:
-            params.extend(('--all', '--osm-file', args.osm_file))
-        else:
-            if args.continue_at == 'load-data':
-                params.append('--load-data')
-            if args.continue_at in ('load-data', 'indexing'):
-                params.append('--index')
-            params.extend(('--create-search-indices', '--create-country-names',
-                           '--setup-website'))
-        if args.osm2pgsql_cache:
-            params.extend(('--osm2pgsql-cache', args.osm2pgsql_cache))
-        if args.reverse_only:
-            params.append('--reverse-only')
-        if args.enable_debug_statements:
-            params.append('--enable-debug-statements')
-        if args.no_partitions:
-            params.append('--no-partitions')
-        if args.no_updates:
-            params.append('--drop')
-        if args.ignore_errors:
-            params.append('--ignore-errors')
-        if args.index_noanalyse:
-            params.append('--index-noanalyse')
-
-        return run_legacy_script(*params, nominatim_env=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):
-        return run_legacy_script('setup.php', '--drop', nominatim_env=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):
-        if args.output.name != '<stdout>':
-            raise NotImplementedError('Only output to stdout is currently implemented.')
-        return run_legacy_script('specialphrases.php', '--wiki-import', nominatim_env=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):
-        params = ['update.php']
-        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 run_legacy_script(*params, nominatim_env=args)
-
-
-class UpdateAddData:
-    """\
-    Add additional data from a file or an online source.
-
-    Data is only imported, not indexed. You need to call `nominatim-update index`
-    to complete the process.
-    """
-
-    @staticmethod
-    def add_args(parser):
-        group_name = parser.add_argument_group('Source')
-        group = group_name.add_mutually_exclusive_group(required=True)
-        group.add_argument('--file', metavar='FILE',
-                           help='Import data from an OSM file')
-        group.add_argument('--diff', metavar='FILE',
-                           help='Import data from an OSM diff file')
-        group.add_argument('--node', metavar='ID', type=int,
-                           help='Import a single node from the API')
-        group.add_argument('--way', metavar='ID', type=int,
-                           help='Import a single way from the API')
-        group.add_argument('--relation', metavar='ID', type=int,
-                           help='Import a single relation from the API')
-        group.add_argument('--tiger-data', metavar='DIR',
-                           help='Add housenumbers from the US TIGER census database.')
-        group = parser.add_argument_group('Extra arguments')
-        group.add_argument('--use-main-api', action='store_true',
-                           help='Use OSM API instead of Overpass to download objects')
-
-    @staticmethod
-    def run(args):
-        if args.tiger_data:
-            os.environ['NOMINATIM_TIGER_DATA_PATH'] = args.tiger_data
-            return run_legacy_script('setup.php', '--import-tiger-data', nominatim_env=args)
-
-        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 UpdateIndex:
-    """\
-    Reindex all new and modified data.
-    """
-
-    @staticmethod
-    def add_args(parser):
-        pass
-
-    @staticmethod
-    def run(args):
-        return run_legacy_script('update.php', '--index', nominatim_env=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('--functions', action='store_true',
-                           help='Update the PL/pgSQL functions in the database')
-        group.add_argument('--wiki-data', action='store_true',
-                           help='Update Wikipedia/data importance numbers.')
-        group.add_argument('--importance', action='store_true',
-                           help='Recompute place importances (expensive!)')
-        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')
-        group.add_argument('--enable-debug-statements', action='store_true',
-                           help='Enable debug warning statements in functions')
-
-    @staticmethod
-    def run(args):
-        if args.postcodes:
-            run_legacy_script('update.php', '--calculate-postcodes',
-                              nominatim_env=args, throw_on_fail=True)
-        if args.word_counts:
-            run_legacy_script('update.php', '--recompute-word-counts',
-                              nominatim_env=args, throw_on_fail=True)
-        if args.address_levels:
-            run_legacy_script('update.php', '--update-address-levels',
-                              nominatim_env=args, throw_on_fail=True)
-        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)
-        if args.wiki_data:
-            run_legacy_script('setup.php', '--import-wikipedia-articles',
-                              nominatim_env=args, throw_on_fail=True)
-        # Attention: importance MUST come after wiki data import.
-        if args.importance:
-            run_legacy_script('update.php', '--recompute-importance',
-                              nominatim_env=args, throw_on_fail=True)
-        if args.website:
-            run_legacy_script('setup.php', '--setup-website',
-                              nominatim_env=args, throw_on_fail=True)
-
-
-class AdminCheckDatabase:
-    """\
-    Check that the database is complete and operational.
-    """
-
-    @staticmethod
-    def add_args(parser):
-        pass # No options
-
-    @staticmethod
-    def run(args):
-        return run_legacy_script('check_import_finished.php', nominatim_env=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')
-        return run_legacy_script(*params, nominatim_env=args)
-
-

 class QueryExport:
     """\
 class QueryExport:
     """\

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

     """
 
     """
 

-    @staticmethod
-    def add_args(parser):
+    def add_args(self, parser: argparse.ArgumentParser) -> None:

         group = parser.add_argument_group('Output arguments')
         group.add_argument('--output-type', default='street',
                            choices=('continent', 'country', 'state', 'county',
         group = parser.add_argument_group('Output arguments')
         group.add_argument('--output-type', default='street',
                            choices=('continent', 'country', 'state', 'county',


@@ -417,16 +152,16 @@ 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')


@@ -438,11 +173,10 @@ class QueryExport:
                            help='Export only children of this OSM relation')
 
 
                            help='Export only children of this OSM relation')
 
 

-    @staticmethod
-    def run(args):
-        params = ['export.php',
-                  '--output-type', args.output_type,
-                  '--output-format', args.output_format]
+    def run(self, args: NominatimArgs) -> int:
+        params: List[Union[int, str]] = [
+                             '--output-type', args.output_type,
+                             '--output-format', args.output_format]

         if args.output_all_postcodes:
             params.append('--output-all-postcodes')
         if args.language:
         if args.output_all_postcodes:
             params.append('--output-all-postcodes')
         if args.language:


@@ -456,46 +190,70 @@ class QueryExport:
         if args.restrict_to_osm_relation:
             params.extend(('--restrict-to-osm-relation', args.restrict_to_osm_relation))
 
         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)
+        return run_legacy_script('export.php', *params, config=args.config)

 
 

-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 development. 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):
-        pass

 
 

-    @staticmethod
-    def run(args): # pylint: disable=W0613
-        print("TODO: searching")
+    def add_args(self, parser: argparse.ArgumentParser) -> None:
+        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.')
+
+
+    def run(self, args: NominatimArgs) -> int:
+        run_php_server(args.server, args.project_dir / 'website')
+        return 0

 
 
 
 

-def nominatim(**kwargs):
+def get_set_parser(**kwargs: Any) -> CommandlineParser:

     """\
     """\

-    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('import', clicmd.SetupAll())
+    parser.add_subcommand('freeze', clicmd.SetupFreeze())
+    parser.add_subcommand('replication', clicmd.UpdateReplication())
+
+    parser.add_subcommand('special-phrases', clicmd.ImportSpecialPhrases())
+
+    parser.add_subcommand('add-data', clicmd.UpdateAddData())
+    parser.add_subcommand('index', clicmd.UpdateIndex())
+    parser.add_subcommand('refresh', clicmd.UpdateRefresh())
+
+    parser.add_subcommand('admin', clicmd.AdminFuncs())

 
 

-    parser.add_subcommand('check-database', AdminCheckDatabase)
-    parser.add_subcommand('warm', AdminWarm)
+    parser.add_subcommand('export', QueryExport())
+    parser.add_subcommand('serve', AdminServe())

 
 

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

 
 

-    parser.add_subcommand('add-data', UpdateAddData)
-    parser.add_subcommand('index', UpdateIndex)
-    parser.add_subcommand('refresh', UpdateRefresh)
+    return parser

 
 

-    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)
+
+def nominatim(**kwargs: Any) -> int:
+    """\
+    Command-line tools for importing, updating, administrating and
+    querying the Nominatim database.
+    """
+    parser = get_set_parser(**kwargs)

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