CLI: get valid --format values via autodiscover

[nominatim.git] / nominatim / config.py

diff --git a/nominatim/config.py b/nominatim/config.py
index 5d56d024660752359c9a6eec71dae455ad7bc4cf..3344a425a5667d58f6e2cd79e916f0b9bbb8c839 100644 (file)
--- a/nominatim/config.py
+++ b/nominatim/config.py
@@ -1,89 +1,186 @@
+# 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.

 """
 Nominatim configuration accessor.
 """
 """
 Nominatim configuration accessor.
 """

+from typing import Dict, Any, List, Mapping, Optional
+import importlib.util

 import logging
 import os
 import logging
 import os

+import sys

 from pathlib import Path
 from pathlib import Path

+import json

 import yaml
 
 from dotenv import dotenv_values
 import yaml
 
 from dotenv import dotenv_values

+from psycopg2.extensions import parse_dsn

 
 

+from nominatim.typing import StrPath

 from nominatim.errors import UsageError
 from nominatim.errors import UsageError

+import nominatim.paths

 
 LOG = logging.getLogger()
 
 LOG = logging.getLogger()

+CONFIG_CACHE : Dict[str, Any] = {}

 
 

-class Configuration:
-    """ Load and manage the project configuration.
+def flatten_config_list(content: Any, section: str = '') -> List[Any]:
+    """ Flatten YAML configuration lists that contain include sections
+        which are lists themselves.
+    """
+    if not content:
+        return []

 
 

-        Nominatim uses dotenv to configure the software. Configuration options
-        are resolved in the following order:
+    if not isinstance(content, list):
+        raise UsageError(f"List expected in section '{section}'.")
+
+    output = []
+    for ele in content:
+        if isinstance(ele, list):
+            output.extend(flatten_config_list(ele, section))
+        else:
+            output.append(ele)

 
 

-         * from the OS environment (or the dirctionary given in `environ`
-         * from the .env file in the project directory of the installation
-         * from the default installation in the configuration directory
+    return output
+
+
+class Configuration:
+    """ This class wraps access to the configuration settings
+        for the Nominatim instance in use.

 
         All Nominatim configuration options are prefixed with 'NOMINATIM_' to
 
         All Nominatim configuration options are prefixed with 'NOMINATIM_' to

-        avoid conflicts with other environment variables.
+        avoid conflicts with other environment variables. All settings can
+        be accessed as properties of the class under the same name as the
+        setting but with the `NOMINATIM_` prefix removed. In addition, there
+        are accessor functions that convert the setting values to types
+        other than string.

     """
 
     """
 

-    def __init__(self, project_dir, config_dir, environ=None):
+    def __init__(self, project_dir: Optional[Path],
+                 environ: Optional[Mapping[str, str]] = None) -> None:

         self.environ = environ or os.environ
         self.project_dir = project_dir
         self.environ = environ or os.environ
         self.project_dir = project_dir

-        self.config_dir = config_dir
-        self._config = dotenv_values(str((config_dir / 'env.defaults').resolve()))
-        if project_dir is not None and (project_dir / '.env').is_file():
-            self._config.update(dotenv_values(str((project_dir / '.env').resolve())))
-
-        # Add defaults for variables that are left empty to set the default.
-        # They may still be overwritten by environment variables.
-        if not self._config['NOMINATIM_ADDRESS_LEVEL_CONFIG']:
-            self._config['NOMINATIM_ADDRESS_LEVEL_CONFIG'] = \
-                str(config_dir / 'address-levels.json')
+        self.config_dir = nominatim.paths.CONFIG_DIR
+        self._config = dotenv_values(str(self.config_dir / 'env.defaults'))
+        if self.project_dir is not None and (self.project_dir / '.env').is_file():
+            self.project_dir = self.project_dir.resolve()
+            self._config.update(dotenv_values(str(self.project_dir / '.env')))

 
         class _LibDirs:
 
         class _LibDirs:

-            pass
+            module: Path
+            osm2pgsql: Path
+            php = nominatim.paths.PHPLIB_DIR
+            sql = nominatim.paths.SQLLIB_DIR
+            data = nominatim.paths.DATA_DIR

 
         self.lib_dir = _LibDirs()
 
         self.lib_dir = _LibDirs()

+        self._private_plugins: Dict[str, object] = {}
+

 
 

-    def set_libdirs(self, **kwargs):
+    def set_libdirs(self, **kwargs: StrPath) -> None:

         """ Set paths to library functions and data.
         """
         for key, value in kwargs.items():
         """ Set paths to library functions and data.
         """
         for key, value in kwargs.items():

-            setattr(self.lib_dir, key, Path(value).resolve())
+            setattr(self.lib_dir, key, Path(value))
+

 
 

-    def __getattr__(self, name):
+    def __getattr__(self, name: str) -> str:

         name = 'NOMINATIM_' + name
 
         if name in self.environ:
             return self.environ[name]
 
         name = 'NOMINATIM_' + name
 
         if name in self.environ:
             return self.environ[name]
 

-        return self._config[name]
+        return self._config[name] or ''

 
 

-    def get_bool(self, name):
+
+    def get_bool(self, name: str) -> bool:

         """ Return the given configuration parameter as a boolean.
         """ Return the given configuration parameter as a boolean.

-            Values of '1', 'yes' and 'true' are accepted as truthy values,
-            everything else is interpreted as false.
+
+            Parameters:
+              name: Name of the configuration parameter with the NOMINATIM_
+                prefix removed.
+
+            Returns:
+              `True` for values of '1', 'yes' and 'true', `False` otherwise.

         """
         """

-        return self.__getattr__(name).lower() in ('1', 'yes', 'true')
+        return getattr(self, name).lower() in ('1', 'yes', 'true')

 
 
 
 

-    def get_int(self, name):
+    def get_int(self, name: str) -> int:

         """ Return the given configuration parameter as an int.
         """ Return the given configuration parameter as an int.

+
+            Parameters:
+              name: Name of the configuration parameter with the NOMINATIM_
+                prefix removed.
+
+            Returns:
+              The configuration value converted to int.
+
+            Raises:
+              ValueError: when the value is not a number.

         """
         try:
         """
         try:

-            return int(self.__getattr__(name))
+            return int(getattr(self, name))

         except ValueError as exp:
             LOG.fatal("Invalid setting NOMINATIM_%s. Needs to be a number.", name)
             raise UsageError("Configuration error.") from exp
 
 
         except ValueError as exp:
             LOG.fatal("Invalid setting NOMINATIM_%s. Needs to be a number.", name)
             raise UsageError("Configuration error.") from exp
 
 

-    def get_libpq_dsn(self):
+    def get_str_list(self, name: str) -> Optional[List[str]]:
+        """ Return the given configuration parameter as a list of strings.
+            The values are assumed to be given as a comma-sparated list and
+            will be stripped before returning them. 
+
+            Parameters:
+              name: Name of the configuration parameter with the NOMINATIM_
+                prefix removed.
+
+            Returns:
+              (List[str]): The comma-split parameter as a list. The
+                elements are stripped of leading and final spaces before
+                being returned.
+              (None): The configuration parameter was unset or empty.
+        """
+        raw = getattr(self, name)
+
+        return [v.strip() for v in raw.split(',')] if raw else None
+
+
+    def get_path(self, name: str) -> Optional[Path]:
+        """ Return the given configuration parameter as a Path.
+
+            Parameters:
+              name: Name of the configuration parameter with the NOMINATIM_
+                prefix removed.
+
+            Returns:
+              (Path): A Path object of the parameter value.
+                  If a relative path is configured, then the function converts this
+                  into an absolute path with the project directory as root path.
+              (None): The configuration parameter was unset or empty.
+        """
+        value = getattr(self, name)
+        if not value:
+            return None
+
+        cfgpath = Path(value)
+
+        if not cfgpath.is_absolute():
+            assert self.project_dir is not None
+            cfgpath = self.project_dir / cfgpath
+
+        return cfgpath.resolve()
+
+
+    def get_libpq_dsn(self) -> str:

         """ Get configured database DSN converted into the key/value format
             understood by libpq and psycopg.
         """
         dsn = self.DATABASE_DSN
 
         """ Get configured database DSN converted into the key/value format
             understood by libpq and psycopg.
         """
         dsn = self.DATABASE_DSN
 

-        def quote_param(param):
+        def quote_param(param: str) -> str:

             key, val = param.split('=')
             val = val.replace('\\', '\\\\').replace("'", "\\'")
             if ' ' in val:
             key, val = param.split('=')
             val = val.replace('\\', '\\\\').replace("'", "\\'")
             if ' ' in val:


@@ -97,33 +194,46 @@ class Configuration:
         return dsn
 
 
         return dsn
 
 

-    def get_import_style_file(self):
+    def get_database_params(self) -> Mapping[str, str]:
+        """ Get the configured parameters for the database connection
+            as a mapping.
+        """
+        dsn = self.DATABASE_DSN
+
+        if dsn.startswith('pgsql:'):
+            return dict((p.split('=', 1) for p in dsn[6:].split(';')))
+
+        return parse_dsn(dsn)
+
+
+    def get_import_style_file(self) -> Path:

         """ Return the import style file as a path object. Translates the
             name of the standard styles automatically into a file in the
             config style.
         """
         """ Return the import style file as a path object. Translates the
             name of the standard styles automatically into a file in the
             config style.
         """

-        style = self.__getattr__('IMPORT_STYLE')
+        style = getattr(self, 'IMPORT_STYLE')

 
         if style in ('admin', 'street', 'address', 'full', 'extratags'):
 
         if style in ('admin', 'street', 'address', 'full', 'extratags'):

-            return self.config_dir / 'import-{}.style'.format(style)
+            return self.config_dir / f'import-{style}.lua'

 
 

-        return Path(style)
+        return self.find_config_file('', 'IMPORT_STYLE')

 
 
 
 

-    def get_os_env(self):
+    def get_os_env(self) -> Dict[str, str]:

         """ Return a copy of the OS environment with the Nominatim configuration
             merged in.
         """
         """ Return a copy of the OS environment with the Nominatim configuration
             merged in.
         """

-        env = dict(self._config)
+        env = {k: v for k, v in self._config.items() if v is not None}

         env.update(self.environ)
 
         return env
 
 
         env.update(self.environ)
 
         return env
 
 

-    def load_sub_configuration(self, filename, config=None):
+    def load_sub_configuration(self, filename: StrPath,
+                               config: Optional[str] = None) -> Any:

         """ Load additional configuration from a file. `filename` is the name
             of the configuration file. The file is first searched in the
         """ Load additional configuration from a file. `filename` is the name
             of the configuration file. The file is first searched in the

-            project directory and then in the global settings dirctory.
+            project directory and then in the global settings directory.

 
             If `config` is set, then the name of the configuration file can
             be additionally given through a .env configuration option. When
 
             If `config` is set, then the name of the configuration file can
             be additionally given through a .env configuration option. When


@@ -140,23 +250,77 @@ class Configuration:
             is loaded using this function and added at the position in the
             configuration tree.
         """
             is loaded using this function and added at the position in the
             configuration tree.
         """

-        assert Path(filename).suffix == '.yaml'
+        configfile = self.find_config_file(filename, config)
+
+        if str(configfile) in CONFIG_CACHE:
+            return CONFIG_CACHE[str(configfile)]
+
+        if configfile.suffix in ('.yaml', '.yml'):
+            result = self._load_from_yaml(configfile)
+        elif configfile.suffix == '.json':
+            with configfile.open('r', encoding='utf-8') as cfg:
+                result = json.load(cfg)
+        else:
+            raise UsageError(f"Config file '{configfile}' has unknown format.")
+
+        CONFIG_CACHE[str(configfile)] = result
+        return result
+
+
+    def load_plugin_module(self, module_name: str, internal_path: str) -> Any:
+        """ Load a Python module as a plugin.
+
+            The module_name may have three variants:
+
+            * A name without any '.' is assumed to be an internal module
+              and will be searched relative to `internal_path`.
+            * If the name ends in `.py`, module_name is assumed to be a
+              file name relative to the project directory.
+            * Any other name is assumed to be an absolute module name.
+
+            In either of the variants the module name must start with a letter.
+        """
+        if not module_name or not module_name[0].isidentifier():
+            raise UsageError(f'Invalid module name {module_name}')
+
+        if '.' not in module_name:
+            module_name = module_name.replace('-', '_')
+            full_module = f'{internal_path}.{module_name}'
+            return sys.modules.get(full_module) or importlib.import_module(full_module)
+
+        if module_name.endswith('.py'):
+            if self.project_dir is None or not (self.project_dir / module_name).exists():
+                raise UsageError(f"Cannot find module '{module_name}' in project directory.")
+
+            if module_name in self._private_plugins:
+                return self._private_plugins[module_name]
+
+            file_path = str(self.project_dir / module_name)
+            spec = importlib.util.spec_from_file_location(module_name, file_path)
+            if spec:
+                module = importlib.util.module_from_spec(spec)
+                # Do not add to global modules because there is no standard
+                # module name that Python can resolve.
+                self._private_plugins[module_name] = module
+                assert spec.loader is not None
+                spec.loader.exec_module(module)

 
 

-        configfile = self._find_config_file(filename, config)
+                return module

 
 

-        return self._load_from_yaml(configfile)
+        return sys.modules.get(module_name) or importlib.import_module(module_name)

 
 
 
 

-    def _find_config_file(self, filename, config=None):
+    def find_config_file(self, filename: StrPath,
+                         config: Optional[str] = None) -> Path:

         """ Resolve the location of a configuration file given a filename and
             an optional configuration option with the file name.
             Raises a UsageError when the file cannot be found or is not
             a regular file.
         """
         if config is not None:
         """ Resolve the location of a configuration file given a filename and
             an optional configuration option with the file name.
             Raises a UsageError when the file cannot be found or is not
             a regular file.
         """
         if config is not None:

-            cfg_filename = self.__getattr__(config)
-            if cfg_filename:
-                cfg_filename = Path(cfg_filename)
+            cfg_value = getattr(self, config)
+            if cfg_value:
+                cfg_filename = Path(cfg_value)

 
                 if cfg_filename.is_absolute():
                     cfg_filename = cfg_filename.resolve()
 
                 if cfg_filename.is_absolute():
                     cfg_filename = cfg_filename.resolve()


@@ -172,7 +336,7 @@ class Configuration:
 
         search_paths = [self.project_dir, self.config_dir]
         for path in search_paths:
 
         search_paths = [self.project_dir, self.config_dir]
         for path in search_paths:

-            if (path / filename).is_file():
+            if path is not None and (path / filename).is_file():

                 return path / filename
 
         LOG.fatal("Configuration file '%s' not found.\nDirectories searched: %s",
                 return path / filename
 
         LOG.fatal("Configuration file '%s' not found.\nDirectories searched: %s",


@@ -180,7 +344,7 @@ class Configuration:
         raise UsageError("Config file not found.")
 
 
         raise UsageError("Config file not found.")
 
 

-    def _load_from_yaml(self, cfgfile):
+    def _load_from_yaml(self, cfgfile: Path) -> Any:

         """ Load a YAML configuration file. This installs a special handler that
             allows to include other YAML files using the '!include' operator.
         """
         """ Load a YAML configuration file. This installs a special handler that
             allows to include other YAML files using the '!include' operator.
         """


@@ -189,18 +353,18 @@ class Configuration:
         return yaml.safe_load(cfgfile.read_text(encoding='utf-8'))
 
 
         return yaml.safe_load(cfgfile.read_text(encoding='utf-8'))
 
 

-    def _yaml_include_representer(self, loader, node):
+    def _yaml_include_representer(self, loader: Any, node: yaml.Node) -> Any:

         """ Handler for the '!include' operator in YAML files.
 
             When the filename is relative, then the file is first searched in the
         """ Handler for the '!include' operator in YAML files.
 
             When the filename is relative, then the file is first searched in the

-            project directory and then in the global settings dirctory.
+            project directory and then in the global settings directory.

         """
         fname = loader.construct_scalar(node)
 
         if Path(fname).is_absolute():
             configfile = Path(fname)
         else:
         """
         fname = loader.construct_scalar(node)
 
         if Path(fname).is_absolute():
             configfile = Path(fname)
         else:

-            configfile = self._find_config_file(loader.construct_scalar(node))
+            configfile = self.find_config_file(loader.construct_scalar(node))

 
         if configfile.suffix != '.yaml':
             LOG.fatal("Format error while reading '%s': only YAML format supported.",
 
         if configfile.suffix != '.yaml':
             LOG.fatal("Format error while reading '%s': only YAML format supported.",