add support for external token analysis modules

[nominatim.git] / nominatim / config.py

diff --git a/nominatim/config.py b/nominatim/config.py
index ed7d35e2448e55c9530406a2439b5acdcefedf3b..7502aff703ebaaf4610ddc4b591d212957c1229c 100644 (file)
--- a/nominatim/config.py
+++ b/nominatim/config.py
@@ -7,19 +7,20 @@
 """
 Nominatim configuration accessor.
 """
-from typing import Dict, Any, List, Mapping, Optional, Union
+from typing import Dict, Any, List, Mapping, Optional
+import importlib.util
 import logging
 import os
+import sys
 from pathlib import Path
 import json
 import yaml
 
 from dotenv import dotenv_values
 
+from nominatim.typing import StrPath
 from nominatim.errors import UsageError
 
-PathOrStr = Union[str, os.PathLike[str]]
-
 LOG = logging.getLogger()
 CONFIG_CACHE : Dict[str, Any] = {}
 
@@ -67,12 +68,17 @@ class Configuration:
             self._config.update(dotenv_values(str((project_dir / '.env').resolve())))
 
         class _LibDirs:
-            pass
+            module: Path
+            osm2pgsql: Path
+            php: Path
+            sql: Path
+            data: Path
 
         self.lib_dir = _LibDirs()
+        self._private_plugins: Dict[str, object] = {}
 
 
-    def set_libdirs(self, **kwargs: PathOrStr) -> None:
+    def set_libdirs(self, **kwargs: StrPath) -> None:
         """ Set paths to library functions and data.
         """
         for key, value in kwargs.items():
@@ -178,11 +184,11 @@ class Configuration:
         return env
 
 
-    def load_sub_configuration(self, filename: PathOrStr,
+    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
-            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
@@ -216,7 +222,50 @@ class Configuration:
         return result
 
 
-    def find_config_file(self, filename: PathOrStr,
+    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)
+
+                return module
+
+        return sys.modules.get(module_name) or importlib.import_module(module_name)
+
+
+    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.
@@ -263,7 +312,7 @@ class Configuration:
         """ 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)