]> git.openstreetmap.org Git - nominatim.git/blobdiff - nominatim/tools/country_info.py
Merge remote-tracking branch 'upstream/master'
[nominatim.git] / nominatim / tools / country_info.py
index 1b61ae685cc6106ac522b00c9c08d582983e2039..d754b4ddb029365b22d2cc7a77ccaeefc49a2719 100644 (file)
@@ -1,3 +1,9 @@
+# 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.
 """
 Functions for importing and managing static country information.
 """
@@ -5,19 +11,128 @@ import psycopg2.extras
 
 from nominatim.db import utils as db_utils
 from nominatim.db.connection import connect
+from nominatim.errors import UsageError
+
+def _flatten_name_list(names):
+    if names is None:
+        return {}
+
+    if not isinstance(names, dict):
+        raise UsageError("Expected key-value list for names in country_settings.py")
+
+    flat = {}
+    for prefix, remain in names.items():
+        if isinstance(remain, str):
+            flat[prefix] = remain
+        elif not isinstance(remain, dict):
+            raise UsageError("Entries in names must be key-value lists.")
+        else:
+            for suffix, name in remain.items():
+                if suffix == 'default':
+                    flat[prefix] = name
+                else:
+                    flat[f'{prefix}:{suffix}'] = name
+
+    return flat
+
+
+
+class _CountryInfo:
+    """ Caches country-specific properties from the configuration file.
+    """
+
+    def __init__(self):
+        self._info = {}
+
+
+    def load(self, config):
+        """ Load the country properties from the configuration files,
+            if they are not loaded yet.
+        """
+        if not self._info:
+            self._info = config.load_sub_configuration('country_settings.yaml')
+            for prop in self._info.values():
+                # Convert languages into a list for simpler handling.
+                if 'languages' not in prop:
+                    prop['languages'] = []
+                elif not isinstance(prop['languages'], list):
+                    prop['languages'] = [x.strip()
+                                         for x in prop['languages'].split(',')]
+                prop['names'] = _flatten_name_list(prop.get('names'))
+
+
+    def items(self):
+        """ Return tuples of (country_code, property dict) as iterable.
+        """
+        return self._info.items()
+
+    def get(self, country_code):
+        """ Get country information for the country with the given country code.
+        """
+        return self._info.get(country_code, {})
+
+
+
+_COUNTRY_INFO = _CountryInfo()
+
+
+def setup_country_config(config):
+    """ Load country properties from the configuration file.
+        Needs to be called before using any other functions in this
+        file.
+    """
+    _COUNTRY_INFO.load(config)
+
+
+def iterate(prop=None):
+    """ Iterate over country code and properties.
+
+        When `prop` is None, all countries are returned with their complete
+        set of properties.
+
+        If `prop` is given, then only countries are returned where the
+        given property is set. The second item of the tuple contains only
+        the content of the given property.
+    """
+    if prop is None:
+        return _COUNTRY_INFO.items()
+
+    return ((c, p[prop]) for c, p in _COUNTRY_INFO.items() if prop in p)
+
 
 def setup_country_tables(dsn, sql_dir, ignore_partitions=False):
     """ Create and populate the tables with basic static data that provides
         the background for geocoding. Data is assumed to not yet exist.
     """
-    db_utils.execute_file(dsn, sql_dir / 'country_name.sql')
     db_utils.execute_file(dsn, sql_dir / 'country_osm_grid.sql.gz')
 
-    if ignore_partitions:
-        with connect(dsn) as conn:
-            with conn.cursor() as cur:
-                cur.execute('UPDATE country_name SET partition = 0')
-            conn.commit()
+    params = []
+    for ccode, props in _COUNTRY_INFO.items():
+        if ccode is not None and props is not None:
+            if ignore_partitions:
+                partition = 0
+            else:
+                partition = props.get('partition')
+            lang = props['languages'][0] if len(
+                props['languages']) == 1 else None
+
+            params.append((ccode, props['names'], lang, partition))
+    with connect(dsn) as conn:
+        with conn.cursor() as cur:
+            psycopg2.extras.register_hstore(cur)
+            cur.execute(
+                """ CREATE TABLE public.country_name (
+                        country_code character varying(2),
+                        name public.hstore,
+                        derived_name public.hstore,
+                        country_default_language_code text,
+                        partition integer
+                    ); """)
+            cur.execute_values(
+                """ INSERT INTO public.country_name
+                    (country_code, name, country_default_language_code, partition) VALUES %s
+                """, params)
+        conn.commit()
 
 
 def create_country_names(conn, tokenizer, languages=None):
@@ -26,12 +141,9 @@ def create_country_names(conn, tokenizer, languages=None):
         empty then only name translations for the given languages are added
         to the index.
     """
-    if languages:
-        languages = languages.split(',')
-
     def _include_key(key):
-        return key == 'name' or \
-               (key.startswith('name:') and (not languages or key[5:] in languages))
+        return ':' not in key or not languages or \
+               key[key.index(':') + 1:] in languages
 
     with conn.cursor() as cur:
         psycopg2.extras.register_hstore(cur)
@@ -41,14 +153,10 @@ def create_country_names(conn, tokenizer, languages=None):
         with tokenizer.name_analyzer() as analyzer:
             for code, name in cur:
                 names = {'countrycode': code}
-                if code == 'gb':
-                    names['short_name'] = 'UK'
-                if code == 'us':
-                    names['short_name'] = 'United States'
 
                 # country names (only in languages as provided)
                 if name:
-                    names.update(((k, v) for k, v in name.items() if _include_key(k)))
+                    names.update({k : v for k, v in name.items() if _include_key(k)})
 
                 analyzer.add_country_names(code, names)