Merge remote-tracking branch 'upstream/master'

[nominatim.git] / nominatim / tools / country_info.py

diff --git a/nominatim/tools/country_info.py b/nominatim/tools/country_info.py
index 675fac5d2018d74d48179700f927e3d4d2816061..f6a6293f46e46573c88d5e288c653b47f66a1cbc 100644 (file)
--- a/nominatim/tools/country_info.py
+++ b/nominatim/tools/country_info.py
@@ -7,20 +7,19 @@
 """
 Functions for importing and managing static country information.
 """
-import json
-from io import StringIO
 import psycopg2.extras
 
 from nominatim.db import utils as db_utils
 from nominatim.db.connection import connect
 
+
 class _CountryInfo:
     """ Caches country-specific properties from the configuration file.
     """
 
     def __init__(self):
         self._info = {}
-
+        self._key_prefix = 'name'
 
     def load(self, config):
         """ Load the country properties from the configuration files,
@@ -35,17 +34,24 @@ class _CountryInfo:
                 elif not isinstance(prop['languages'], list):
                     prop['languages'] = [x.strip()
                                          for x in prop['languages'].split(',')]
-                if 'names' not in prop:
-                    prop['names']['name'] = {}
+                if 'names' not in prop or prop['names'] is None:
+                    prop['names'] = {self._key_prefix: {}}
 
     def items(self):
         """ Return tuples of (country_code, property dict) as iterable.
         """
         return self._info.items()
 
+    def key_prefix(self):
+        """ Return the prefix that will be attached to the keys of the country
+            names values when storing them in the database
+        """
+        return self._key_prefix
+
 
 _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
@@ -66,25 +72,24 @@ def setup_country_tables(dsn, sql_dir, ignore_partitions=False):
     """
     db_utils.execute_file(dsn, sql_dir / 'country_osm_grid.sql.gz')
 
-    def prepend_name_to_keys(name):
-        return {'name:'+k: v for k, v in name.items()}
+    def add_prefix_to_keys(names, prefix):
+        return {prefix+':'+k: v for k, v in names.items()}
 
-    params, country_names_data = [], ''
+    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, partition, lang))
-
-            name = prepend_name_to_keys(props.get('names').get('name'))
-            name = json.dumps(name , ensure_ascii=False, separators=(', ', '=>'))
-            country_names_data += ccode + '\t' + name[1:-1] + '\n'
-
+            lang = props['languages'][0] if len(
+                props['languages']) == 1 else None
+            name = add_prefix_to_keys(props.get('names').get(
+                _COUNTRY_INFO.key_prefix()), _COUNTRY_INFO.key_prefix())
+            params.append((ccode, name, 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),
@@ -93,13 +98,10 @@ def setup_country_tables(dsn, sql_dir, ignore_partitions=False):
                         country_default_language_code text,
                         partition integer
                     ); """)
-            data = StringIO(country_names_data)
-            cur.copy_from(data, 'country_name', columns=('country_code', 'name'))
             cur.execute_values(
-                """ UPDATE country_name
-                    SET partition = part, country_default_language_code = lang
-                    FROM (VALUES %s) AS v (cc, part, lang)
-                    WHERE country_code = v.cc""", params)
+                """ INSERT INTO public.country_name
+                    (country_code, name, country_default_language_code, partition) VALUES %s
+                """, params)
         conn.commit()
 
 
@@ -113,8 +115,9 @@ def create_country_names(conn, tokenizer, languages=None):
         languages = languages.split(',')
 
     def _include_key(key):
-        return key.startswith('name:') and \
-            key[5:] in languages or key[5:] == 'default'
+        return key == _COUNTRY_INFO.key_prefix() or \
+            (key.startswith(_COUNTRY_INFO.key_prefix()+':') and
+             (not languages or key[len(_COUNTRY_INFO.key_prefix())+1:] in languages))
 
     with conn.cursor() as cur:
         psycopg2.extras.register_hstore(cur)
@@ -131,7 +134,8 @@ def create_country_names(conn, tokenizer, languages=None):
 
                 # 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)