add japanese sanitizer

[nominatim.git] / nominatim / tokenizer / sanitizers / clean_housenumbers.py

diff --git a/nominatim/tokenizer/sanitizers/clean_housenumbers.py b/nominatim/tokenizer/sanitizers/clean_housenumbers.py
index b65880c38e5179e167b8a2711e31b079e900d8ab..ae6349a9c846cba2d9fa403471fe0c0f5b335e17 100644 (file)
--- a/nominatim/tokenizer/sanitizers/clean_housenumbers.py
+++ b/nominatim/tokenizer/sanitizers/clean_housenumbers.py
@@ -5,30 +5,52 @@
 # Copyright (C) 2022 by the Nominatim developer community.
 # For a full list of authors see the git log.
 """
-Sanitizer that cleans and normalizes housenumbers.
+Sanitizer that preprocesses address tags for house numbers. The sanitizer
+allows to
+
+* define which tags are to be considered house numbers (see 'filter-kind')
+* split house number lists into individual numbers (see 'delimiters')
 
 Arguments:
     delimiters: Define the set of characters to be used for
-                splitting a list of housenumbers into parts. (default: ',;')
-
+                splitting a list of house numbers into parts. (default: ',;')
+    filter-kind: Define the address tags that are considered to be a
+                 house number. Either takes a single string or a list of strings,
+                 where each string is a regular expression. An address item
+                 is considered a house number if the 'kind' fully matches any
+                 of the given regular expressions. (default: 'housenumber')
+    convert-to-name: Define house numbers that should be treated as a name
+                     instead of a house number. Either takes a single string
+                     or a list of strings, where each string is a regular
+                     expression that must match the full house number value.
 """
-from nominatim.tokenizer.sanitizers.helpers import create_split_regex
+from typing import Callable, Iterator, List
+
+from nominatim.tokenizer.sanitizers.base import ProcessInfo
+from nominatim.data.place_name import PlaceName
+from nominatim.tokenizer.sanitizers.config import SanitizerConfig
 
 class _HousenumberSanitizer:
 
-    def __init__(self, config):
-        self.kinds = config.get('filter-kind', ('housenumber', ))
-        self.split_regexp = create_split_regex(config)
+    def __init__(self, config: SanitizerConfig) -> None:
+        self.filter_kind = config.get_filter('filter-kind', ['housenumber'])
+        self.split_regexp = config.get_delimiter()
+
+        self.filter_name = config.get_filter('convert-to-name', 'FAIL_ALL')
 
 
-    def __call__(self, obj):
+    def __call__(self, obj: ProcessInfo) -> None:
         if not obj.address:
             return
 
-        new_address = []
+        new_address: List[PlaceName] = []
         for item in obj.address:
-            if item.kind in self.kinds:
-                new_address.extend(item.clone(kind='housenumber', name=n) for n in self.sanitize(item.name))
+            if self.filter_kind(item.kind):
+                if self.filter_name(item.name):
+                    obj.names.append(item.clone(kind='housenumber'))
+                else:
+                    new_address.extend(item.clone(kind='housenumber', name=n)
+                                       for n in self.sanitize(item.name))
             else:
                 # Don't touch other address items.
                 new_address.append(item)
@@ -36,7 +58,7 @@ class _HousenumberSanitizer:
         obj.address = new_address
 
 
-    def sanitize(self, value):
+    def sanitize(self, value: str) -> Iterator[str]:
         """ Extract housenumbers in a regularized format from an OSM value.
 
             The function works as a generator that yields all valid housenumbers
@@ -47,11 +69,11 @@ class _HousenumberSanitizer:
                 yield from self._regularize(hnr)
 
 
-    def _regularize(self, hnr):
+    def _regularize(self, hnr: str) -> Iterator[str]:
         yield hnr
 
 
-def create(config):
+def create(config: SanitizerConfig) -> Callable[[ProcessInfo], None]:
     """ Create a housenumber processing function.
     """