Merge pull request #2784 from lonvia/doscs-customizing-icu-tokenizer

[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 c229716f8e8353d8d8788670b767ea0c9cebc22f..417d68d2025777b944d1944371dea3d9268c0616 100644 (file)
--- a/nominatim/tokenizer/sanitizers/clean_housenumbers.py
+++ b/nominatim/tokenizer/sanitizers/clean_housenumbers.py
@@ -24,11 +24,16 @@ Arguments:
                      or a list of strings, where each string is a regular
                      expression that must match the full house number value.
 """
                      or a list of strings, where each string is a regular
                      expression that must match the full house number value.
 """

+from typing import Callable, Iterator, List

 import re
 
 import re
 

+from nominatim.tokenizer.sanitizers.base import ProcessInfo
+from nominatim.data.place_name import PlaceName
+from nominatim.tokenizer.sanitizers.config import SanitizerConfig
+

 class _HousenumberSanitizer:
 
 class _HousenumberSanitizer:
 

-    def __init__(self, config):
+    def __init__(self, config: SanitizerConfig) -> None:

         self.filter_kind = config.get_filter_kind('housenumber')
         self.split_regexp = config.get_delimiter()
 
         self.filter_kind = config.get_filter_kind('housenumber')
         self.split_regexp = config.get_delimiter()
 


@@ -37,13 +42,13 @@ class _HousenumberSanitizer:
 
 
 
 
 
 

-    def __call__(self, obj):
+    def __call__(self, obj: ProcessInfo) -> None:

         if not obj.address:
             return
 
         if not obj.address:
             return
 

-        new_address = []
+        new_address: List[PlaceName] = []

         for item in obj.address:
         for item in obj.address:

-            if self.filter_kind(item):
+            if self.filter_kind(item.kind):

                 if self._treat_as_name(item.name):
                     obj.names.append(item.clone(kind='housenumber'))
                 else:
                 if self._treat_as_name(item.name):
                     obj.names.append(item.clone(kind='housenumber'))
                 else:


@@ -56,7 +61,7 @@ class _HousenumberSanitizer:
         obj.address = new_address
 
 
         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
         """ Extract housenumbers in a regularized format from an OSM value.
 
             The function works as a generator that yields all valid housenumbers


@@ -67,16 +72,15 @@ class _HousenumberSanitizer:
                 yield from self._regularize(hnr)
 
 
                 yield from self._regularize(hnr)
 
 

-    @staticmethod
-    def _regularize(hnr):
+    def _regularize(self, hnr: str) -> Iterator[str]:

         yield hnr
 
 
         yield hnr
 
 

-    def _treat_as_name(self, housenumber):
+    def _treat_as_name(self, housenumber: str) -> bool:

         return any(r.fullmatch(housenumber) is not None for r in self.is_name_regexp)
 
 
         return any(r.fullmatch(housenumber) is not None for r in self.is_name_regexp)
 
 

-def create(config):
+def create(config: SanitizerConfig) -> Callable[[ProcessInfo], None]:

     """ Create a housenumber processing function.
     """
 
     """ Create a housenumber processing function.
     """