From a79a3210e69c648e3e5c1123b4049ce0419b484c Mon Sep 17 00:00:00 2001
From: Sarah Hoffmann <lonvia@denofr.de>
Date: Sat, 5 Feb 2022 15:13:45 +0100
Subject: [PATCH] implement is-a-name option for housenumbers

---
 .../tokenizer/sanitizers/clean_housenumbers.py    | 15 +++++++++++++--
 settings/icu_tokenizer.yaml                       |  6 ++++--
 test/bdd/db/query/housenumbers.feature            | 14 ++++++++++++++
 3 files changed, 31 insertions(+), 4 deletions(-)

diff --git a/nominatim/tokenizer/sanitizers/clean_housenumbers.py b/nominatim/tokenizer/sanitizers/clean_housenumbers.py
index 49f9b4f0..48021793 100644
--- a/nominatim/tokenizer/sanitizers/clean_housenumbers.py
+++ b/nominatim/tokenizer/sanitizers/clean_housenumbers.py
@@ -29,6 +29,10 @@ class _HousenumberSanitizer:
         self.filter_kind = create_kind_filter(config, 'housenumber')
         self.split_regexp = create_split_regex(config)
 
+        nameregexps = config.get('is-a-name', [])
+        self.is_name_regexp = [re.compile(r) for r in nameregexps]
+
+
 
     def __call__(self, obj):
         if not obj.address:
@@ -37,8 +41,11 @@ class _HousenumberSanitizer:
         new_address = []
         for item in obj.address:
             if self.filter_kind(item):
-                new_address.extend(item.clone(kind='housenumber', name=n)
-                                   for n in self.sanitize(item.name))
+                if self.treat_as_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)
@@ -62,6 +69,10 @@ class _HousenumberSanitizer:
         yield hnr
 
 
+    def _treat_as_name(self, housenumber):
+        return any(r.fullmatch(housenumber) is not None for r in self.is_name_regexp)
+
+
 def create(config):
     """ Create a housenumber processing function.
     """
diff --git a/settings/icu_tokenizer.yaml b/settings/icu_tokenizer.yaml
index bf51f563..7f53c5a7 100644
--- a/settings/icu_tokenizer.yaml
+++ b/settings/icu_tokenizer.yaml
@@ -25,13 +25,15 @@ transliteration:
     - "[^a-z0-9[:Space:]] >"
     - ":: NFC ()"
 sanitizers:
-    - step: split-name-list
-    - step: strip-brace-terms
     - step: clean-housenumbers
       filter-kind:
         - housenumber
         - conscriptionnumber
         - streetnumber
+      is-a-name:
+        - (\A|.*,)[^\d,]{3,}(,.*|\Z)
+    - step: split-name-list
+    - step: strip-brace-terms
     - step: tag-analyzer-by-language
       filter-kind: [".*name.*"]
       whitelist: [bg,ca,cs,da,de,el,en,es,et,eu,fi,fr,gl,hu,it,ja,mg,ms,nl,no,pl,pt,ro,ru,sk,sl,sv,tr,uk,vi]
diff --git a/test/bdd/db/query/housenumbers.feature b/test/bdd/db/query/housenumbers.feature
index 63bd8984..bbb43f17 100644
--- a/test/bdd/db/query/housenumbers.feature
+++ b/test/bdd/db/query/housenumbers.feature
@@ -53,3 +53,17 @@ Feature: Searching of house numbers
         | 2;4;12 |
         | 2,4,12 |
         | 2, 4, 12 |
+
+
+    Scenario: A name mapped as a housenumber is found
+        Given the places
+         | osm | class    | type | housenr | geometry |
+         | N1  | building | yes  | Warring | 9        |
+        And the places
+         | osm | class   | type | name       | geometry |
+         | W10 | highway | path | Chester St | 1,2,3    |
+        When importing
+        When sending search query "Chester St Warring"
+        Then results contain
+         | osm |
+         | N1  |
-- 
2.39.5