From a574bb2f56328e3596536c232edc4ed0e4073fd3 Mon Sep 17 00:00:00 2001 From: Sarah Hoffmann Date: Thu, 7 May 2026 13:45:18 +0200 Subject: [PATCH] add sanitizer for suspicious ref values Mostly for filtering ref tags from German post boxes. --- settings/icu_tokenizer.yaml | 5 +++++ test/bdd/features/db/import/naming.feature | 14 ++++++++++++++ 2 files changed, 19 insertions(+) diff --git a/settings/icu_tokenizer.yaml b/settings/icu_tokenizer.yaml index 45d21c2b..69bcda43 100644 --- a/settings/icu_tokenizer.yaml +++ b/settings/icu_tokenizer.yaml @@ -43,6 +43,11 @@ sanitizers: - step: clean-tiger-tags - step: split-name-list delimiters: ; + - step: delete-tags + filter-kind: ref + name: + - ".*,.*" + - ".{41,}" - step: strip-brace-terms - step: tag-analyzer-by-language filter-kind: [".*name.*"] diff --git a/test/bdd/features/db/import/naming.feature b/test/bdd/features/db/import/naming.feature index 944c2de7..1c76b849 100644 --- a/test/bdd/features/db/import/naming.feature +++ b/test/bdd/features/db/import/naming.feature @@ -102,3 +102,17 @@ Feature: Import and search of names Then the result set contains | object | | N3 | + + + Scenario: ref tags with full address are not searchable + Given the places + | osm | class | type | name+ref | geometry | + | N1 | amenity | postbox | Hoopter Straße 207, 21423 Winsen | 23 45 | + When importing + Then placex contains + | object | name+ref | + | N1 | Hoopter Straße 207, 21423 Winsen | + And search_name contains exactly + | object | + When geocoding "Hoopter Straße" + Then exactly 0 results are returned -- 2.47.3