From 01c2c26160f602c867470eeeb8f8cd1184d3e85a Mon Sep 17 00:00:00 2001
From: Itz-Agasta <rupamgolui69@gmail.com>
Date: Sat, 9 May 2026 21:05:43 +0530
Subject: [PATCH] Improves geocoder reranking logic for locale-specific names

- Refines how result labels are aggregated for reranking by using sets of localized names instead of concatenated strings.

- Updates tests to reflect improved handling of locale-specific queries.
---
 src/nominatim_api/search/geocoder.py          | 34 ++++++++-----------
 .../features/db/query/search_simple.feature   |  6 ++++
 2 files changed, 21 insertions(+), 19 deletions(-)

diff --git a/src/nominatim_api/search/geocoder.py b/src/nominatim_api/search/geocoder.py
index cbc64998..e67a278e 100644
--- a/src/nominatim_api/search/geocoder.py
+++ b/src/nominatim_api/search/geocoder.py
@@ -7,7 +7,7 @@
 """
 Public interface to the search code.
 """
-from typing import List, Any, Optional, Iterator, Tuple, Dict
+from typing import List, Any, Optional, Iterator, Tuple, Dict, Set
 import itertools
 import re
 import difflib
@@ -175,27 +175,25 @@ class ForwardGeocoder:
 
         return final
 
-    def _get_result_rerank_text(self, result: SearchResult) -> str:
+    def _get_result_rerank_text(self, result: SearchResult) -> Set[str]:
         if not self.params.locales:
-            return result.display_name or ''
+            return {result.display_name} if result.display_name else set()
 
-        label_parts: List[str] = []
+        label_parts: Set[str] = set()
         if result.address_rows:
             for line in result.address_rows:
                 if line.isaddress and line.names:
                     address_name = self.params.locales.display_name(line.names)
-                    if address_name and (
-                        not label_parts or label_parts[-1] != address_name
-                    ):
-                        label_parts.append(address_name)
+                    if address_name:
+                        label_parts.add(address_name)
 
         if label_parts:
-            return ', '.join(label_parts)
+            return label_parts
 
         if result.names:
-            return self.params.locales.display_name(result.names)
+            return {self.params.locales.display_name(result.names)}
 
-        return result.display_name or ''
+        return {result.display_name} if result.display_name else set()
 
     def rerank_by_query(self, query: QueryStruct, results: SearchResults) -> None:
         """ Adjust the accuracy of the localized result according to how well
@@ -218,7 +216,7 @@ class ForwardGeocoder:
             # (e.g., name:en) are included in the match pool.
             rerank_text = self._get_result_rerank_text(result)
             norm = self.query_analyzer.normalize_text(
-                ' '.join((rerank_text, result.country_code or ''))
+                ' '.join((*rerank_text, result.country_code or ''))
             )
             words = set((w for w in re.split('[-,: ]+', norm) if w))
             if not words:
@@ -233,13 +231,11 @@ class ForwardGeocoder:
             # to offset this.
             if result.rank_address == 4:
                 if self.params.locales and result.names:
-                    loc_names = [result.names[t] for t in self.params.locales.name_tags
-                                 if t in result.names]
-                    if loc_names:
-                        norm_loc = self.query_analyzer.normalize_text(' '.join(loc_names))
-                        loc_words = set(w for w in re.split('[-,: ]+', norm_loc) if w)
-                        if loc_words and loc_words.isdisjoint(qwords):
-                            result.accuracy += result.calculated_importance() * 0.5
+                    # Exclude country code from disjoint check
+                    check_words = (words - {result.country_code.lower()}) \
+                        if result.country_code else words
+                    if check_words and check_words.isdisjoint(qwords):
+                        result.accuracy += result.calculated_importance() * 0.5
                 else:
                     distance *= 2
             result.accuracy += distance * 0.3 / sum(len(w) for w in qwords)
diff --git a/test/bdd/features/db/query/search_simple.feature b/test/bdd/features/db/query/search_simple.feature
index 8ff6a44b..5f7a1fbb 100644
--- a/test/bdd/features/db/query/search_simple.feature
+++ b/test/bdd/features/db/query/search_simple.feature
@@ -114,3 +114,9 @@ Feature: Searching of simple objects
         Then result 0 contains
          | object |
          | N1     |
+        When geocoding "Athens"
+         | accept-language |
+         | de              |
+        Then result 0 contains
+         | object |
+         | N2     |
-- 
2.47.3