prefilter bad results before adding details and reranking

[nominatim.git] / docs / develop / Tokenizers.md

diff --git a/docs/develop/Tokenizers.md b/docs/develop/Tokenizers.md
index c82071b62aa82dd3d8246307472528bbbc441174..050371771c27eb21ffa7158efbc261e8ae35154d 100644 (file)
--- a/docs/develop/Tokenizers.md
+++ b/docs/develop/Tokenizers.md
@@ -135,13 +135,13 @@ and implement the abstract functions defined there.
 
 ::: nominatim.tokenizer.base.AbstractTokenizer
     options:
-        heading_level: 3
+        heading_level: 6
 
 ### Python Analyzer Class
 
 ::: nominatim.tokenizer.base.AbstractAnalyzer
     options:
-        heading_level: 3
+        heading_level: 6
 
 ### PL/pgSQL Functions