🔧 (highlighter) Make translation and text fielfd opt-out

Author: simonwoerpel

openaleph_search/query/base.py

@@ -359,13 +359,11 @@ def get_highlight(self) -> dict[str, Any]:
                 self.HIGHLIGHT_FIELD, text, self.parser.highlight_count
             ),
             Field.NAMES: get_highlighter(Field.NAMES),
-            Field.TRANSLATION: get_highlighter(Field.TRANSLATION, text),
         }
-        # text field is a copy_to catch-all; it provides a fallback for
-        # entities whose content field text is not in _source (e.g.
-        # HyperText where indexText is popped from properties).
-        if Field.TEXT not in fields:
+        if settings.highlighter_text_field and Field.TEXT not in fields:
             fields[Field.TEXT] = get_highlighter(Field.TEXT, text)
+        if settings.highlighter_translation_field:
+            fields[Field.TRANSLATION] = get_highlighter(Field.TRANSLATION, text)
         # Add filter value highlights to the main highlight field only.
         # Only highlight filter values that are human-readable text.
         # Skip short-code groups (countries, languages, etc.) to avoid

openaleph_search/query/highlight.py

@@ -81,7 +81,7 @@ def get_highlighter(
         highlighter = {
             "type": "unified",  # Good for mixed content
             "fragment_size": 200,  # Longer to capture full names/titles
-            "number_of_fragments": 3,
+            "number_of_fragments": 1,
             "fragmenter": "simple",  # Don't break names awkwardly
             "pre_tags": [""],  # No markup
             "post_tags": [""],  # No markup
@@ -93,8 +93,7 @@ def get_highlighter(
     if field == Field.NAMES:
         return {
             "type": "plain",
-            "number_of_fragments": 3,
-            "max_analyzed_offset": 999999,  # probably many names
+            "number_of_fragments": 1,
             "pre_tags": [""],  # No markup
             "post_tags": [""],  # No markup
         }

openaleph_search/settings.py

@@ -86,7 +86,9 @@ class Settings(BaseSettings):
     highlighter_phrase_limit: int = 64
     highlighter_boundary_max_scan: int = 100
     highlighter_no_match_size: int = 300
-    highlighter_max_analyzed_offset: int = 999999
+    highlighter_text_field: bool = True
+    highlighter_translation_field: bool = True
+    highlighter_max_analyzed_offset: int = 100000
 
     # More Like This defaults
     mlt_min_doc_freq: int = 1

tests/test_search_query.py

@@ -143,8 +143,7 @@ def test_highlight(self):
                     },
                     "names": {
                         "type": "plain",
-                        "number_of_fragments": 3,
-                        "max_analyzed_offset": 999999,
+                        "number_of_fragments": 1,
                         "pre_tags": [""],
                         "post_tags": [""],
                     },