Enhancement: add highlighting to title + content searches

src/documents/search/_backend.py

@@ -21,6 +21,7 @@ from guardian.shortcuts import get_users_with_perms
 
 from documents.search._normalize import ascii_fold
 from documents.search._query import build_permission_filter
+from documents.search._query import parse_simple_highlight_query
 from documents.search._query import parse_simple_text_query
 from documents.search._query import parse_simple_title_query
 from documents.search._query import parse_user_query
@@ -335,6 +336,17 @@ class TantivyBackend:
         else:
             return parse_user_query(self._index, query, tz)
 
+    def _parse_highlight_query(
+        self,
+        query: str,
+        search_mode: SearchMode,
+    ) -> tantivy.Query:
+        if search_mode is SearchMode.TEXT:
+            # title does not supported highlight for now
+            return parse_simple_highlight_query(self._index, query, ["content"])
+        else:
+            return self._parse_query(query, search_mode)
+
     def _apply_permission_filter(
         self,
         query: tantivy.Query,
@@ -549,6 +561,7 @@ class TantivyBackend:
 
         self._ensure_open()
         user_query = self._parse_query(query, search_mode)
+        highlight_query = self._parse_highlight_query(query, search_mode)
 
         # For notes_text snippet generation, we need a query that targets the
         # notes_text field directly. user_query may contain JSON-field terms
@@ -601,7 +614,7 @@ class TantivyBackend:
                 if snippet_generator is None:
                     snippet_generator = tantivy.SnippetGenerator.create(
                         searcher,
-                        user_query,
+                        highlight_query,
                         self._schema,
                         "content",
                     )
@@ -610,7 +623,7 @@ class TantivyBackend:
                 if content_html:
                     highlights["content"] = content_html
 
-                if "notes_text" in doc_dict:
+                if search_mode is SearchMode.QUERY and "notes_text" in doc_dict:
                     # Use notes_text (plain text) for snippet generation — tantivy's
                     # SnippetGenerator does not support JSON fields.
                     if notes_snippet_generator is None:

src/documents/search/_query.py

@@ -452,6 +452,14 @@ _FIELD_BOOSTS = {"title": 2.0}
 _SIMPLE_FIELD_BOOSTS = {"simple_title": 2.0}
 
 
+def _simple_query_tokens(raw_query: str) -> list[str]:
+    tokens = [
+        ascii_fold(token.lower())
+        for token in _SIMPLE_QUERY_TOKEN_RE.findall(raw_query, timeout=_REGEX_TIMEOUT)
+    ]
+    return [token for token in tokens if token]
+
+
 def _build_simple_field_query(
     index: tantivy.Index,
     field: str,
@@ -547,11 +555,7 @@ def parse_simple_query(
 
     Query string is escaped and normalized to be treated as "simple" text query.
     """
-    tokens = [
-        ascii_fold(token.lower())
-        for token in _SIMPLE_QUERY_TOKEN_RE.findall(raw_query, timeout=_REGEX_TIMEOUT)
-    ]
-    tokens = [token for token in tokens if token]
+    tokens = _simple_query_tokens(raw_query)
     if not tokens:
         return tantivy.Query.empty_query()
 
@@ -564,6 +568,30 @@ def parse_simple_query(
     return tantivy.Query.boolean_query(field_queries)
 
 
+def parse_simple_highlight_query(
+    index: tantivy.Index,
+    raw_query: str,
+    fields: list[str],
+) -> tantivy.Query:
+    """Build a snippet-friendly query for simple text/title searches.
+
+    Simple search matching uses regex queries over the normalized shadow fields to
+    support substring matches. Tantivy's SnippetGenerator does not produce
+    highlights for that query shape, so for snippet generation we build a plain
+    term query over the real stored text fields instead.
+    """
+
+    tokens = _simple_query_tokens(raw_query)
+    if not tokens:
+        return tantivy.Query.empty_query()
+
+    return index.parse_query(
+        " ".join(tokens),
+        fields,
+        field_boosts={field: _FIELD_BOOSTS.get(field, 1.0) for field in fields},
+    )
+
+
 def parse_simple_text_query(
     index: tantivy.Index,
     raw_query: str,

src/documents/tests/search/test_backend.py

@@ -563,6 +563,26 @@ class TestFieldHandling:
 class TestHighlightHits:
     """Test highlight_hits returns proper HTML strings, not raw Snippet objects."""
 
+    def test_highlights_simple_text_mode_returns_html_string(
+        self,
+        backend: TantivyBackend,
+    ):
+        """Simple text search should still produce content highlights for exact-token hits."""
+        doc = Document.objects.create(
+            title="Highlight Test",
+            content="The quick brown fox jumps over the lazy dog",
+            checksum="HH0",
+            pk=89,
+        )
+        backend.add_or_update(doc)
+
+        hits = backend.highlight_hits("quick", [doc.pk], search_mode=SearchMode.TEXT)
+
+        assert len(hits) == 1
+        highlights = hits[0]["highlights"]
+        assert "content" in highlights
+        assert "<b>" in highlights["content"]
+
     def test_highlights_content_returns_html_string(self, backend: TantivyBackend):
         """highlight_hits must return HTML strings (from Snippet.to_html()), not Snippet objects."""
         doc = Document.objects.create(