From 30b82ba95f925d5bbcd7698f399830dcaab8b73a Mon Sep 17 00:00:00 2001
From: Igor Lins e Silva <4753812+igorls@users.noreply.github.com>
Date: Tue, 2 Jun 2026 21:50:08 -0300
Subject: [PATCH] fix: avoid qdrant lexical full scan on empty text hits

---
 mempalace/backends/qdrant.py |  4 +++-
 tests/test_qdrant_backend.py | 20 ++++++++++++++++++++
 2 files changed, 23 insertions(+), 1 deletion(-)

diff --git a/mempalace/backends/qdrant.py b/mempalace/backends/qdrant.py
index d43b6ed..2600b0f 100644
--- a/mempalace/backends/qdrant.py
+++ b/mempalace/backends/qdrant.py
@@ -1027,18 +1027,20 @@ class QdrantCollection(BaseCollection):
         q_filter = None if _requires_local_filter(where) else _qdrant_filter(where)
         rows = []
         text_filter = _text_any_filter(query)
+        text_filter_success = False
         if text_filter:
             try:
                 rows = self._scroll_all(
                     qdrant_filter=_combine_filters(q_filter, text_filter),
                     with_vector=False,
                 )
+                text_filter_success = True
             except BackendError:
                 logger.debug(
                     "Qdrant text filter failed; falling back to lexical scan", exc_info=True
                 )
                 rows = []
-        if not rows:
+        if not text_filter_success:
             rows = self._scroll_all(qdrant_filter=q_filter, with_vector=False)
         rows = [row for row in rows if _matches_where(row["metadata"], where)]
         scores = _bm25_scores(query, [row["document"] for row in rows])
diff --git a/tests/test_qdrant_backend.py b/tests/test_qdrant_backend.py
index 827adc7..d7c873b 100644
--- a/tests/test_qdrant_backend.py
+++ b/tests/test_qdrant_backend.py
@@ -78,6 +78,7 @@ class _FakeQdrantClient:
     def __init__(self, _config):
         self.collections = {}
         self.query_calls = []
+        self.scroll_calls = []
         self.created_indexes = []
         _FakeQdrantClient.instances.append(self)
 
@@ -143,6 +144,7 @@ class _FakeQdrantClient:
         offset=None,
         with_vector=False,
     ):
+        self.scroll_calls.append(qdrant_filter)
         points = list(self.collections.get(collection, {"points": {}})["points"].values())
         points = [point for point in points if _fake_match_filter(point, qdrant_filter)]
         start = int(offset or 0)
@@ -312,6 +314,24 @@ def test_qdrant_complex_filters_use_exact_local_fallback(tmp_path, fake_qdrant):
     assert fake_client.query_calls == []
 
 
+def test_qdrant_lexical_empty_text_filter_does_not_full_scan(tmp_path, fake_qdrant):
+    _backend, col = _collection(tmp_path)
+    col.upsert(
+        ids=["a", "b"],
+        documents=["alpha backend note", "beta frontend note"],
+        metadatas=[{"wing": "project"}, {"wing": "project"}],
+        embeddings=[[1, 0], [0, 1]],
+    )
+    fake_client = fake_qdrant.instances[0]
+    fake_client.scroll_calls.clear()
+
+    hits = col.lexical_search(query="missingterm", n_results=2).hits
+
+    assert hits == []
+    assert len(fake_client.scroll_calls) == 1
+    assert "text_any" in str(fake_client.scroll_calls[0])
+
+
 def test_qdrant_dimension_mismatch(tmp_path, fake_qdrant):
     _backend, col = _collection(tmp_path)
     col.upsert(ids=["a"], documents=["one"], metadatas=[{}], embeddings=[[1, 0]])