Really fix the regression in flat search cutoff.

typesense · Apr 26, 2024 · 58c8198 · 58c8198
1 parent 397518c
commit 58c8198
Show file tree

Hide file tree

Showing 2 changed files with 27 additions and 3 deletions.
diff --git a/src/index.cpp b/src/index.cpp
@@ -3069,6 +3069,10 @@ Option<bool> Index::search(std::vector<query_tokens_t>& field_query_tokens, cons
                 group_by_field_it_vec = get_group_by_field_iterators(group_by_fields);
             }
 
+            std::sort(dist_results.begin(), dist_results.end(), [](const auto& p1, const auto& p2) {
+                return p1.first < p2.first;
+            });
+
             for (auto& dist_result : dist_results) {
                 auto& seq_id = dist_result.second.seq_id;
                 auto references = std::move(dist_result.second.reference_filter_results);
@@ -3077,7 +3081,7 @@ Option<bool> Index::search(std::vector<query_tokens_t>& field_query_tokens, cons
                     continue;
                 }
 
-                if(vector_query.query_doc_given && nearest_ids.size() >= k-1) {
+                if(vector_query.query_doc_given && nearest_ids.size() == k-1) {
                     // When id based vector query is made, we ask for K+1 results to account for the query
                     // record itself being returned. However, when the filter clause does not match the
                     // query record, we will end up returning 1 extra hit.

diff --git a/test/collection_vector_search_test.cpp b/test/collection_vector_search_test.cpp
@@ -768,7 +768,7 @@ TEST_F(CollectionVectorTest, VecSearchWithFiltering) {
     ASSERT_EQ(10, results["hits"].size());
 
     // with points:<10, flat-search
-    results = coll1->search("*", {}, "points:<10", {}, {}, {0}, 20, 1, FREQUENCY, {true}, Index::DROP_TOKENS_THRESHOLD,
+    results = coll1->search("*", {}, "points:<10", {}, {}, {0}, 3, 1, FREQUENCY, {true}, Index::DROP_TOKENS_THRESHOLD,
                             spp::sparse_hash_set<std::string>(),
                             spp::sparse_hash_set<std::string>(), 10, "", 30, 5,
                             "", 10, {}, {}, {}, 0,
@@ -778,13 +778,33 @@ TEST_F(CollectionVectorTest, VecSearchWithFiltering) {
                             false, true, "vec:([0.96826, 0.94, 0.39557, 0.306488], flat_search_cutoff: 1000)").get();
 
     ASSERT_EQ(10, results["found"].get<size_t>());
-    ASSERT_EQ(10, results["hits"].size());
+    ASSERT_EQ(3, results["hits"].size());
     ASSERT_FLOAT_EQ(3.409385e-05, results["hits"][0]["vector_distance"].get<float>());
     ASSERT_EQ("1", results["hits"][0]["document"]["id"].get<std::string>());
 
     ASSERT_FLOAT_EQ(0.016780376, results["hits"][1]["vector_distance"].get<float>());
     ASSERT_EQ("5", results["hits"][1]["document"]["id"].get<std::string>());
 
+    results = coll1->search("*", {}, "points:<10", {}, {}, {0}, 3, 1, FREQUENCY, {true}, Index::DROP_TOKENS_THRESHOLD,
+                            spp::sparse_hash_set<std::string>(),
+                            spp::sparse_hash_set<std::string>(), 10, "", 30, 5,
+                            "", 10, {}, {}, {}, 0,
+                            "<mark>", "</mark>", {}, 1000, true, false, true, "", false, 6000 * 1000, 4, 7,
+                            fallback,
+                            4, {off}, 32767, 32767, 2,
+                            false, true, "vec:([], id: 3, flat_search_cutoff: 1000)").get();
+
+    ASSERT_EQ(3, results["hits"].size());
+
+    LOG(INFO) << results["hits"][0];
+    LOG(INFO) << results["hits"][1];
+
+    ASSERT_EQ("9", results["hits"][0]["document"]["id"].get<std::string>());
+    ASSERT_FLOAT_EQ(0.050603985, results["hits"][0]["vector_distance"].get<float>());
+
+    ASSERT_EQ("5", results["hits"][1]["document"]["id"].get<std::string>());
+    ASSERT_FLOAT_EQ(0.100155532, results["hits"][1]["vector_distance"].get<float>());
+
     // single point
 
     results = coll1->search("*", {}, "points:1", {}, {}, {0}, 20, 1, FREQUENCY, {true}, Index::DROP_TOKENS_THRESHOLD,