sparse: add StaticHasRawData support (#1188)

Signed-off-by: Shawn Wang <shawn.wang@zilliz.com>

Author: sparknack

src/index/index_static.cc

@@ -116,8 +116,8 @@ IndexStaticFaced<DataType>::HasRawData(const IndexType& indexType, const IndexVe
     }
 
     static std::set<knowhere::IndexType> has_raw_data_index_set = {
-        IndexEnum::INDEX_FAISS_BIN_IDMAP,  IndexEnum::INDEX_FAISS_BIN_IVFFLAT,     IndexEnum::INDEX_FAISS_IVFFLAT,
-        IndexEnum::INDEX_FAISS_IVFFLAT_CC, IndexEnum::INDEX_SPARSE_INVERTED_INDEX, IndexEnum::INDEX_SPARSE_WAND};
+        IndexEnum::INDEX_FAISS_BIN_IDMAP, IndexEnum::INDEX_FAISS_BIN_IVFFLAT, IndexEnum::INDEX_FAISS_IVFFLAT,
+        IndexEnum::INDEX_FAISS_IVFFLAT_CC};
 
     static std::set<knowhere::IndexType> has_raw_data_index_alias_set = {"IVFBIN", "BINFLAT", "IVFFLAT", "IVFFLATCC"};
 

src/index/sparse/sparse_index_node.cc

@@ -245,6 +245,11 @@ class SparseInvertedIndexNode : public IndexNode {
         return expected<DataSetPtr>::Err(Status::not_implemented, "GetVectorByIds not implemented");
     }
 
+    static bool
+    StaticHasRawData(const knowhere::BaseConfig& /*config*/, const IndexVersion& /*version*/) {
+        return false;
+    }
+
     [[nodiscard]] bool
     HasRawData(const std::string& metric_type) const override {
         return false;
@@ -541,6 +546,14 @@ class SparseInvertedIndexNodeCC : public SparseInvertedIndexNode<T, use_wand> {
         return res;
     }
 
+    static bool
+    StaticHasRawData(const knowhere::BaseConfig& config, const IndexVersion& version) {
+        if (!config.metric_type.has_value()) {
+            return false;
+        }
+        return IsMetricType(config.metric_type.value(), metric::IP);
+    }
+
     [[nodiscard]] bool
     HasRawData(const std::string& metric_type) const override {
         return IsMetricType(metric_type, metric::IP);

tests/ut/test_index_check.cc

@@ -209,8 +209,8 @@ TEST_CASE("Test index has raw data", "[IndexHasRawData]") {
         CHECK_FALSE(knowhere::IndexStaticFaced<fp32>::HasRawData(IndexEnum::INDEX_CUVS_CAGRA, ver, {}));
 #endif
         // sparse index
-        CHECK(knowhere::IndexStaticFaced<fp32>::HasRawData(IndexEnum::INDEX_SPARSE_INVERTED_INDEX, ver, {}));
-        CHECK(knowhere::IndexStaticFaced<fp32>::HasRawData(IndexEnum::INDEX_SPARSE_WAND, ver, {}));
+        CHECK_FALSE(knowhere::IndexStaticFaced<fp32>::HasRawData(IndexEnum::INDEX_SPARSE_INVERTED_INDEX, ver, {}));
+        CHECK_FALSE(knowhere::IndexStaticFaced<fp32>::HasRawData(IndexEnum::INDEX_SPARSE_WAND, ver, {}));
     }
 
     SECTION("Special test") {

tests/ut/test_sparse.cc

@@ -137,6 +137,8 @@ TEST_CASE("Test Mem Sparse Index With Float Vector", "[float metrics]") {
             REQUIRE(idx.Build(train_ds, json) == knowhere::Status::success);
             REQUIRE(idx.Size() > 0);
             REQUIRE(idx.Count() == nb);
+            REQUIRE(idx.HasRawData(metric) ==
+                    knowhere::IndexStaticFaced<knowhere::fp32>::HasRawData(name, version, json));
 
             knowhere::BinarySet bs;
             REQUIRE(idx.Serialize(bs) == knowhere::Status::success);
@@ -562,7 +564,7 @@ TEST_CASE("Test Mem Sparse Index CC", "[float metrics]") {
 
     SECTION("Test GetVectorByIds") {
         std::vector<int64_t> ids = {0, 1, 2};
-        REQUIRE(idx.HasRawData(metric));
+        REQUIRE(idx.HasRawData(metric) == knowhere::IndexStaticFaced<knowhere::fp32>::HasRawData(name, version, json));
         auto results = idx.GetVectorByIds(GenIdsDataSet(3, ids));
         REQUIRE(results.has_value());
         auto xb = (knowhere::sparse::SparseRow<float>*)train_ds->GetTensor();