Refactoring

hnswlib/hnswalg.h

@@ -64,7 +64,7 @@ class HierarchicalNSW : public AlgorithmInterface<dist_t> {
     mutable std::atomic<long> metric_distance_computations{0};
     mutable std::atomic<long> metric_hops{0};
 
-    bool replace_deleted_ = false;  // flag to replace deleted elements (marked as deleted) during insertions
+    bool allow_replace_deleted_ = false;  // flag to replace deleted elements (marked as deleted) during insertions
 
     std::mutex deleted_elements_lock;  // lock for deleted_elements
     std::unordered_set<tableint> deleted_elements;  // contains internal ids of deleted elements
@@ -79,8 +79,8 @@ class HierarchicalNSW : public AlgorithmInterface<dist_t> {
         const std::string &location,
         bool nmslib = false,
         size_t max_elements = 0,
-        bool replace_deleted = false)
-        : replace_deleted_(replace_deleted) {
+        bool allow_replace_deleted = false)
+        : allow_replace_deleted_(allow_replace_deleted) {
         loadIndex(location, s, max_elements);
     }
 
@@ -91,11 +91,11 @@ class HierarchicalNSW : public AlgorithmInterface<dist_t> {
         size_t M = 16,
         size_t ef_construction = 200,
         size_t random_seed = 100,
-        bool replace_deleted = false)
+        bool allow_replace_deleted = false)
         : link_list_locks_(max_elements),
             label_op_locks_(MAX_LABEL_OPERATION_LOCKS),
             element_levels_(max_elements),
-            replace_deleted_(replace_deleted) {
+            allow_replace_deleted_(allow_replace_deleted) {
         max_elements_ = max_elements;
         num_deleted_ = 0;
         data_size_ = s->get_data_size();
@@ -707,7 +707,7 @@ class HierarchicalNSW : public AlgorithmInterface<dist_t> {
         for (size_t i = 0; i < cur_element_count; i++) {
             if (isMarkedDeleted(i)) {
                 num_deleted_ += 1;
-                if (replace_deleted_) deleted_elements.insert(i);
+                if (allow_replace_deleted_) deleted_elements.insert(i);
             }
         }
 
@@ -771,7 +771,7 @@ class HierarchicalNSW : public AlgorithmInterface<dist_t> {
             unsigned char *ll_cur = ((unsigned char *)get_linklist0(internalId))+2;
             *ll_cur |= DELETE_MARK;
             num_deleted_ += 1;
-            if (replace_deleted_) {
+            if (allow_replace_deleted_) {
                 std::unique_lock <std::mutex> lock_deleted_elements(deleted_elements_lock);
                 deleted_elements.insert(internalId);
             }
@@ -813,7 +813,7 @@ class HierarchicalNSW : public AlgorithmInterface<dist_t> {
             unsigned char *ll_cur = ((unsigned char *)get_linklist0(internalId)) + 2;
             *ll_cur &= ~DELETE_MARK;
             num_deleted_ -= 1;
-            if (replace_deleted_) {
+            if (allow_replace_deleted_) {
                 std::unique_lock <std::mutex> lock_deleted_elements(deleted_elements_lock);
                 deleted_elements.erase(internalId);
             }
@@ -847,7 +847,7 @@ class HierarchicalNSW : public AlgorithmInterface<dist_t> {
     * If replacement of deleted elements is enabled: replaces previously deleted point if any, updating it with new point
     */
     void addPoint(const void *data_point, labeltype label, bool replace_deleted = false) {
-        if ((replace_deleted_ == false) && (replace_deleted == true)) {
+        if ((allow_replace_deleted_ == false) && (replace_deleted == true)) {
             throw std::runtime_error("Replacement of deleted elements is disabled in constructor");
         }
 
@@ -1053,7 +1053,7 @@ class HierarchicalNSW : public AlgorithmInterface<dist_t> {
             auto search = label_lookup_.find(label);
             if (search != label_lookup_.end()) {
                 tableint existingInternalId = search->second;
-                if (replace_deleted_) {
+                if (allow_replace_deleted_) {
                     if (isMarkedDeleted(existingInternalId)) {
                         throw std::runtime_error("Can't use addPoint to update deleted elements if replacement of deleted elements is enabled.");
                     }

python_bindings/bindings.cpp

@@ -194,12 +194,12 @@ class Index {
         size_t M,
         size_t efConstruction,
         size_t random_seed,
-        bool replace_deleted) {
+        bool allow_replace_deleted) {
         if (appr_alg) {
             throw std::runtime_error("The index is already initiated.");
         }
         cur_l = 0;
-        appr_alg = new hnswlib::HierarchicalNSW<dist_t>(l2space, maxElements, M, efConstruction, random_seed, replace_deleted);
+        appr_alg = new hnswlib::HierarchicalNSW<dist_t>(l2space, maxElements, M, efConstruction, random_seed, allow_replace_deleted);
         index_inited = true;
         ep_added = false;
         appr_alg->ef_ = default_ef;
@@ -224,12 +224,12 @@ class Index {
     }
 
 
-    void loadIndex(const std::string &path_to_index, size_t max_elements, bool replace_deleted) {
+    void loadIndex(const std::string &path_to_index, size_t max_elements, bool allow_replace_deleted) {
       if (appr_alg) {
           std::cerr << "Warning: Calling load_index for an already inited index. Old index is being deallocated." << std::endl;
           delete appr_alg;
       }
-      appr_alg = new hnswlib::HierarchicalNSW<dist_t>(l2space, path_to_index, false, max_elements, replace_deleted);
+      appr_alg = new hnswlib::HierarchicalNSW<dist_t>(l2space, path_to_index, false, max_elements, allow_replace_deleted);
       cur_l = appr_alg->cur_element_count;
       index_inited = true;
     }
@@ -415,7 +415,7 @@ class Index {
             "ef"_a = appr_alg->ef_,
             "has_deletions"_a = (bool)appr_alg->num_deleted_,
             "size_links_per_element"_a = appr_alg->size_links_per_element_,
-            "replace_deleted"_a = appr_alg->replace_deleted_,
+            "allow_replace_deleted"_a = appr_alg->allow_replace_deleted_,
 
             "label_lookup_external"_a = py::array_t<hnswlib::labeltype>(
                 { appr_alg->label_lookup_.size() },  // shape
@@ -578,19 +578,19 @@ class Index {
         }
 
         // process deleted elements
-        bool replace_deleted = false;
-        if (d.contains("replace_deleted")) {
-            replace_deleted = d["replace_deleted"].cast<bool>();
+        bool allow_replace_deleted = false;
+        if (d.contains("allow_replace_deleted")) {
+            allow_replace_deleted = d["allow_replace_deleted"].cast<bool>();
         }
-        appr_alg->replace_deleted_= replace_deleted;
+        appr_alg->allow_replace_deleted_= allow_replace_deleted;
 
         appr_alg->num_deleted_ = 0;
         bool has_deletions = d["has_deletions"].cast<bool>();
         if (has_deletions) {
             for (size_t i = 0; i < appr_alg->cur_element_count; i++) {
                 if (appr_alg->isMarkedDeleted(i)) {
                     appr_alg->num_deleted_ += 1;
-                    if (replace_deleted) appr_alg->deleted_elements.insert(i);
+                    if (allow_replace_deleted) appr_alg->deleted_elements.insert(i);
                 }
             }
         }
@@ -886,7 +886,7 @@ PYBIND11_PLUGIN(hnswlib) {
             py::arg("M") = 16,
             py::arg("ef_construction") = 200,
             py::arg("random_seed") = 100,
-            py::arg("replace_deleted") = false)
+            py::arg("allow_replace_deleted") = false)
         .def("knn_query",
             &Index<float>::knnQuery_return_numpy,
             py::arg("data"),
@@ -908,7 +908,7 @@ PYBIND11_PLUGIN(hnswlib) {
             &Index<float>::loadIndex,
             py::arg("path_to_index"),
             py::arg("max_elements") = 0,
-            py::arg("replace_deleted") = false)
+            py::arg("allow_replace_deleted") = false)
         .def("mark_deleted", &Index<float>::markDeleted, py::arg("label"))
         .def("unmark_deleted", &Index<float>::unmarkDeleted, py::arg("label"))
         .def("resize_index", &Index<float>::resizeIndex, py::arg("new_size"))

python_bindings/tests/bindings_test_replace.py

@@ -40,7 +40,7 @@ def testRandomSelf(self):
 
         # Declaring index
         hnsw_index = hnswlib.Index(space='l2', dim=dim)
-        hnsw_index.init_index(max_elements=max_num_elements, ef_construction=200, M=16, replace_deleted=True)
+        hnsw_index.init_index(max_elements=max_num_elements, ef_construction=200, M=16, allow_replace_deleted=True)
 
         hnsw_index.set_ef(100)
         hnsw_index.set_num_threads(4)
@@ -106,7 +106,7 @@ def testRandomSelf(self):
         hnsw_index = hnswlib.Index(space='l2', dim=dim)  # the space can be changed - keeps the data, alters the distance function.
         hnsw_index.set_num_threads(4)
         print(f"Loading index from {index_path}")
-        hnsw_index.load_index(index_path, max_elements=max_num_elements, replace_deleted=True)
+        hnsw_index.load_index(index_path, max_elements=max_num_elements, allow_replace_deleted=True)
 
         # Insert batch 4
         print("Inserting batch 4 by replacing deleted elements")

python_bindings/tests/bindings_test_stress_mt_replace.py

@@ -31,7 +31,7 @@ def testRandomSelf(self):
         # Declaring index
         for _ in range(100):
             hnsw_index = hnswlib.Index(space='l2', dim=dim)
-            hnsw_index.init_index(max_elements=max_num_elements, ef_construction=200, M=16, replace_deleted=True)
+            hnsw_index.init_index(max_elements=max_num_elements, ef_construction=200, M=16, allow_replace_deleted=True)
 
             hnsw_index.set_ef(100)
             hnsw_index.set_num_threads(50)