Two main changes:

1. searchKnn will fist check if the graph is empty
2. searchKnn will return a min-heap

The test code in sift_1b is changed and tested.

Author: orrorcol

hnswlib/bruteforce.h

@@ -84,7 +84,10 @@ namespace hnswlib {
         }
 
 
-        std::priority_queue<std::pair<dist_t, labeltype >> searchKnn(const void *query_data, size_t k) const {
+        retType<dist_t> searchKnn(const void *query_data, size_t k) const {
+            retType<dist_t> result;
+            if (cur_element_count == 0) return result;
+
             std::priority_queue<std::pair<dist_t, labeltype >> topResults;
             for (int i = 0; i < k; i++) {
                 dist_t dist = fstdistfunc_(query_data, data_ + size_per_element_ * i, dist_func_param_);
@@ -103,7 +106,12 @@ namespace hnswlib {
                 }
 
             }
-            return topResults;
+            while (!topResults.empty()) {
+                auto each = topResults.top();
+                result.push(each);
+                topResults.pop();
+            }
+            return result;
         };
 
         void saveIndex(const std::string &location) {

hnswlib/hnswalg.h

@@ -895,7 +895,10 @@ namespace hnswlib {
             return cur_c;
         };
 
-        std::priority_queue<std::pair<dist_t, labeltype >> searchKnn(const void *query_data, size_t k) const {
+        retType<dist_t> searchKnn(const void *query_data, size_t k) const {
+            retType<dist_t> result;
+            if (cur_element_count == 0) return result;
+
             tableint currObj = enterpoint_node_;
             dist_t curdist = fstdistfunc_(query_data, getDataByInternalId(enterpoint_node_), dist_func_param_);
 
@@ -934,16 +937,15 @@ namespace hnswlib {
                         currObj, query_data, std::max(ef_, k));
                 top_candidates.swap(top_candidates1);
             }
-            std::priority_queue<std::pair<dist_t, labeltype >> results;
             while (top_candidates.size() > k) {
                 top_candidates.pop();
             }
             while (top_candidates.size() > 0) {
                 std::pair<dist_t, tableint> rez = top_candidates.top();
-                results.push(std::pair<dist_t, labeltype>(rez.first, getExternalLabel(rez.second)));
+                result.push(std::pair<dist_t, labeltype>(rez.first, getExternalLabel(rez.second)));
                 top_candidates.pop();
             }
-            return results;
+            return result;
         };
 
 

hnswlib/hnswlib.h

@@ -24,12 +24,29 @@
 #endif
 
 #include <queue>
+#include <vector>
 
 #include <string.h>
 
 namespace hnswlib {
     typedef size_t labeltype;
 
+    template <typename T>
+    class pairGreater {
+    public:
+        bool operator()(const T& p1, const T& p2) {
+            return p1.first > p2.first;
+        }
+    };
+
+    template <typename T>
+    using retType = std::priority_queue<
+                std::pair<T, labeltype>,
+                std::vector<std::pair<T, labeltype>>,
+                pairGreater<std::pair<T, labeltype>>
+            >;
+
+
     template<typename T>
     static void writeBinaryPOD(std::ostream &out, const T &podRef) {
         out.write((char *) &podRef, sizeof(T));
@@ -61,7 +78,7 @@ namespace hnswlib {
     class AlgorithmInterface {
     public:
         virtual void addPoint(const void *datapoint, labeltype label)=0;
-        virtual std::priority_queue<std::pair<dist_t, labeltype >> searchKnn(const void *, size_t) const = 0;
+        virtual retType<dist_t> searchKnn(const void *, size_t) const = 0;
         virtual void saveIndex(const std::string &location)=0;
         virtual ~AlgorithmInterface(){
         }

sift_1b.cpp

@@ -147,10 +147,10 @@ static size_t getCurrentRSS() {
 
 static void
 get_gt(unsigned int *massQA, unsigned char *massQ, unsigned char *mass, size_t vecsize, size_t qsize, L2SpaceI &l2space,
-       size_t vecdim, vector<std::priority_queue<std::pair<int, labeltype >>> &answers, size_t k) {
+       size_t vecdim, vector<retType<int>> &answers, size_t k) {
 
 
-    (vector<std::priority_queue<std::pair<int, labeltype >>>(qsize)).swap(answers);
+    (vector<retType<int>>(qsize)).swap(answers);
     DISTFUNC<int> fstdistfunc_ = l2space.get_dist_func();
     cout << qsize << "\n";
     for (int i = 0; i < qsize; i++) {
@@ -162,15 +162,15 @@ get_gt(unsigned int *massQA, unsigned char *massQ, unsigned char *mass, size_t v
 
 static float
 test_approx(unsigned char *massQ, size_t vecsize, size_t qsize, HierarchicalNSW<int> &appr_alg, size_t vecdim,
-            vector<std::priority_queue<std::pair<int, labeltype >>> &answers, size_t k) {
+            vector<retType<int>> &answers, size_t k) {
     size_t correct = 0;
     size_t total = 0;
     //uncomment to test in parallel mode:
     //#pragma omp parallel for
     for (int i = 0; i < qsize; i++) {
 
-        std::priority_queue<std::pair<int, labeltype >> result = appr_alg.searchKnn(massQ + vecdim * i, k);
-        std::priority_queue<std::pair<int, labeltype >> gt(answers[i]);
+        retType<int> result = appr_alg.searchKnn(massQ + vecdim * i, k);
+        retType<int> gt(answers[i]);
         unordered_set<labeltype> g;
         total += gt.size();
 
@@ -196,7 +196,7 @@ test_approx(unsigned char *massQ, size_t vecsize, size_t qsize, HierarchicalNSW<
 
 static void
 test_vs_recall(unsigned char *massQ, size_t vecsize, size_t qsize, HierarchicalNSW<int> &appr_alg, size_t vecdim,
-               vector<std::priority_queue<std::pair<int, labeltype >>> &answers, size_t k) {
+               vector<retType<int>> &answers, size_t k) {
     vector<size_t> efs;// = { 10,10,10,10,10 };
     for (int i = k; i < 30; i++) {
         efs.push_back(i);
@@ -231,7 +231,7 @@ inline bool exists_test(const std::string &name) {
 void sift_test1B() {
 
 
-	int subset_size_milllions = 200;
+	int subset_size_milllions = 1;
 	int efConstruction = 40;
 	int M = 16;
 
@@ -351,7 +351,7 @@ void sift_test1B() {
     }
 
 
-    vector<std::priority_queue<std::pair<int, labeltype >>> answers;
+    vector<retType<int>> answers;
     size_t k = 1;
     cout << "Parsing gt:\n";
     get_gt(massQA, massQ, mass, vecsize, qsize, l2space, vecdim, answers, k);