[ENH] Break out the integrity check and fix obo

Chasing a bug in the integrity check.  Prior to Monday, it was assert(data[j] > 0);
This checks and throws if data[j] <= 0.

CMakeLists.txt

@@ -62,5 +62,7 @@ if(CMAKE_PROJECT_NAME STREQUAL PROJECT_NAME)
 
     add_executable(api_tests tests/cpp/api_test.cpp)
     target_link_libraries(api_tests hnswlib)
-endif()
 
+    add_executable(hnsw-fsck hnsw-fsck.cpp)
+    target_link_libraries(hnsw-fsck hnswlib)
+endif()

hnsw-fsck.cpp

@@ -0,0 +1,40 @@
+#include <iostream>
+#include "hnswlib/hnswlib.h"
+
+int main(int argc, const char* argv[]) {
+    if (argc != 4) {
+        std::cerr << "USAGE: hnsw-fsck <index_path> <space_name> <dims>\n";
+        return 1;
+    }
+    std::string index_path(argv[1]);
+    std::string space_name(argv[2]);
+    int dim = atoi(argv[3]);
+    std::string index_file = index_path;
+    hnswlib::SpaceInterface<float> *l2space;
+    bool normalize = false;
+
+    if (space_name == "l2")
+    {
+        l2space = new hnswlib::L2Space(dim);
+        normalize = false;
+    }
+    else if (space_name == "ip")
+    {
+        l2space = new hnswlib::InnerProductSpace(dim);
+        // For IP, we expect the vectors to be normalized
+        normalize = false;
+    }
+    else if (space_name == "cosine")
+    {
+        l2space = new hnswlib::InnerProductSpace(dim);
+        normalize = true;
+    }
+    else
+    {
+        std::cerr << "Unknown space name: " << space_name << std::endl;
+        return 2;
+    }
+
+    auto appr_alg = new hnswlib::HierarchicalNSW<float>(l2space, index_file, false, 0, false/*allow_replace_deleted*/, normalize, true /*is_persistent_index*/);
+    appr_alg->checkIntegrity();
+}

hnswlib/hnswalg.h

@@ -9,6 +9,7 @@
 #include <unordered_set>
 #include <set>
 #include <list>
+#include <sstream>
 
 namespace hnswlib
 {
@@ -1835,8 +1836,10 @@ namespace hnswlib
                     std::unordered_set<tableint> s;
                     for (int j = 0; j < size; j++)
                     {
-                        if (data[j] < 0 || data[j] >= cur_element_count || data[j] == i)
-                            throw std::runtime_error("HNSW Integrity failure: invalid neighbor index");
+                        if (data[j] >= cur_element_count)
+                            throw std::runtime_error("HNSW Integrity failure: invalid neighbor index data[j] >= cur_element_count");
+                        if (data[j] == i)
+                            throw std::runtime_error("HNSW Integrity failure: invalid neighbor index data[j] == i");
                         inbound_connections_num[data[j]]++;
                         s.insert(data[j]);
                         connections_checked++;
@@ -1850,8 +1853,14 @@ namespace hnswlib
                 int min1 = inbound_connections_num[0], max1 = inbound_connections_num[0];
                 for (int i = 0; i < cur_element_count; i++)
                 {
+                    /*
                     // This should always be true regardless the data is corrupted or not
-                    assert(inbound_connections_num[i] > 0);
+                    if (inbound_connections_num[i] <= 0) {
+                        std::ostringstream ostr;
+                        ostr << "HNSW Integrity failure: inbound_connections_num[" << i << "] = " << inbound_connections_num[i] << " <= 0";
+                        throw std::runtime_error(ostr.str());
+                    }
+                    */
                     min1 = std::min(inbound_connections_num[i], min1);
                     max1 = std::max(inbound_connections_num[i], max1);
                 }

tests/cpp/updates_test.cpp

@@ -250,6 +250,7 @@ int main(int argc, char **argv)
     // Adding enterpoint:
 
     appr_alg.addPoint((void *)dummy_batch.data(), (size_t)0);
+    appr_alg.checkIntegrity();
 
     StopW stopw = StopW();