Docs: C++ docstrings & param names

cpp/README.md

@@ -164,7 +164,7 @@ As in many other bulk-processing APIs, the `executor` and `progress` are optiona
 In its high-level interface, USearch supports a variety of metrics, including the most popular ones:
 
 - `metric_cos_gt<scalar_t>` for "Cosine" or "Angular" distance.
-- `metric_ip_gt<scalar_t>` for "Inner Product" or "Dot Product" distance.
+- `metric_ip_gt<scalar_t>` for "Inner Product" or "Dot Product" distance with normalized vectors.
 - `metric_l2sq_gt<scalar_t>` for the squared "L2" or "Euclidean" distance.
 - `metric_jaccard_gt<scalar_t>` for "Jaccard" distance between two ordered sets of unique elements.
 - `metric_hamming_gt<scalar_t>` for "Hamming" distance, as the number of shared bits in hashes.

include/usearch/index_dense.hpp

@@ -1970,6 +1970,7 @@ class index_dense_gt {
 
         result.computed_distances = computed_distances;
         result.visited_members = visited_members;
+        result.clusters = unique_clusters;
 
         (void)progress;
         return result;

include/usearch/index_plugins.hpp

@@ -1303,6 +1303,8 @@ struct casts_punned_t {
 
 /**
  *  @brief  Inner (Dot) Product distance.
+ *          Vectors should be normalized to unit length,
+ *          otherwise `::metric_cos_gt` should be used instead.
  */
 template <typename scalar_at = float, typename result_at = scalar_at> struct metric_ip_gt {
     using scalar_t = scalar_at;