Load pre-trained pivot table

* @brief Gets the number of threads for IO operations. * * This function retrieves the number of threads to use for disk index IO during the

NOTE :: good efficiency when total_vec_size is integral multiple of 64

push back

Read disk graph data

Asynchronously queue a read request from a file into a buffer slice. Wraps the unsafe Windows API function `ReadFile`, making it safe to call only wh

Register the io context for a thread if it hasn't been registered.

Reserve capacity

Run Bruteforce index tests

Run HNSW index tests

Run HNSW index tests

Run HNSW index tests

Run Lloyds until max_reps or stopping criterion If you pass NULL for closest_docs and closest_center, it will NOT return the results, else it will ass

Save the data on a file.

Save data to file # Arguments `filename` - filename where the data is `data` - information data `npts` - number of points `ndims` - point dimension `a

Save the graph index on a file as an adjacency list. For each point, first store the number of neighbors, and then the neighbor list (each as 4 byte u

Assume memory allocated for pivot_data as new float[num_centers * dim] and select randomly num_centers points as pivots

Utility method to determine whether the provided dtype is supported by `diskannpy`, and if so, the canonical dtype we will use internally (e.

register for notifications