du_points = DistributedInMemoryDataUnit("Points", pilot=pilot) du_points.load(points) f = open("centers.csv") centers = f.readlines() f.close() du_centers = DistributedInMemoryDataUnit("Centers") du_centers.load(centers) number_of_centroids_points=len(centers) end_data_load = time.time() time_measures["DataLoadTime"] = end_data_load-end_start_pilot for iteration in range(0,NUM_ITERATIONS): iteration_start = time.time() future = du_points.map_pilot("KMeans.closestPoint", du_centers.name, number_of_compute_units=2) output_dus = future.result() new_centers = [] for du in output_dus: future=du.reduce_pilot("KMeans.averagePoints") result_du = future.result() new_centers.append(result_du) du_centers = DistributedInMemoryDataUnit("Centers-%d"%(iteration+1)).merge(new_centers) iteration_end = time.time() time_measures["Iteration-%d"%iteration] = iteration_end - iteration_start end = time.time() time_measures["Runtime"] = end-start #################################################################################################################