def main(): centrality_data_folder = sys.argv[1] intervals = load_json(centrality_data_folder) file_prefix = sys.argv[2] out_folder = sys.argv[3] if not os.path.exists(out_folder): os.makedirs(out_folder) day = 0 top_list_prev = [] top_list = [] ret_sort = [] num_nodes = 1 for inter in intervals["centrality_test"]["intervals"]: print("[preproc day = " + str(day) +" ]") if inter["interval"]["graph_stat"]["num_nodes"] != 0: top_list, ret_sort = ccfcr.pre_proc(centrality_data_folder + "/centrality_scores", file_prefix, day) else: print str(day) + ": empty" day+=1 continue if day != 0: full_prev_toplist, full_toplist = ccfcr.proc_kendall(top_list_prev, top_list, ret_sort) write_out_toplists(out_folder + "/" + file_prefix + "_", day, full_prev_toplist, full_toplist) day+=1 top_list_prev = top_list
def main(): baseline_folder = sys.argv[1] # for baseline centrality_folder = sys.argv[2] # for origi file_prefix = sys.argv[3] baseline_type = sys.argv[4] from_interval = int(sys.argv[5]) to_interval = int(sys.argv[6]) out_folder = sys.argv[7] if not os.path.exists(out_folder): os.makedirs(out_folder) intervals = load_json(centrality_folder) top_list_other = [] top_list = [] ret_sort = [] intervals_list = intervals["centrality_test"]["intervals"] for day in range(from_interval, to_interval+1): print("[preproc day = " + str(day) +" ]") if intervals_list[day]["interval"]["graph_stat"]["num_nodes"] != 0: top_list_other, ret_sort_other = ccfcr.pre_proc(baseline_folder, file_prefix + "_baseline" + baseline_type, day) # for baseline top_list, ret_sort = ccfcr.pre_proc(centrality_folder + "/centrality_scores", file_prefix, day) # for origi else: print str(day) + ": empty" day+=1 continue if day != 0: full_prev_toplist, full_toplist = ccfcr.proc_kendall(top_list_other, top_list, ret_sort) write_out_toplists(out_folder + "/" + file_prefix + "_", day, full_prev_toplist, full_toplist)
def test_kendall_all(): prev_data_full, curr_data_full = ccfcr.proc_kendall(prev_data, curr_data_d, sorted_ids_d) p_res =kmc.kendall_all(prev_data_full, curr_data_full) c_res = ccfcr.kendall(prev_data, curr_data_d, sorted_ids_d) assert abs(c_res[0] - p_res[0]) < epsilon assert abs(c_res[1] - p_res[1]) < epsilon