import json import pandas as pd file_names = [#'./llm_for_code/datasets/codecontests/CN_py800_catmix_5rand_4fold_full_oripairs_matched.json' # './dataset_lfs/CC_900_py_catmix_5rand_4fold_oripairs_temp.json', # './dataset_lfs/CN_py800_catmix_5rand_4fold_200_oripairs_temp.json', # './dataset_lfs/CN_py800_catmix_5rand_4fold_200_7m2_oripairs_temp.json', # './dataset_lfs/CN_py800_catmix_5rand_4fold_200_7m3_oripairs_temp.json', # './dataset_lfs/CN_py800_catmix_5rand_4fold_200_7m4_oripairs_temp.json', # './dataset_lfs/CC_900_py_catmix_5rand_4fold_0_2_oripairs.json', # './dataset_lfs/CC_900_py_catmix_5rand_4fold_101_3_oripairs_temp.json', # './dataset_lfs/CC_900_py_catmix_5rand_4fold_101_3_rem_oripairs_temp.json', # './dataset_lfs/CC_900_py_catmix_5rand_4fold_101_3_rem2_oripairs_temp.json', # './dataset_lfs/CC_900_py_catmix_5rand_4fold_200_0_oripairs_temp.json', # './dataset_lfs/CC_900_py_catmix_5rand_4fold_200_101_2_oripairs_temp.json', # './dataset_lfs/CC_900_py_catmix_5rand_4fold_200_101_oripairs_temp.json', # './dataset_lfs/CC_900_py_catmix_5rand_4fold_200_110_oripairs.json', # './dataset_lfs/CC_900_py_catmix_5rand_4fold_1s_1k_oripairs_temp.json', # './dataset_lfs/CC_900_py_catmix_5rand_4fold_1s_1k_2_oripairs_temp.json', # './dataset_lfs/CC_900_py_catmix_5rand_4fold_1s_1k_3_oripairs_temp.json', # './dataset_lfs/CC_900_py_catmix_5rand_4fold_100_4_oripairs_temp.json', #"./dataset_lfs/allpairs_data_val_900_unperturbed.json", './dataset_lfs/CC_900_py_catmix_5rand_4fold_train_all_2_oripairs_temp.json', './dataset_lfs/CC_900_py_catmix_5rand_4fold_train_all_3_oripairs_temp.json', './dataset_lfs/CC_900_py_catmix_5rand_4fold_train_all_4_oripairs_temp.json', './dataset_lfs/CC_900_py_catmix_5rand_4fold_train_all_5_oripairs_temp.json', './dataset_lfs/CC_900_py_catmix_5rand_4fold_train_all_oripairs_temp.json', './dataset_lfs/CC_900_py_catmix_5rand_4fold_train_0_conc_1_oripairs_temp.json', './dataset_lfs/CC_900_py_catmix_5rand_4fold_train_0_conc_oripairs_temp.json', './dataset_lfs/CC_900_py_catmix_5rand_4fold_train_101_oripairs_temp.json', './dataset_lfs/CC_900_py_catmix_5rand_4fold_train_2_oripairs.json', './dataset_lfs/CC_900_py_catmix_5rand_4fold_train_all_conc_2_oripairs.json', './dataset_lfs/CC_900_py_catmix_5rand_4fold_train_all_conc_3_oripairs.json', #'./dataset_lfs/CC_900_py_catmix_5rand_4fold_train_all_conc_2_oripairs_temp.json', './dataset_lfs/CC_900_py_catmix_5rand_4fold_train_3_oripairs.json', './dataset_lfs/CC_900_py_catmix_5rand_4fold_train_all_conc_500_1_oripairs_temp.json', './dataset_lfs/CC_900_py_catmix_5rand_4fold_train_all_conc_500_2_oripairs_temp.json', './dataset_lfs/CC_900_py_catmix_5rand_4fold_train_0s_conc_1_oripairs.json', './dataset_lfs/CC_900_py_catmix_5rand_4fold_train_0_conc_1_oripairs_temp.json', './dataset_lfs/CC_900_py_catmix_5rand_4fold_train_0_conc_oripairs_temp.json', './dataset_lfs/CC_900_py_catmix_5rand_4fold_train_all_conc_0s_500_0_oripairs.json', './dataset_lfs/CC_900_py_catmix_5rand_4fold_train_conc_pos_2_oripairs_temp.json', './dataset_lfs/CC_900_py_catmix_5rand_4fold_train_conc_pos_oripairs_temp.json', './dataset_lfs/CC_900_py_catmix_5rand_4fold_train_all_5_oripairs_temp.json', './dataset_lfs/CC_900_py_catmix_5rand_4fold_train_all_conc_1_oripairs_temp.json', './dataset_lfs/CC_900_py_catmix_5rand_4fold_train_all_conc_10s_500_0_oripairs_temp.json', './dataset_lfs/CC_900_py_catmix_5rand_4fold_train_all_conc_10s_500_2_oripairs_temp.json', './dataset_lfs/CC_900_py_catmix_5rand_4fold_train_all_conc_2_oripairs.json', './dataset_lfs/CC_900_py_catmix_5rand_4fold_train_all_conc_3_oripairs.json', './dataset_lfs/CC_900_py_catmix_5rand_4fold_train_all_conc_oripairs_temp.json', './dataset_lfs/CC_900_py_catmix_5rand_4fold_train_conc_5_oripairs_temp.json', ] file_names_per = [#'./dataset_lfs/CC_900_py_catmix_5rand_4fold_pairs_0_temp.json', # './dataset_lfs/CN_py800_catmix_5rand_4fold_200_pairs_0_temp.json', # './dataset_lfs/CN_py800_catmix_5rand_4fold_200_7m2_pairs_0_temp.json', # './dataset_lfs/CN_py800_catmix_5rand_4fold_200_7m3_pairs_0_temp.json', # './dataset_lfs/CN_py800_catmix_5rand_4fold_200_7m4_pairs_0_temp.json', # './dataset_lfs/CC_900_py_catmix_5rand_4fold_0_2_pairs_0.json', # './dataset_lfs/CC_900_py_catmix_5rand_4fold_101_3_pairs_0_temp.json', # './dataset_lfs/CC_900_py_catmix_5rand_4fold_101_3_rem_pairs_0_temp.json', # './dataset_lfs/CC_900_py_catmix_5rand_4fold_101_3_rem2_pairs_0_temp.json', # './dataset_lfs/CC_900_py_catmix_5rand_4fold_200_0_pairs_0_temp.json', # './dataset_lfs/CC_900_py_catmix_5rand_4fold_200_101_2_pairs_0_temp.json', # './dataset_lfs/CC_900_py_catmix_5rand_4fold_200_101_pairs_0_temp.json', # './dataset_lfs/CC_900_py_catmix_5rand_4fold_200_110_pairs_0.json', # './dataset_lfs/CC_900_py_catmix_5rand_4fold_1s_1k_pairs_0_temp.json', # './dataset_lfs/CC_900_py_catmix_5rand_4fold_1s_1k_2_pairs_0_temp.json', # './dataset_lfs/CC_900_py_catmix_5rand_4fold_1s_1k_3_pairs_0_temp.json', # './dataset_lfs/CC_900_py_catmix_5rand_4fold_100_4_pairs_0_temp.json', './dataset_lfs/CC_900_py_catmix_5rand_4fold_train_all_2_pairs_0_temp.json', './dataset_lfs/CC_900_py_catmix_5rand_4fold_train_all_3_pairs_0_temp.json', './dataset_lfs/CC_900_py_catmix_5rand_4fold_train_all_4_pairs_0_temp.json', './dataset_lfs/CC_900_py_catmix_5rand_4fold_train_all_5_pairs_0_temp.json', './dataset_lfs/CC_900_py_catmix_5rand_4fold_train_all_pairs_0_temp.json', './dataset_lfs/CC_900_py_catmix_5rand_4fold_train_0_conc_1_pairs_0_temp.json', './dataset_lfs/CC_900_py_catmix_5rand_4fold_train_0_conc_pairs_0_temp.json', './dataset_lfs/CC_900_py_catmix_5rand_4fold_train_101_pairs_0_temp.json', './dataset_lfs/CC_900_py_catmix_5rand_4fold_train_2_pairs_0.json', './dataset_lfs/CC_900_py_catmix_5rand_4fold_train_all_conc_2_pairs_0.json', './dataset_lfs/CC_900_py_catmix_5rand_4fold_train_all_conc_3_pairs_0.json', #'./dataset_lfs/CC_900_py_catmix_5rand_4fold_train_all_conc_2_pairs_0_temp.json', './dataset_lfs/CC_900_py_catmix_5rand_4fold_train_3_pairs_0.json', './dataset_lfs/CC_900_py_catmix_5rand_4fold_train_all_conc_500_1_pairs_0_temp.json', './dataset_lfs/CC_900_py_catmix_5rand_4fold_train_all_conc_500_2_pairs_0_temp.json', './dataset_lfs/CC_900_py_catmix_5rand_4fold_train_0s_conc_1_pairs_0.json', './dataset_lfs/CC_900_py_catmix_5rand_4fold_train_0_conc_1_pairs_0_temp.json', './dataset_lfs/CC_900_py_catmix_5rand_4fold_train_0_conc_pairs_0_temp.json', './dataset_lfs/CC_900_py_catmix_5rand_4fold_train_all_conc_0s_500_0_pairs_0.json', './dataset_lfs/CC_900_py_catmix_5rand_4fold_train_conc_pos_2_pairs_0_temp.json', './dataset_lfs/CC_900_py_catmix_5rand_4fold_train_conc_pos_pairs_0_temp.json', './dataset_lfs/CC_900_py_catmix_5rand_4fold_train_all_5_pairs_0_temp.json', './dataset_lfs/CC_900_py_catmix_5rand_4fold_train_all_conc_1_pairs_0_temp.json', './dataset_lfs/CC_900_py_catmix_5rand_4fold_train_all_conc_10s_500_0_pairs_0_temp.json', './dataset_lfs/CC_900_py_catmix_5rand_4fold_train_all_conc_10s_500_2_pairs_0_temp.json', './dataset_lfs/CC_900_py_catmix_5rand_4fold_train_all_conc_2_pairs_0.json', './dataset_lfs/CC_900_py_catmix_5rand_4fold_train_all_conc_3_pairs_0.json', './dataset_lfs/CC_900_py_catmix_5rand_4fold_train_all_conc_pairs_0_temp.json', './dataset_lfs/CC_900_py_catmix_5rand_4fold_train_conc_5_pairs_0_temp.json', ] #file_names = [ './dataset_lfs/CC_900_py_catmix_5rand_4fold_test_varc2_oripairs_temp.json', # #'./dataset_lfs/CC_900_py_catmix_5rand_4fold_val_oripairs.json' # './dataset_lfs/CC_900_py_catmix_5rand_4fold_test_varc_oripairs_temp.json', # './llm_for_code/datasets/codecontests/CC_900_py_catmix_5rand_4fold_test_varc0s1_oripairs.json', # './dataset_lfs/CC_900_py_catmix_5rand_4fold_test_varc10s_oripairs_temp.json', #'./dataset_lfs/CC_900_py_catmix_5rand_4fold_train_all_2_pairs_0_temp_fulltrydup_train.json' #] # file_names = [ 'CC_1200_py_catmix_5rand_4fold_test_1s_v_oripairs.json', # 'CC_1200_py_catmix_5rand_4fold_test_all_v_oripairs.json', # 'CC_1200_py_catmix_5rand_4fold_test_all_v2_oripairs.json', # 'CC_1200_py_catmix_5rand_4fold_test_pos1s_v_oripairs.json', # 'CC_1200_py_catmix_5rand_4fold_test_v2_1s_oripairs_temp.json' # ] # file_names_per = [ 'CC_1200_py_catmix_5rand_4fold_test_1s_v_pairs_0.json', # 'CC_1200_py_catmix_5rand_4fold_test_all_v_pairs_0.json', # 'CC_1200_py_catmix_5rand_4fold_test_all_v2_pairs_0.json', # 'CC_1200_py_catmix_5rand_4fold_test_pos1s_v_pairs_0.json', # 'CC_1200_py_catmix_5rand_4fold_test_v2_1s_pairs_0_temp.json' # ] # file_names = [ 'CC_2200_py_catmix_5rand_1fold_test_1s_5_oripairs.json', # 'CC_2200_py_catmix_5rand_1fold_test_all_2_oripairs.json', # 'CC_2200_py_catmix_5rand_1fold_test_all_3_oripairs.json', # 'CC_2200_py_catmix_5rand_1fold_test_all_4_oripairs.json', # 'CC_2200_py_catmix_5rand_1fold_test_all_oripairs.json', # 'CC_2200_py_catmix_5rand_1fold_test_ls_2_oripairs.json', # 'CC_2200_py_catmix_5rand_1fold_test_ls_oripairs.json', # 'CC_2200_py_catmix_5rand_1fold_test_0s_2_oripairs.json', # 'CC_2200_py_catmix_5rand_1fold_test_0s_neg_oripairs.json', # 'CC_2200_py_catmix_5rand_1fold_test_0s_neg2_oripairs.json', # 'CC_2200_py_catmix_5rand_1fold_test_0s_neg22_oripairs.json', # 'CC_2200_py_catmix_5rand_1fold_test_0s_oripairs.json', # 'CC_2200_py_catmix_5rand_1fold_test_0s_pos_oripairs.json', # ] # file_names_per = [ 'CC_2200_py_catmix_5rand_1fold_test_1s_5_pairs_0.json', # 'CC_2200_py_catmix_5rand_1fold_test_all_2_pairs_0.json', # 'CC_2200_py_catmix_5rand_1fold_test_all_3_pairs_0.json', # 'CC_2200_py_catmix_5rand_1fold_test_all_4_pairs_0.json', # 'CC_2200_py_catmix_5rand_1fold_test_all_pairs_0.json', # 'CC_2200_py_catmix_5rand_1fold_test_ls_2_pairs_0.json', # 'CC_2200_py_catmix_5rand_1fold_test_ls_pairs_0.json', # 'CC_2200_py_catmix_5rand_1fold_test_0s_2_pairs_0.json', # 'CC_2200_py_catmix_5rand_1fold_test_0s_neg_pairs_0.json', # 'CC_2200_py_catmix_5rand_1fold_test_0s_neg2_pairs_0.json', # 'CC_2200_py_catmix_5rand_1fold_test_0s_neg22_pairs_0.json', # 'CC_2200_py_catmix_5rand_1fold_test_0s_pairs_0.json', # 'CC_2200_py_catmix_5rand_1fold_test_0s_pos_pairs_0.json', # ] #file_names_per = [ './dataset_lfs/CC_900_py_catmix_5rand_4fold_test_varc2_pairs_0_temp.json', # # './dataset_lfs/CC_900_py_catmix_5rand_4fold_val_pairs_0.json' # './dataset_lfs/CC_900_py_catmix_5rand_4fold_test_varc_pairs_0_temp.json', # './llm_for_code/datasets/codecontests/CC_900_py_catmix_5rand_4fold_test_varc0s1_pairs_0.json', # './dataset_lfs/CC_900_py_catmix_5rand_4fold_test_varc10s_pairs_0_temp.json', #'./dataset_lfs/CC_900_py_catmix_5rand_4fold_train_all_2_pairs_0_temp_fulltrydup_train.json' #] # file_names = [ # 'CC_1600_py_catmix_5rand_4fold_test_all_v2_oripairs.json', # 'CC_1600_py_catmix_5rand_4fold_test_all_oripairs_temp.json', # 'CC_1600_py_catmix_5rand_4fold_test_1s_oripairs.json', # 'CC_1600_py_catmix_5rand_4fold_test_pos1s_v_oripairs.json', # 'CC_1600_py_catmix_5rand_4fold_test_pos1s_oripairs.json', # 'CC_1600_py_catmix_5rand_4fold_test_neg1s_v_oripairs.json', # 'CC_1600_py_catmix_5rand_4fold_test_neg1s_oripairs_temp.json', # 'CC_1600_py_catmix_5rand_4fold_test_neg1s_2_oripairs_temp.json', #] # file_names_per = [ # 'CC_1600_py_catmix_5rand_4fold_test_all_v2_pairs_0.json', # 'CC_1600_py_catmix_5rand_4fold_test_all_pairs_0_temp.json', # 'CC_1600_py_catmix_5rand_4fold_test_1s_pairs_0.json', # 'CC_1600_py_catmix_5rand_4fold_test_pos1s_v_pairs_0.json', # 'CC_1600_py_catmix_5rand_4fold_test_pos1s_pairs_0.json', # 'CC_1600_py_catmix_5rand_4fold_test_neg1s_v_pairs_0.json', # 'CC_1600_py_catmix_5rand_4fold_test_neg1s_pairs_0_temp.json', # 'CC_1600_py_catmix_5rand_4fold_test_neg1s_2_pairs_0_temp.json', # ] #file_names = ['CC_900_py_catmix_5rand_1fold_conctest_all_ind_oripairs.json', 'CC_900_py_catmix_5rand_1fold_conctest_1s_ind_oripairs.json',] #file_names_per = ['CC_900_py_catmix_5rand_1fold_conctest_all_ind_pairs_0.json' ,'CC_900_py_catmix_5rand_1fold_conctest_1s_ind_pairs_0.json'] # file_names = ['CC_900_py_catmix_5rand_1fold_conctest_1s_loop_oripairs_temp.json', # 'CC_900_py_catmix_5rand_1fold_conctest_all_loop_oripairs_temp.json'] # file_names_per = ['CC_900_py_catmix_5rand_1fold_conctest_1s_loop_pairs_0_temp.json', # 'CC_900_py_catmix_5rand_1fold_conctest_all_loop_pairs_0_temp.json',] # file_names = ['CC_900_py_catmix_5rand_1fold_conctest_boolinv_1s_oripairs.json', # 'CC_900_py_catmix_5rand_1fold_conctest_boolinv_1s2_oripairs.json', # 'CC_900_py_catmix_5rand_1fold_conctest_boolinv_1s3_oripairs.json', # 'CC_900_py_catmix_5rand_1fold_conctest_boolinv_all_oripairs.json', # 'CC_900_py_catmix_5rand_1fold_conctest_boolinv_all2_oripairs.json'] # file_names_per = ['CC_900_py_catmix_5rand_1fold_conctest_boolinv_1s_pairs_0.json', # 'CC_900_py_catmix_5rand_1fold_conctest_boolinv_1s2_pairs_0.json', # 'CC_900_py_catmix_5rand_1fold_conctest_boolinv_1s3_pairs_0.json', # 'CC_900_py_catmix_5rand_1fold_conctest_boolinv_all_pairs_0.json', # 'CC_900_py_catmix_5rand_1fold_conctest_boolinv_all2_pairs_0.json'] # file_names = ['CC_900_py_catmix_exact9rand_1fold_conctest_all_oripairs.json', # & 7 # 'CC_900_py_catmix_exact9rand_1fold_conctest_0s_oripairs.json', # 'CC_900_py_catmix_exact9rand_1fold_conctest_1s_oripairs.json',] # file_names_per = ['CC_900_py_catmix_exact9rand_1fold_conctest_all_pairs_0.json', # 'CC_900_py_catmix_exact9rand_1fold_conctest_0s_pairs_0.json', # 'CC_900_py_catmix_exact9rand_1fold_conctest_1s_pairs_0.json',] # file_names = ['CC_900_py_catmix_exact1rand_1fold_conctest_all_oripairs.json', # & 3 # #'CC_900_py_catmix_exact3rand_1fold_conctest_0s_oripairs.json', # 'CC_900_py_catmix_exact1rand_1fold_conctest_1s_oripairs.json',] # file_names_per = ['CC_900_py_catmix_exact1rand_1fold_conctest_all_pairs_0.json', # #'CC_900_py_catmix_exact3rand_1fold_conctest_0s_pairs_0.json', # 'CC_900_py_catmix_exact1rand_1fold_conctest_1s_pairs_0.json',] # file_names = ['CC_900_py_catmix_exact5rand_1fold_conctest_all_oripairs_temp.json', # & 7 # 'CC_900_py_catmix_exact5rand_1fold_conctest_0s_oripairs.json', # 'CC_900_py_catmix_exact5rand_1fold_conctest_1s_oripairs_temp.json', # 'CC_900_py_catmix_exact1rand_5fold_conctest_1s2_oripairs_temp.json', # 'CC_900_py_catmix_exact1rand_5fold_conctest_all2_oripairs_temp.json',] # file_names_per = ['CC_900_py_catmix_exact5rand_1fold_conctest_all_pairs_0_temp.json', # 'CC_900_py_catmix_exact5rand_1fold_conctest_0s_pairs_0.json', # 'CC_900_py_catmix_exact5rand_1fold_conctest_1s_pairs_0_temp.json', # 'CC_900_py_catmix_exact1rand_5fold_conctest_1s2_pairs_0_temp.json', # 'CC_900_py_catmix_exact1rand_5fold_conctest_all2_pairs_0_temp.json',] # file_names = [ 'CC_900_py_catmix_5rand_1fold_conctest_expr_0s_4_oripairs.json', # 'CC_900_py_catmix_5rand_1fold_conctest_expr_1s_4_oripairs.json', # 'CC_900_py_catmix_5rand_1fold_conctest_expr_all_4_oripairs.json', # ] # file_names_per = ['CC_900_py_catmix_5rand_1fold_conctest_expr_0s_4_pairs_0.json', # 'CC_900_py_catmix_5rand_1fold_conctest_expr_1s_4_pairs_0.json', # 'CC_900_py_catmix_5rand_1fold_conctest_expr_all_4_pairs_0.json', # ] # file_names = ['CC_900_py_catmix_5rand_1fold_train_conctest_loopvarexpr_all_oripairs_temp.json', # 'CC_900_py_catmix_5rand_1fold_train_conctest_loopvarexpr_0s_oripairs.json', # 'CC_900_py_catmix_5rand_1fold_train_conctest_loopvarexpr_1s_oripairs_temp.json', # 'CC_900_py_catmix_5rand_1fold_train_conctest_loopvarexpr_all_2_oripairs_temp.json', # 'CC_900_py_catmix_5rand_1fold_train_conctest_loopvarexpr_all_3_oripairs_temp.json', # 'CC_900_py_catmix_5rand_1fold_train_conctest_loopvarexpr_all_7_oripairs_temp.json'] # file_names_per = ['CC_900_py_catmix_5rand_1fold_train_conctest_loopvarexpr_all_pairs_0_temp.json', # 'CC_900_py_catmix_5rand_1fold_train_conctest_loopvarexpr_0s_pairs_0.json', # 'CC_900_py_catmix_5rand_1fold_train_conctest_loopvarexpr_1s_pairs_0_temp.json', # 'CC_900_py_catmix_5rand_1fold_train_conctest_loopvarexpr_all_2_pairs_0_temp.json', # 'CC_900_py_catmix_5rand_1fold_train_conctest_loopvarexpr_all_3_pairs_0_temp.json', # 'CC_900_py_catmix_5rand_1fold_train_conctest_loopvarexpr_all_7_pairs_0_temp.json'] # file_names = ['CC_900_py_catmix_5rand_1fold_test_conctest_all_sptask_1s_13_oripairs_temp.json', # 'CC_900_py_catmix_5rand_1fold_test_conctest_all_sptask_1s_14_oripairs_temp.json'] # file_names_per = ['CC_900_py_catmix_5rand_1fold_test_conctest_all_sptask_1s_13_pairs_0_temp.json', # 'CC_900_py_catmix_5rand_1fold_test_conctest_all_sptask_1s_14_pairs_0_temp.json'] file_names = ['CC_900_py_catmix_1rand_1fold_train_conctest_all_11_oripairs_temp.json', 'CC_900_py_catmix_1rand_1fold_train_conctest_all_12_oripairs_temp.json', 'CC_900_py_catmix_1rand_1fold_train_conctest_0s_15_oripairs.json', 'CC_900_py_catmix_1rand_1fold_train_conctest_0s_16_oripairs.json',] file_names_per = ['CC_900_py_catmix_1rand_1fold_train_conctest_all_11_pairs_0_temp.json', 'CC_900_py_catmix_1rand_1fold_train_conctest_all_12_pairs_0_temp.json', 'CC_900_py_catmix_1rand_1fold_train_conctest_0s_15_pairs_0.json', 'CC_900_py_catmix_1rand_1fold_train_conctest_0s_16_pairs_0.json',] file_names = ['CC_900_py_catmix_5rand_1fold_train_conctest_iffp_all_2_oripairs.json'] file_names_per = ['CC_900_py_catmix_5rand_1fold_train_conctest_iffp_all_2_pairs_0.json'] file_names = ['AtCoder_D_catmix_5rand_oripairs.json', 'AtCoder_D_catmix_5rand_1s_oripairs.json' ] file_names_per = ['AtCoder_D_catmix_5rand_pairs_0.json', 'AtCoder_D_catmix_5rand_1s_pairs_0.json'] file_names = ['CC_900_py_catmix_5rand_4fold_pyjava_bothper_oripairs.json', 'CC_900_py_catmix_5rand_4fold_pyjava_bothper_1s_3_oripairs.json', 'CC_900_py_catmix_5rand_4fold_pyjava_bothper_1s_oripairs.json', 'CC_900_py_catmix_5rand_4fold_pyjava_bothper_1s_2_oripairs.json', #'CC_900_py_catmix_5rand_4fold_pyjava_bothper_2_oripairs.json', ] file_names_per = ['CC_900_py_catmix_5rand_4fold_pyjava_bothper_pairs_0.json', 'CC_900_py_catmix_5rand_4fold_pyjava_bothper_1s_3_pairs_0.json', 'CC_900_py_catmix_5rand_4fold_pyjava_bothper_1s_pairs_0.json', 'CC_900_py_catmix_5rand_4fold_pyjava_bothper_1s_2_pairs_0.json', #'CC_900_py_catmix_5rand_4fold_pyjava_bothper_2_pairs_0.json', ] file_names = ['CC_900_py_catmix_5rand_4fold_pyjava_bothper_wneg_10s_3_oripairs.json', 'CC_900_py_catmix_5rand_4fold_pyjava_bothper_wneg_10s_2_oripairs_temp.json', 'CC_900_py_catmix_5rand_4fold_pyjava_bothper_wneg_10s_4_oripairs_temp.json', 'CC_900_py_catmix_5rand_4fold_pyjava_bothper_wneg_10s_oripairs_temp.json'] file_names_per = ['CC_900_py_catmix_5rand_4fold_pyjava_bothper_wneg_10s_3_pairs_0.json', 'CC_900_py_catmix_5rand_4fold_pyjava_bothper_wneg_10s_2_pairs_0_temp.json', 'CC_900_py_catmix_5rand_4fold_pyjava_bothper_wneg_10s_4_pairs_0_temp.json', 'CC_900_py_catmix_5rand_4fold_pyjava_bothper_wneg_10s_pairs_0_temp.json'] file_names = ["CC_900_py_catmix_5rand_1fold_pyjava_bothper_all_oripairs_temp.json", "CC_900_py_catmix_5rand_1fold_pyjava_bothper_all_2_oripairs_temp.json", "CC_900_py_catmix_5rand_1fold_pyjava_bothper_1s_oripairs_temp.json", "CC_900_py_catmix_5rand_1fold_pyjava_bothper_1s_2_oripairs_temp.json", "CC_900_py_catmix_5rand_1fold_pyjava_only_1s_50_oripairs_temp.json", "CC_900_py_catmix_5rand_1fold_pyjava_only_11s_oripairs_temp.json"] file_names_per = ["CC_900_py_catmix_5rand_1fold_pyjava_bothper_all_pairs_0_temp.json", "CC_900_py_catmix_5rand_1fold_pyjava_bothper_all_2_pairs_0_temp.json", "CC_900_py_catmix_5rand_1fold_pyjava_bothper_1s_pairs_0_temp.json", "CC_900_py_catmix_5rand_1fold_pyjava_bothper_1s_2_pairs_0_temp.json", "CC_900_py_catmix_5rand_1fold_pyjava_only_1s_50_pairs_0_temp.json", "CC_900_py_catmix_5rand_1fold_pyjava_only_11s_pairs_0_temp.json"] file_names = ["CC_900_py_catmix_5rand_1fold_pycpp_11s_oripairs_temp.json", "CC_900_py_catmix_5rand_1fold_pycpp_1s_2_oripairs_temp.json", "CC_900_py_catmix_5rand_1fold_pycpp_1s_oripairs_temp.json", "CC_900_py_catmix_5rand_1fold_pycpp_all_oripairs_temp.json", "CC_900_py_catmix_5rand_1fold_pycpp_all_2_oripairs_temp.json", "CC_900_py_catmix_5rand_1fold_pycpp_all_3s_oripairs_temp.json", ] file_names_per = ["CC_900_py_catmix_5rand_1fold_pycpp_11s_pairs_0_temp.json", "CC_900_py_catmix_5rand_1fold_pycpp_1s_2_pairs_0_temp.json", "CC_900_py_catmix_5rand_1fold_pycpp_1s_pairs_0_temp.json", "CC_900_py_catmix_5rand_1fold_pycpp_all_pairs_0_temp.json", "CC_900_py_catmix_5rand_1fold_pycpp_all_2_pairs_0_temp.json", "CC_900_py_catmix_5rand_1fold_pycpp_all_3s_pairs_0_temp.json", ] # file_names = ["CC_900_py_catmix_5rand_1fold_java_1s_2_oripairs_temp.json", # "CC_900_py_catmix_5rand_1fold_java_all_oripairs.json", # "CC_900_py_catmix_5rand_1fold_java_only_1s_oripairs_temp.json", # "CC_900_py_catmix_5rand_1fold_java_only_all_oripairs_temp.json"] # file_names_per = ["CC_900_py_catmix_5rand_1fold_java_1s_2_pairs_0_temp.json", # "CC_900_py_catmix_5rand_1fold_java_all_pairs_0.json", # "CC_900_py_catmix_5rand_1fold_java_only_1s_pairs_0_temp.json", # "CC_900_py_catmix_5rand_1fold_java_only_all_pairs_0_temp.json"] # file_names = ['CC_900_py_catmix_5rand_1fold_train_conctest_iffp_all_oripairs_temp.json', # 'CC_900_py_catmix_5rand_1fold_train_conctest_iffp_all_3_oripairs_temp.json', # 'CC_900_py_catmix_5rand_1fold_train_conctest_iffp_all_2_oripairs_temp.json', # 'CC_900_py_catmix_1rand_1fold_train_conctest_0s_15_oripairs_temp.json', # 'CC_900_py_catmix_1rand_1fold_train_conctest_0s_16_oripairs_temp.json'] # file_names_per = ['CC_900_py_catmix_5rand_1fold_train_conctest_iffp_all_pairs_0_temp.json', # 'CC_900_py_catmix_5rand_1fold_train_conctest_iffp_all_3_pairs_0_temp.json', # 'CC_900_py_catmix_5rand_1fold_train_conctest_iffp_all_2_pairs_0_temp.json', # 'CC_900_py_catmix_1rand_1fold_train_conctest_0s_15_pairs_0_temp.json', # 'CC_900_py_catmix_1rand_1fold_train_conctest_0s_16_pairs_0_temp.json'] file1 = './llm_for_code/datasets/codecontests/CC_900_py_catmix_5rand_1fold_pycpp_11s_oripairs_temp_fulltry.json' #CC_900_py_catmix_5rand_4fold_pyjava_bothper_oripairs_fulltry.json' #CC_900_py_catmix_5rand_4fold_oripairs_temp_fulltry.json' file2 = './llm_for_code/datasets/codecontests/CC_900_py_catmix_5rand_1fold_pycpp_11s_pairs_0_temp_fulltry.json' #CC_900_py_catmix_5rand_4fold_pyjava_bothper_pairs_0_fulltry.json' #CC_900_py_catmix_5rand_4fold_pairs_0_temp_fulltry.json' #file1 = './CC_2200_py_catmix_5rand_1fold_test_1s_5_oripairs_temp_fulltry.json' #'./llm_for_code/datasets/codecontests/CC_1200_py_catmix_5rand_4fold_test_1s_v_oripairs_fulltry.json' #file2 = './CC_2200_py_catmix_5rand_1fold_test_1s_5_pairs_0_temp_fulltry.json' #'./llm_for_code/datasets/codecontests/CC_1200_py_catmix_5rand_4fold_test_1s_v_pairs_0_fulltry.json' filt1 = [] with open(file1, 'r') as file: for line in file: filt1.append(json.loads(line)) filt2 = [] with open(file2, 'r') as file: for line in file: filt2.append(json.loads(line)) filtdf = pd.DataFrame(filt1[:]) filtdf_per = pd.DataFrame(filt2[:]) original_columns_df2 = filtdf_per.columns.tolist() filtdf_per.columns = [f"{col}_df2" for col in filtdf_per.columns] filtdf_combined = pd.concat([filtdf, filtdf_per], axis=1) data = [] data_per = [] for fn in file_names: with open("./dataset_lfs/"+fn, 'r') as file: for line in file: data.append(json.loads(line)) for fn in file_names_per: with open("./dataset_lfs/"+fn, 'r') as file: for line in file: data_per.append(json.loads(line)) print(len(data), len(data_per)) def replace_third(val): parts = val.split('_') if len(parts) >= 3: parts[2] = 'K' return '_'.join(parts) df = pd.DataFrame(data) #df['category'] = df['category'].apply(replace_third) df_per = pd.DataFrame(data_per) original_columns_df2 = df_per.columns.tolist() df_per.columns = [f"{col}_df2" for col in df_per.columns] df_combined = pd.concat([df, df_per], axis=1) print(df_combined.head(), len(df_combined)) df_combined = df_combined[~df_combined.apply(tuple, axis=1).isin(filtdf_combined.apply(tuple, axis=1))] prohib_probid= [ x.get('probid') for x in filtdf['prog1'] if isinstance(x, dict) and 'probid' in x ] prohib_probid += [ x.get('probid') for x in filtdf['prog2'] if isinstance(x, dict) and 'probid' in x ] # # # Extract 'A1' from df and filter # df_combined = df_combined[ ~df['prog1'].apply(lambda x: x.get('probid') if isinstance(x, dict) else None).isin(prohib_probid)] # df_combined = df_combined[ ~df['prog2'].apply(lambda x: x.get('probid') if isinstance(x, dict) else None).isin(prohib_probid)] # df_combined = df_combined[ ~df['prog1'].apply(lambda x: x.get('probid') if isinstance(x, dict) else None).str.startswith('p0').fillna(False)] print("filtered, ", len(df_combined)) # Step 2: Update only the 'cat' column #df_combined['category'] = df_combined['category'].str[:-1] + df_combined['category_df2'].str[-1] # Define a custom function to compare dictionaries in prog1 and prog2 def compare_dicts(row): #print(row) return (str(row['prog1_df2']['scode']), str(row['prog2_df2']['scode'])) # Convert dictionaries to tuple of strings for comparison # Apply the custom function and mark duplicates df_combined = df_combined.dropna(subset=['prog1_df2']) df_combined['unique_check'] = df_combined.apply(compare_dicts, axis=1) # Drop duplicates based on the unique_check column df_unique = df_combined.drop_duplicates(subset=['unique_check']) # Drop the helper column used for comparison df_unique = df_unique.drop(columns=['unique_check']) print(len(data),len(df_unique) ," are unique") #df = pd.DataFrame(data) print(df_unique['category'].value_counts()) def get_sample_size(category): catlen = {'2_0_0_1': (200-135-31), '2_0_0_0': (200-145-38),'1_1_0_1': (1000-845-155) , '1_1_0_0': (1000-818-182), '0_1_0_1': (200-200), '0_1_0_0': (200-200), '1_0_0_0' : (1000-742-144), '1_0_0_1':(1000-725-162) } # if category[0] != '0': # return 0 # catlen = {'0_1_0_0': (200-45), '0_1_0_1' : (200-36)} print(catlen[category]) return catlen[category] # For example: sample 10% of the group size, with a max of 50 # if category == '1_0_0_1' or category == '1_0_0_0': # return 1000 """ pycpp 1_1_0_1 845 1_1_0_0 818 1_0_0_0 742 1_0_0_1 725 2_0_0_0 145 2_0_0_1 135 """ if category[0] == '1': return 300 # return 0 return 65 # Apply the custom sampling sampled_df = ( df_unique.groupby('category', group_keys=False) .apply(lambda x: x.sample(n=min(get_sample_size(x['category'].iloc[0]),len(x)), replace=False)) ) #sampled_df1 = df[~(df['category'].str[0] == '1')] #sampled_df = sampled_df[sampled_df['category'].str.startswith('2')] # Step 1: Identify and separate unique and duplicate rows #df_unique = df_combined.drop_duplicates(subset=['unique_check']) # df_duplicates = df_combined[~df_combined.index.isin(df_unique.index)] # df_duplicates = df_duplicates.drop(columns=['unique_check']) # # Step 2: Sample from df_unique first # # Step 3: Identify categories that still need more samples # remaining_counts = ( # sampled_df['category'].value_counts().reindex(df_combined['category'].unique(), fill_value=0) # ) # print("remaining_counts: ", remaining_counts) # # Step 4: Fill missing samples from df_duplicates # additional_samples = None # additional_samples = [] # for category, count in remaining_counts.items(): # required = get_sample_size(category) - count # if required > 0: # extra_samples = df_duplicates[df_duplicates['category'] == category].sample( # n=min(required, len(df_duplicates[df_duplicates['category'] == category])), # replace=False # ) # additional_samples.append(extra_samples) # # Concatenate the extra samples with the originally sampled DataFrame # if additional_samples: # additional_samples_df = pd.concat(additional_samples) # sampled_df = pd.concat([sampled_df, additional_samples_df]) # # #Final sampled dataframe sampled_df = sampled_df.reset_index(drop=True) print(sampled_df['category'].value_counts(), len(sampled_df)) df1_final = sampled_df.iloc[:, :df.shape[1]] df2_final = sampled_df.iloc[:, df.shape[1]:] print(df1_final.head(), df2_final.head()) df2_final.columns = original_columns_df2 with open(file_names[0][:-5]+"_fulltry_rem.json" , 'w+') as outfile: for _, sample in df1_final.iterrows(): outfile.write(f"{json.dumps(sample.to_dict())}\n") with open(file_names_per[0][:-5]+"_fulltry_rem.json" , 'w+') as outfile: for _, sample in df2_final.iterrows(): outfile.write(f"{json.dumps(sample.to_dict())}\n") print("changed file using", file_names_per[0])