print("\tperformance: exec_time, comp_time") print("\tnumber of samples: 25, 50, 100") print("\t#runs: 10, 100") print("\talgorithm: ceal") exit() wf = sys.argv[1] perfn = sys.argv[2] num_smpl = int(sys.argv[3]) num_run = int(sys.argv[4]) algo = sys.argv[5] dir_name = '../plot/pct_repl/' + wf + '_' + perfn + '/' if (wf == 'lv'): lmp_mdl = mdlr.train_mdl(sp.df_lmp, sp.lmp_confn, perfn) vr_mdl = mdlr.train_mdl(sp.df_vr, sp.vr_confn, perfn) cpnt_mdls = [lmp_mdl, vr_mdl] cpnt_confns = (sp.lmp_confn, sp.vr_confn) confn = sp.lv_confn if (perfn == 'exec_time'): if (num_smpl == 50): if (algo == 'ceal'): pct_rand = 0.1 num_iter = 7 else: print("Error: unknown algorithm!") exit() else: # 100 samples if (algo == 'ceal'): pct_rand = 0.05
import numpy as np import pandas as pd import xgboost as xgb import sample as sp import modeler as mdlr gs_mdl = mdlr.train_mdl(sp.df_gs, sp.gs_confn, 'exec_time') pdf_mdl = mdlr.train_mdl(sp.df_pdf, sp.pdf_confn, 'exec_time') cpnt_mdls = [gs_mdl, pdf_mdl, sp.df_gplot, sp.df_pplot] for i in range(len(cpnt_mdls)): if isinstance(cpnt_mdls[i], xgb.sklearn.XGBRegressor): print("xgb.sklearn.XGBRegressor") elif isinstance(cpnt_mdls[i], float): print("float") elif isinstance(cpnt_mdls[i], pd.core.frame.DataFrame): print("pd.core.frame.DataFrame") else: print(f"Unknown type of cpnt_mdls[{i}]") print(3 * np.ones(5))