# -*- coding: utf-8 -*- """ Created on Mon May 8 14:41:09 2023 @author: gsun """ import numpy as np import pandas as pd from sklearn.metrics import matthews_corrcoef from collections import defaultdict from collections import Counter from sklearn.metrics import roc_auc_score,accuracy_score from sklearn.metrics import confusion_matrix import math,time from imblearn.combine import SMOTEENN from imblearn.under_sampling import RandomUnderSampler from imblearn.over_sampling import RandomOverSampler, SMOTE from sklearn.feature_selection import SelectFromModel,SelectKBest,chi2,f_classif,mutual_info_classif,SelectFdr,SelectFpr import os from sklearn import preprocessing from xgboost import XGBClassifier from sklearn.metrics import f1_score,recall_score,precision_score import random from sklearn.preprocessing import StandardScaler from sklearn.ensemble import RandomForestClassifier import warnings from tabulate import tabulate import matplotlib.pyplot as plt def f(x,y): if x == y: return 1 else: return 0 def cal_metrics(y,pred,metric): if metric == 'acc': return accuracy_score(y,pred) if metric == 'recall f': return recall_score(y,pred,average=None)[0] if metric == 'recall p': p= recall_score(y,pred,average=None) if len(p) ==2: return p[1] else: return -1 if metric == 'precision f': return precision_score(y,pred,average=None)[0] if metric == 'precision p': p = precision_score(y,pred,average=None) if len(p) ==2: return p[1] else: return -1 if metric == 'f1 f': return f1_score(y,pred,average=None)[0] if metric == 'f1 p': return f1_score(y,pred,average=None)[1] def compare(pr_df): metrics = ['precision f','precision p','recall f','recall p'] same = [[],[],[],[]] diff = [[],[],[],[]] overall = [[],[],[],[]] performance = [overall,same,diff] shapes = [] for i,df in enumerate(pr_df): shapes.append(df.shape[0]) pr_now = df.loc[df['last_label']!=df['now_label']].reset_index() pr_same = df.loc[df['last_label']==df['now_label']].reset_index() pr_status = [df,pr_same,pr_now] headers = ['Overall','|','PR_Same','|','PR!=NOW','|'] print_data = [] for metric in metrics: data_row =[metric] for j in range(0,len(pr_status)): df = pr_status[j] y = df['now_label'] pred = df['predict'] perf = cal_metrics(y,pred,metric) data_row+=[perf] data_row+="|" if metric == 'precision f': performance[j][0].append(perf) elif metric == 'precision p': performance[j][1].append(perf) elif metric == 'recall f': performance[j][2].append(perf) elif metric == 'recall p': performance[j][3].append(perf) print_data.append(data_row) print(tabulate(print_data, headers=headers,floatfmt=".3f")) shapes.append(300) # hardcode project a shapes=list(map(lambda x:1000*x/max(shapes),shapes)) return shapes,overall,same,diff #%% def add_project_a_sbs(pr_as_pred,overall,same,diff,pr_shapes): # the performance of SBS on project a is hardcoded since we cannot share the confidential dataset pr_as_pred_a = [[0.7092896174863388],[0.8209959623149394],[0.7092896174863388],[0.8209959623149394]] overall_a = [[0.19419237749546278],[0.5632432432432433],[0.11693989071038251],[0.7012113055181696]] same_a = [[0.1782178217821782],[0.5901759530791789],[0.1386748844375963],[0.6598360655737705]] diff_a = [[0.3695652173913043],[0.4876543209876543],[0.06390977443609022],[0.8909774436090225]] for i in range(0,4): pr_as_pred[i]+=pr_as_pred_a[i] overall[i]+=overall_a[i] same[i]+=same_a[i] diff[i]+=diff_a[i] pr_shapes[0] += [1724] pr_shapes[1] += [12003] return pr_as_pred,overall,same,diff,pr_shapes def add_project_a_bf(pr_as_pred,overall,same,diff,pr_shapes): # the performance of BF on project a is hardcoded since we cannot share the confidential dataset pr_as_pred_a = [[0.7092896174863388],[0.8209959623149394],[0.7092896174863388],[0.8209959623149394]] overall_a = [[0.88], [0.6405910473707084], [0.09617486338797815], [0.9919246298788694]] same_a = [[1.0], [0.6850084222346996], [0.13559322033898305], [1.0]] diff_a = [[0.0], [0.48846153846153845], [0.0], [0.9548872180451128]] for i in range(0,4): pr_as_pred[i]+=pr_as_pred_a[i] overall[i]+=overall_a[i] same[i]+=same_a[i] diff[i]+=diff_a[i] pr_shapes[0] += [1724] pr_shapes[1] += [12003] return pr_as_pred,overall,same,diff,pr_shapes #%% def plot_scatter_plot(x,y,fsize,shapes,x_label = "",y_label = "",title =""): plt.rcParams.update({'font.size': 22}) titles = ['Precision Fail','Precision Pass', 'Recall Fail', 'Recall Pass'] x = np.array(x) y = np.array(y) colors = np.random.rand(len(x[0])) colors = list(range(5,26)) colors = list(map(lambda x:x/20,colors)) i,j = 0,0 plt.figure() r_x = np.arange(0.0,1.1 , 0.1) fig, ax = plt.subplots(2, 2, sharex=True, sharey=True,figsize=(fsize,fsize)) for i in range(0,2): for j in range(0,2): ax[i,j].plot(x[i*2+j][-1],y[i*2+j][-1],'yD',markersize=shapes[-1]/10) ax[i,j].set(xlim=(0, 1), ylim=(0, 1)) ax[i,j].plot([0, 1], [0, 1], ls="--", c=".3",alpha=0.5) ax[i,j].plot([0, 0.95], [0.05, 1], ls="-.", c=".5",linewidth=1,alpha=0.5) ax[i,j].plot([0.05, 1], [0, 0.95], ls="-.", c=".5",linewidth=1,alpha=0.5) ax[i,j].scatter(x[i*2+j], y[i*2+j], s=shapes, edgecolors='gray', c=colors, alpha=0.5,cmap='gray') ax[i,j].set_title(titles[i*2+j],fontsize=22) ax[i,j].fill_between(r_x,r_x, color='grey', alpha=0.1, interpolate=True) ax[i,j].set_xlabel(x_label,fontsize=22) ax[i,j].set_ylabel(y_label,fontsize=22) fig.suptitle(title, fontsize=16) fig.tight_layout() plt.show() #%% def bar_projects(pr_shapes): combined = list(zip(pr_shapes[1],pr_shapes[0])) sorted_combined = sorted(combined, key=lambda x: x[0]) x_sorted, y_sorted = zip(*sorted_combined) percentages = [100*a/b for a,b in zip(y_sorted, x_sorted)] textures = ['x' if i ==20 else '//' for i in range(len(percentages))] plt.figure(figsize=(15,8)) # Define the data plt.grid(which='both',axis='y',linewidth=0.5) colors=['#BFBFBF']*20+['#FFFF14'] plt.bar(range(len(percentages))[:-1], percentages[:-1],edgecolor='gray',alpha=0.8,color=colors[:-1],hatch=textures[:-1],label="Open-Source") plt.bar(range(len(percentages))[-1], percentages[-1],edgecolor='gray',alpha=0.8,color=colors[-1],hatch=textures[-1],label="Project A") tick_labels = list(range(0,21)) plt.xticks(range(len(percentages)), tick_labels) plt.xlabel('Projects Ordered by Size', fontsize = 32) plt.ylabel('Percentage (%)', fontsize = 32) plt.legend() plt.show()