import matplotlib.pyplot as plt

import seaborn as sns

import pandas as pd

import os

def plot_result(tool,Base, Fair):

try:

if Fair:

EvaluationsFolderPath = './Results/Evaluations_Fair/'

else:

EvaluationsFolderPath = './Results/Evaluations/'

ToolsCapacity = pd.read_excel('./Mapping/ToolsCapacity.xlsx',sheet_name='DASP',index_col='Tool')

labels = []

if len(Base) == 1 and Base[0].lower() == 'all':

Base = [f.name for f in os.scandir(EvaluationsFolderPath) if f.is_dir()]

if len(tool) == 1 and tool[0].lower() == 'all':

tool = [f.name.split('.')[0] for f in os.scandir(EvaluationsFolderPath+Base[0]) if f.is_file() and '.csv' in f.name]

tool = sorted(tool)

Base = sorted(Base)

resultDF= get_performance_results(tool,Base,Fair)

plot_performance_results(resultDF,tool,ToolsCapacity)

return resultDF

except Exception as err:

print(f"Unexpected {err=}, {type(err)=}")

raise

def get_labelsList(ToolsCapacity):

labels = []

DASP_Labels = ['Reentrancy','Access Control','Arithmetic','Unchecked Return Values','DoS','Bad Randomness','Front-Running','Time manipulation','Short Address Attack']

for v in DASP_Labels:

if 1 in ToolsCapacity[v].tolist():

labels.append(v)

return labels

def plot_performance_results(resultDF,tool,ToolsCapacity):

Base = resultDF.Base.unique().tolist()

if len(tool) > 1 and len(Base) > 1: # Many tools x Many Bases

plot_ManyTool_ManyBase(resultDF,tool,ToolsCapacity)

elif len(tool)== 1 and len(Base) > 1: # One tool x Many Bases >> ToBeAdded

print(len(resultDF))

elif len(tool)> 1 and len(Base) == 1: # Many tools x One Base

plot_ManyTool_OneBase(resultDF,tool,ToolsCapacity)

else: # One tool x One Base

plot_OneTool_OneBase(resultDF,tool)

def plot_ManyTool_ManyBase(resultDF,tool,ToolsCapacity):

Base = resultDF.Base.unique().tolist()

Vulnerabilities = get_labelsList(ToolsCapacity)

dict_resultDF = resultDF.to_dict('records')

rows = len(Vulnerabilities)

cols = len(Base)

fig, axs = plt.subplots(rows,cols, figsize=(25, 30))

plt.rc('xtick', labelsize='large')

plt.rc('ytick', labelsize=6)

for v in Vulnerabilities:

for b in Base:

if (b =='Doublade' and Vulnerabilities.index(v)+1 not in [1,2,4,5]) or (b =='SolidiFI' and Vulnerabilities.index(v)+1 not in [1,2,3,4,7,8]):

axs[Vulnerabilities.index(v),Base.index(b)].axis('off')

else:

Precision_Scores = []

Recall_Scores = []

#store Precision and Recall in one list

for index in range(0,len(dict_resultDF)):

if dict_resultDF[index]['Base'] == b and dict_resultDF[index]['Label'] == v:

for t in tool:

Precision_Scores.append(dict_resultDF[index][t+'_Precision'])

Recall_Scores.append(dict_resultDF[index][t+'_Recall'])

continue

# Plot bar chart

bar_width = 0.4

x_range_Precision_Scores = [idx - 0.2 for idx in range(len(tool))]

x_range_Recall_Scores = [idx for idx in range(len(tool))]

axs[Vulnerabilities.index(v),Base.index(b)].bar(x_range_Precision_Scores,Precision_Scores,width=bar_width,color='lightblue')

axs[Vulnerabilities.index(v),Base.index(b)].bar(x_range_Recall_Scores,Recall_Scores,width=bar_width,color='steelblue')

axs[Vulnerabilities.index(v),Base.index(b)].grid(True, color = "grey", which='major', linewidth = "0.3", linestyle = "-.")

axs[Vulnerabilities.index(v),Base.index(b)].grid(True, color="grey", which='minor', linestyle=':', linewidth="0.5")

axs[Vulnerabilities.index(v),Base.index(b)].minorticks_on()

axs[Vulnerabilities.index(v),Base.index(b)].set_yticks((0,0.5,1))

ax = axs[Vulnerabilities.index(v),Base.index(b)]

for p in ax.patches:

if p.get_height() > 0:

ax.text(p.get_x()+0,

p.get_height()* .5 ,

'{0:.2f}'.format(p.get_height()),

color='black', rotation='vertical', size='small')

axs[Vulnerabilities.index(v),Base.index(b)].set_ylabel(v, rotation=90,fontsize=7)

axs[Vulnerabilities.index(v),Base.index(b)].set_xticks(range(len(tool)))

axs[Vulnerabilities.index(v),Base.index(b)].set_xticklabels(tool,rotation = 30,fontsize=7)

pad = 5

for ax, col in zip(axs[0], Base):

ax.annotate(col,xy=(0.5, 1), xytext=(0, pad),

xycoords='axes fraction', textcoords='offset points',

size='large', ha='center', va='baseline')

fig.legend(["Precision", "Recall"],loc="lower left", ncol=1)

#fig.subplots_adjust(left=0, top=1)

fig.tight_layout()

plt.show()

def plot_ManyTool_OneBase(resultDF,tool,ToolsCapacity):

Base = resultDF.Base.unique().tolist()

Vulnerabilities = get_labelsList(ToolsCapacity)

dict_resultDF = resultDF.to_dict('records')

rows = cols =3

fig, axs = plt.subplots(rows,cols, figsize=(15, 13))

plt.rc('xtick', labelsize=10)

plt.rc('ytick', labelsize=10)

x=y=0

for v in Vulnerabilities:

Precision_Scores = []

Recall_Scores = []

#store Precision and Recall in one list

for index in range(0,len(dict_resultDF)):

if dict_resultDF[index]['Label'] == v:

#print('b is:', b, 'and Base is:', dict_resultDF[index]['Base'])

for t in tool:

Precision_Scores.append(dict_resultDF[index][t+'_Precision'])

Recall_Scores.append(dict_resultDF[index][t+'_Recall'])

continue

# Plot bar chart

bar_width = 0.4

x_range_Precision_Scores = [idx - bar_width/2 for idx in range(len(tool))]

x_range_Recall_Scores = [idx + bar_width/2 for idx in range(len(tool))]

axs[x,y].bar(x_range_Precision_Scores,Precision_Scores,width=bar_width,color='lightblue')

axs[x,y].bar(x_range_Recall_Scores,Recall_Scores,width=bar_width,color='steelblue')

axs[x,y].grid(True, color = "grey", which='major', linewidth = "0.3", linestyle = "-.")

axs[x,y].grid(True, color="grey", which='minor', linestyle=':', linewidth="0.5")

axs[x,y].minorticks_on()

axs[x,y].set_yticks((0,0.5,1))

ax = axs[x,y]

for p in ax.patches:

if p.get_height() > 0:

ax.text(p.get_x()+0,

p.get_height()* .5 ,

'{0:.2f}'.format(p.get_height()),

color='black', rotation='vertical', size='large')

axs[x,y].set_xlabel('Tool')

axs[x,y].set_ylabel('Score')

axs[x,y].set_title( v, fontsize=10)

axs[x,y].set_xticks(range(len(tool)))

axs[x,y].set_xticklabels(tool,rotation = 75)

if (y+1)%3 == 0:

y=0

x +=1

else:

y +=1

fig.legend(["Precision", "Recall"])

fig.tight_layout()

axs.flat[-1].set_visible(False)

plt.show()

def plot_OneVuln_ManyTool(Vulnerability,Base,tool,Precision_Recall_scores):

#create DataFrame

Precision_Recall_DF = pd.DataFrame({'Tool': tool*2,

'Score': Precision_Recall_scores,

'Evaluation Metrics': ['Precision']*6 + ['Recall']*6})

#set seaborn plotting aesthetics

sns.set(style='ticks')

#create grouped bar charts

g, axes = plt.subplots(9,3)

ax = axes[0,0]

g=sns.catplot(x='Tool', y='Score', hue='Evaluation Metrics', data=Precision_Recall_DF, kind='bar', height=4, aspect=2.5, palette="PuBu")

for p in ax.patches:

ax.text(p.get_x() + 0.15,

p.get_height()* .5 ,

'{0:.3f}'.format(p.get_height()),

color='black', rotation='vertical', size='small')

plt.title('Tools performance in detecting '+ Vulnerability + ' on ' + Base + ' dataset', fontsize=12)

plt.grid(True, color = "grey", which='major', linewidth = "0.3", linestyle = "-.")

plt.grid(True, color="grey", which='minor', linestyle=':', linewidth="0.5");

plt.minorticks_on()

plt.xticks(rotation = 90)

plt.show()

def plot_OneTool_OneBase(resultDF,tool):

Base = resultDF.Base.unique().tolist()

#store Precision and Recall in one list

Precision_Recall_scores =[]

Precision_Scores = resultDF[tool[0]+'_Precision'].to_list()

Recall_Scores = resultDF[tool[0]+'_Recall'].to_list()

Precision_Recall_scores = Precision_Scores + Recall_Scores

NoOfLabels = len(resultDF)

#create DataFrame

Precision_Recall_And_DF = pd.DataFrame({'Vulnerability': resultDF['Label'].to_list()*2,

'Score': Precision_Recall_scores,

'Evaluation Metrics': ['Precision']*NoOfLabels + ['Recall']*NoOfLabels})

#set seaborn plotting aesthetics

sns.set(style='ticks')

#create grouped bar chart

g=sns.catplot(x='Vulnerability', y='Score', hue='Evaluation Metrics', data=Precision_Recall_And_DF, kind='bar', height=4, aspect=2.5, palette="PuBu")

ax = g.facet_axis(0,0)

for p in ax.patches:

ax.text(p.get_x() + 0.15,

p.get_height()* .5 ,

'{0:.2f}'.format(p.get_height()),

color='black', rotation='vertical', size='small')

plt.title('Precision and Recall for ' + tool[0] + ' per vulnerability on ' + Base[0], fontsize=12)

plt.grid(True, color = "grey", which='major', linewidth = "0.3", linestyle = "-.")

plt.grid(True, color="grey", which='minor', linestyle=':', linewidth="0.5");

plt.minorticks_on()

plt.xticks(rotation = 90)

plt.show()

def get_performance_results(tool,Base,Fair):

resultDF = buildDF(tool)

if Fair:

EvaluationsFolderPath = './Results/Evaluations_Fair/'

else:

EvaluationsFolderPath = './Results/Evaluations/'

for b in Base:

subResultDF = pd.DataFrame(columns = resultDF.columns.to_list())

flag = True

for t in tool:

toolResult = pd.read_csv(EvaluationsFolderPath+b+'/'+t+'.csv')

for index, row in toolResult.iterrows():

if flag:

subResultDF.at[index,'Base'] = toolResult.at[index,'Base'].rsplit('.')[0]

subResultDF.at[index,'Label'] = toolResult.at[index,'Label']

subResultDF.at[index,'In Base'] = toolResult.at[index,'In Base']

subResultDF.at[index,'Detectable By ' + t] = toolResult.at[index,'Detectable By Tool']

subResultDF.at[index,t+'_Recall'] = toolResult.at[index,'Recall']

subResultDF.at[index,t+'_Precision'] = toolResult.at[index,'Precision']

flag = False

resultDF = pd.concat([resultDF,subResultDF])

resultDF.reset_index(inplace=True, drop=True)

return resultDF

def buildDF(tool):

resultDF = pd.DataFrame(columns = ['Base','Label','In Base'])

for t in tool:

resultDF['Detectable By ' + t] = ''

resultDF[t+'_Recall'] = ''

resultDF[t+'_Precision'] = ''

return resultDF

#plot_result(['All'],['SolidiFI','Doublade','JiuZhou','SBcurated'])