mebeingme31/DIC-PHASE3
0
1import gradio as gr2 3import pandas as pd4import pickle5import matplotlib.pyplot as plt6import pathlib7import matplotlib.lines as mlines8 9global_pr_qm=[]10global_da_ph=[]11answer=''12white_circle = mlines.Line2D([], [], color='white', markeredgecolor='black', marker='o', markersize=10, label='New Point', linestyle='None')13red_circle = mlines.Line2D([], [], color='red', marker='o', markersize=10, label='Legitimate', linestyle='None')14green_circle = mlines.Line2D([], [], color='green', marker='o', markersize=10, label='Phishing', linestyle='None')15 16def visualisations():17 global global_pr_qm18 with open('dataframe.pkl', 'rb') as file:19 df = pickle.load(file)20 plt.figure(figsize=(8, 5))21 22 red_points = df[df['is_legitimate'] == 0]23 blue_points = df[df['is_legitimate'] == 1]24 25 plt.scatter(red_points['nb_qm'], red_points['page_rank'], color='red' )26 plt.scatter(blue_points['nb_qm'], blue_points['page_rank'], color='green')27 if len(global_pr_qm)!=0:28 # if answer.split(',')[0]=='Legitimate':29 # ccccc='green'30 # else:31 # ccccc='red'32 33 ccccc=[]34 for i in answer[:-1].split(','):35 if i.startswith('Le'):36 ccccc.append('green')37 else:38 ccccc.append('red')39 40 x_coords, y_coords = zip(*global_pr_qm)41 plt.scatter(x_coords, y_coords,color=ccccc , s=100, edgecolor='black', zorder=5)42 43 # plt.scatter(global_pr_qm[0][0],global_pr_qm[0][1],color=ccccc , s=100, edgecolor='black', label='New Point', zorder=5)44 plt.title('scatter plot')45 plt.xlabel('Number of Question marks')46 plt.ylabel('page_rank')47 # plt.legend(title="Status")48 plt.legend(title="Status",handles=[white_circle, red_circle, green_circle])49 50 return plt.gcf() 51 52def visualisations2():53 global global_da_ph54 with open('dataframe.pkl', 'rb') as file:55 df = pickle.load(file)56 plt.figure(figsize=(8, 5))57 58 red_points = df[df['is_legitimate'] == 0]59 blue_points = df[df['is_legitimate'] == 1]60 61 plt.scatter(red_points['domain_age'], red_points['phish_hints'], color='red')62 plt.scatter(blue_points['domain_age'], blue_points['phish_hints'], color='green')63 64 if len(global_da_ph)!=0:65 66 ccccc=[]67 for i in answer[:-1].split(','):68 if i.startswith('Le'):69 ccccc.append('green')70 else:71 ccccc.append('red')72 73 # if answer.split(',')[0]=='Legitimate':74 # ccccc='green'75 # else:76 # ccccc='red'77 x_coords, y_coords = zip(*global_da_ph)78 plt.scatter(x_coords, y_coords,color=ccccc , s=100, edgecolor='black', zorder=5)79 80 plt.title('scatter plot')81 plt.xlabel('domain_age')82 plt.ylabel('phish_hints')83 # plt.legend(title="Status")84 plt.legend(title="Status",handles=[white_circle, red_circle, green_circle])85 return plt.gcf() 86 87 88def greet(aa):89 df_for_test=pd.DataFrame(columns=['length_url', 'length_hostname', 'ip', 'nb_dots', 'nb_hyphens', 'nb_at',90 'nb_qm', 'nb_and', 'nb_eq', 'nb_slash', 'nb_colon', 'nb_semicolumn',91 'nb_www', 'nb_com', 'nb_dslash', 'http_in_path', 'https_token',92 'ratio_digits_url', 'ratio_digits_host', 'tld_in_path',93 'tld_in_subdomain', 'abnormal_subdomain', 'nb_subdomains',94 'prefix_suffix', 'shortening_service', 'nb_external_redirection',95 'length_words_raw', 'shortest_word_host', 'shortest_word_path',96 'longest_words_raw', 'longest_word_host', 'longest_word_path',97 'avg_words_raw', 'avg_word_host', 'avg_word_path', 'phish_hints',98 'domain_in_brand', 'brand_in_subdomain', 'brand_in_path',99 'suspecious_tld', 'statistical_report', 'nb_hyperlinks',100 'ratio_inthyperlinks', 'ratio_exthyperlinks', 'nb_extcss',101 'ratio_extredirection', 'external_favicon', 'links_in_tags',102 'ratio_intmedia', 'ratio_extmedia', 'popup_window', 'safe_anchor',103 'empty_title', 'domain_in_title', 'domain_with_copyright',104 'whois_registered_domain', 'domain_registration_length', 'domain_age',105 'web_traffic', 'dns_record', 'google_index', 'page_rank'])106 df_for_test.loc[0]=eval(aa)107 108 with open('minMaxScalerForTestingData.pkl','rb') as f:109 scaler_objects=pickle.load(f)110 111 for i in list(scaler_objects.keys()):112 if i[0] in df_for_test.columns:113 df_for_test[i[0]]=scaler_objects[i].transform(df_for_test[i[0]].values.reshape(-1,1))114 115 with open('all_models.pkl', 'rb') as file:116 all_models = pickle.load(file)117 118 def predict_all_models(data):119 if all_models['SVM'].predict([data])[0] ==1:120 return 'Legitimate'121 return 'Phishing' 122 return predict_all_models(df_for_test.iloc[0].tolist()) 123 124def combined_interface(url=None):125 if url is None:126 greeting = "Please enter a feature vector to analyze." 127 else:128 greeting = greet(url) 129 return visualisations(), greeting130 131def load_on_start():132 global global_pr_qm,global_da_ph,answer133 answer=''134 global_pr_qm=[]135 global_da_ph=[]136 return visualisations2(),visualisations(), "Please enter a feature vector to analyze."137 138def upload_file(files):139 global global_da_ph,global_pr_qm140 global answer141 try:142 a=pd.read_csv(files)143 # print(a.columns)144 145 for i in range(len(a)):146 global_da_ph.append((a.loc[i,'domain_age'],a.loc[i,'phish_hints']))147 global_pr_qm.append((a.loc[i,'nb_qm'],a.loc[i,'page_rank']))148 149 answer=answer+greet(str(a.loc[i].values.tolist()))+','150 except Exception as e:151 print('Error:', e)152 return "Failed to process the file."153 154 155 print(files)156 # visualisations2()157 return answer[:-1],visualisations2(),visualisations()158 159 160with gr.Blocks() as demo:161 # plot_output2 = None162 with gr.Row():163 with gr.Column():164 # file_output = gr.File()165 text_output2 = gr.Textbox(label='Status of Website')166 upload_button=gr.UploadButton(label='upload csv',file_count="single")167 # plot_output2= gr.Plot()168 169 gr.DownloadButton("Download Input Template", value=pathlib.Path('FeatureVector.csv'))170 # url_input = gr.Textbox(label="Feature Vector")171 # submit_btn = gr.Button("Submit")172 173 with gr.Row():174 text_output = gr.Textbox(label='Status of Website',visible=False)175 176 gr.Markdown('# Visualisations')177 with gr.Row():178 with gr.Column():179 plot_output = gr.Plot()180 with gr.Column():181 gr.Markdown("## We can clearly understand that as the number of question marks are increasing (indicating more queries in URL), there are a lot of phishing sites indicating that phishing websites generally have more number of query parameters in a URL when compared to a legitimate one.") 182 with gr.Row():183 with gr.Column():184 plot_output2= gr.Plot()185 upload_button.upload(upload_file, upload_button,[text_output2,plot_output2,plot_output])186 with gr.Column():187 gr.Markdown("## We can see that as the age of the domain is increasing there are relatively less phishing site which indicates that Older domains might have fewer phishing hints due to established legitimacy over time. We can also interpret this as websites which are just created have more phish_hints potentially being a phishing website.") 188 # submit_btn.click(189 # fn=lambda url: (visualisations2(),visualisations(), greet(url)),190 # inputs=url_input,191 # outputs=[plot_output, plot_output2,text_output]192 # )193 194 195 demo.load(load_on_start, inputs=None, outputs=[plot_output2,plot_output, text_output])196 197 198demo.launch()199 200# import pandas as pd201 202# df_for_test=pd.DataFrame(columns=['length_url', 'length_hostname', 'ip', 'nb_dots', 'nb_hyphens', 'nb_at',203# 'nb_qm', 'nb_and', 'nb_eq', 'nb_slash', 'nb_colon', 'nb_semicolumn',204# 'nb_www', 'nb_com', 'nb_dslash', 'http_in_path', 'https_token',205# 'ratio_digits_url', 'ratio_digits_host', 'tld_in_path',206# 'tld_in_subdomain', 'abnormal_subdomain', 'nb_subdomains',207# 'prefix_suffix', 'shortening_service', 'nb_external_redirection',208# 'length_words_raw', 'shortest_word_host', 'shortest_word_path',209# 'longest_words_raw', 'longest_word_host', 'longest_word_path',210# 'avg_words_raw', 'avg_word_host', 'avg_word_path', 'phish_hints',211# 'domain_in_brand', 'brand_in_subdomain', 'brand_in_path',212# 'suspecious_tld', 'statistical_report', 'nb_hyperlinks',213# 'ratio_inthyperlinks', 'ratio_exthyperlinks', 'nb_extcss',214# 'ratio_extredirection', 'external_favicon', 'links_in_tags',215# 'ratio_intmedia', 'ratio_extmedia', 'popup_window', 'safe_anchor',216# 'empty_title', 'domain_in_title', 'domain_with_copyright',217# 'whois_registered_domain', 'domain_registration_length', 'domain_age',218# 'web_traffic', 'dns_record', 'google_index', 'page_rank'])219 220# df_for_test.to_csv('./FeatureVector.csv')221 222# df=pd.read_csv('FeatureVector.csv')223# df.loc[0]=[68.0, 40.0, 0.0, 6.0, 0.0, 0.0, 0.0, 0.0, 0.0, 4.0, 1.0, 0.0, 0.0,2.0, 0.0, 0.0, 1.0, 0.029411765, 0.05, 0.0, 0.0, 0.0, 3.0, 0.0,0.0, 0.0, 7.0, 5.0, 3.0, 17.0, 17.0, 6.0, 6.857142857, 10.33333333,4.25, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 4.0, 1.0, 0.0, 0.0, 0.0, 0.0,100.0, 0.0, 0.0, 0.0, 100.0, 0.0, 1.0, 0.0, 0.0, 2381.0, 0, 0.0,0.0, 1.0, 2.0]224# df.loc[1]=[55.0, 15.0, 0.0, 2.0, 2.0, 0.0, 0.0, 0.0, 0.0, 5.0, 1.0, 0.0, 1.0,0.0, 0.0, 0.0, 1.0, 0.0, 0.0, 0.0, 0.0, 0.0, 2.0, 0.0, 0.0, 0.0,6.0, 3.0, 4.0, 11.0, 7.0, 11.0, 6.333333333, 5.0, 7.0, 0.0, 0.0,0.0, 0.0, 0.0, 0.0, 102.0, 0.470588235, 0.529411765, 0.0,0.537037037, 0.0, 76.47058824, 0.0, 100.0, 0.0, 0.0, 0.0, 0.0, 1.0,0.0, 224.0, 8175.0, 8725.0, 0.0, 0.0, 6.0,]225# df.to_csv('test.csv',index=False)226