app.py

import streamlit as st
import pandas as pd
import numpy as np
import pickle
import base64
import seaborn as sns
import matplotlib.pyplot as plt

import symbol

st.write("""
# WebPhishing Detection App

WebPhishing Detection App adalah sebuah aplikasi untuk mendeteksi sebuah Phishing pada situs web. 
Aplikasi ini menggunakan berbagai macam paramater untuk menujukan bahwa situs Phishing atau Normal


""")


url_dataset = f'<a href="phishing.csv">Download Dataset CSV File</a>'
st.markdown(url_dataset, unsafe_allow_html=True)

def user_input_features() :
    UsingIP = st.sidebar.selectbox('UsingIP', ('YA', 'Tidak'))
    LongURL = st.sidebar.selectbox('LongURL', ('YA', 'Tidak'))
    ShortURL = st.sidebar.selectbox('Short URL', ('YA', 'Tidak'))
    Symbol = st.sidebar.selectbox('Symbol', ('YA', 'Tidak'))
    Redirecting = st.sidebar.selectbox('Redirecting', ('YA', 'Tidak'))
    PrefixSuffix = st.sidebar.selectbox('PrefixSuffix', ('YA', 'Tidak'))
    SubDomains = st.sidebar.selectbox('SubDomains', ('YA', 'Tidak'))
    HTTPS = st.sidebar.selectbox('HTTPS', ('YA', 'Tidak'))
    DomainRegLen = st.sidebar.selectbox('DomainRegLen', ('YA', 'Tidak'))
    Favicon = st.sidebar.selectbox('Favicon', ('YA', 'Tidak'))
    NonStdPort = st.sidebar.selectbox('NonStdPort', ('YA', 'Tidak'))
    HTTPSDomainURL = st.sidebar.selectbox('HTTPSDomainURL', ('YA', 'Tidak'))
    RequestURL = st.sidebar.selectbox('RequestURL', ('YA', 'Tidak'))
    AnchorURL = st.sidebar.selectbox('AnchorURL', ('YA', 'Tidak'))
    LinksInScriptTags = st.sidebar.selectbox('LinksInScriptTags', ('YA', 'Tidak'))
    ServerFormHandler = st.sidebar.selectbox('ServerFormHandler', ('YA', 'Tidak'))
    InfoEmail = st.sidebar.selectbox('InfoEmail', ('YA', 'Tidak'))
    AbnormalURL = st.sidebar.selectbox('AbnormalURL', ('YA', 'Tidak'))
    WebsiteForwarding = st.sidebar.selectbox('WebsiteForwarding', ('YA', 'Tidak'))
    StatusBarCust = st.sidebar.selectbox('StatusBarCust', ('YA', 'Tidak'))
    DisableRightClick = st.sidebar.selectbox('DisableRightClick', ('YA', 'Tidak'))
    UsingPopupWindow = st.sidebar.selectbox('UsingPopupWindow', ('YA', 'Tidak'))
    IframeRedirection = st.sidebar.selectbox('IframeRedirection', ('YA', 'Tidak'))
    AgeofDomain = st.sidebar.selectbox('AgeofDomain', ('YA', 'Tidak'))
    DNSRecording = st.sidebar.selectbox('DNSRecording', ('YA', 'Tidak'))
    WebsiteTraffic = st.sidebar.selectbox('WebsiteTraffic', ('YA', 'Tidak'))
    PageRank = st.sidebar.selectbox('PageRank', ('YA', 'Tidak'))
    GoogleIndex = st.sidebar.selectbox('GoogleIndex', ('YA', 'Tidak'))
    LinksPointingToPage = st.sidebar.selectbox('LinksPointingToPage', ('YA', 'Tidak'))
    StatsReport = st.sidebar.selectbox('StatsReport', ('YA', 'Tidak'))
  
   # phishingYT01 = 1
    #if(phishingYT == 'Left') :
        #phishingYT01 = 0
   
    usingip = -1
    if(UsingIP == 'YA') :
        usingip = 1
    longurl = -1
    if(LongURL == 'YA') :
        longurl =1
    shorturl = -1
    if(ShortURL == 'YA') :
        shorturl =1
    symbol = -1
    if(Symbol == 'YA') :
        symbol =1
    redirecting = -1
    if(Redirecting == 'YA') :
        redirecting =1
    subdomains = -1
    if(SubDomains == 'YA') :
        subdomains =1
    prefixsuffix = -1
    if(PrefixSuffix == 'YA') :
        prefixsuffix =1
    https = -1
    if(HTTPS == 'YA') :
        https =1
    domainreglen = -1
    if(DomainRegLen == 'YA') :
        domainreglen =1
    favicon = -1
    if(Favicon == 'YA') :
        favicon =1
    nonstdport = -1
    if(NonStdPort == 'YA') :
        nonstdport =1
    httpsdomainurl = -1
    if(HTTPSDomainURL == 'YA') :
        httpsdomainurl =1
    requesturl = -1
    if(RequestURL == 'YA') :
        requesturl =1
    anchorurl = -1
    if(AnchorURL == 'YA') :
        anchorurl =1
    linksinscripttags = -1
    if(LinksInScriptTags == 'YA') :
        linksinscripttags =1
    serverformhandler = -1
    if(ServerFormHandler == 'YA') :
        serverformhandler =1
    infoemail = -1
    if(InfoEmail == 'YA') :
        infoemail  =1
    abnormalurl = -1
    if(AbnormalURL == 'YA') :
        abnormalurl =1
    websiteforwarding = 0
    if(WebsiteForwarding == 'Tidak') :
        websiteforwarding = 0
    statusbarcust = -1
    if(StatusBarCust == 'YA') :
        statusbarcust  =1
    disablerightclick = -1
    if(DisableRightClick == 'YA') :
        disablerightclick  =1
    usingpopupwindow = -1
    if(UsingPopupWindow == 'YA') :
        usingpopupwindow  =1
    iframeredirection = -1
    if(IframeRedirection == 'YA') :
        iframeredirection  =1
    ageofdomain = -1
    if(AgeofDomain == 'YA') :
        ageofdomain  =1
    dnsrecording = -1
    if(DNSRecording == 'YA') :
        dnsrecording  =1
    websitetraffic = -1
    if(WebsiteTraffic == 'YA') :
        websitetraffic  =1
    pagerank= -1
    if(PageRank== 'YA') :
        pagerank  =1
    googleindex = -1
    if(GoogleIndex == 'YA') :
        googleindex  =1
    linkspointingtopage= -1
    if(LinksPointingToPage== 'YA') :
        linkspointingtopage  =1
    statsreport= -1
    if(StatsReport== 'YA') :
        statsreport  =1
   
   #data = {'phishingYT':[phishingYT01], 

    data = { 
    'UsingIP':[usingip],
    'LongURL':[longurl],
    'ShortURL':[shorturl],
    'Symbol@':[symbol],
    'Redirecting//':[redirecting],
    'SubDomains':[subdomains],
    'PrefixSuffix-':[prefixsuffix],
    'HTTPS':[https],
    'DomainRegLen':[domainreglen],
    'Favicon':[favicon],
    'NonStdPort':[nonstdport],
    'HTTPSDomainURL':[httpsdomainurl],
    'RequestURL':[requesturl],
    'AnchorURL':[anchorurl],
    'LinksInScriptTags':[linksinscripttags],
    'ServerFormHandler':[serverformhandler ],
    'InfoEmail':[infoemail],
    'AbnormalURL':[abnormalurl],
    'WebsiteForwarding':[websiteforwarding],
    'StatusBarCust':[statusbarcust],
    'DisableRightClick':[disablerightclick],
    'UsingPopupWindow':[usingpopupwindow],
    'IframeRedirection':[iframeredirection],
    'AgeofDomain':[ageofdomain],
    'DNSRecording':[dnsrecording],
    'WebsiteTraffic':[websitetraffic],
    'PageRank':[pagerank],
    'GoogleIndex':[googleindex],
    'LinksPointingToPage':[linkspointingtopage],
    'StatsReport':[statsreport]}

    features = pd.DataFrame(data)
    return features

input_df = user_input_features()

phishing_raw = pd.read_csv('phishing.csv')
phishing_raw.fillna(0, inplace=True)
phishing = phishing_raw.drop(columns=['class'])
df = pd.concat([input_df, phishing],axis=0)

df = df[:1] # Selects only the first row (the user input data)
df.fillna(0, inplace=True)

features = ['UsingIP', 'LongURL', 'ShortURL', 'Symbol@', 'Redirecting//',
       'PrefixSuffix-', 'SubDomains', 'HTTPS', 'DomainRegLen', 'Favicon',
       'NonStdPort', 'HTTPSDomainURL', 'RequestURL', 'AnchorURL',
       'LinksInScriptTags', 'ServerFormHandler', 'InfoEmail', 'AbnormalURL',
       'WebsiteForwarding', 'StatusBarCust', 'DisableRightClick',
       'UsingPopupWindow', 'IframeRedirection', 'AgeofDomain', 'DNSRecording',
       'WebsiteTraffic', 'PageRank', 'GoogleIndex', 'LinksPointingToPage',
       'StatsReport']

df = df[features]

st.subheader('User Input features')
st.write(df)
load_clf = pickle.load(open('PhishingDETECT_clf.pkl', 'rb'))
detection = load_clf.predict(df)
if(detection < 0) :
    detection = 0
detection_proba = load_clf.predict_proba(df)
phishing_labels = np.array(['Normal', 'Phishing'])
st.subheader('Detection')
st.write(phishing_labels[detection])
st.subheader('Detection Probability')
df_prob = pd.DataFrame(data=detection_proba, index=['Probability'], columns=phishing_labels)
st.write(df_prob)