def result():
url = request.args['url']
result = phishing_detection.getResult(url)
return result
def result():
urlname = request.args['name']
result = phishing_detection.getResult(urlname)
return result
def getURL():
if request.method == 'POST':
dec=[0,0,0]
urlname = request.form['url']
url = request.form['url']
print(url)
tokenizerFolder = "tokenizer"
savedModelDirectory = "saved_models"
websiteToTest = url
threshold = 0.5
tokenizer = ByteLevelBPETokenizer(
tokenizerFolder + "/tokenizer.tok-vocab.json",
tokenizerFolder + "/tokenizer.tok-merges.txt",
)
tokenizerVocabSize = tokenizer.get_vocab_size()
print("Tokenizer files have been loaded and the vocab size is %d..." % tokenizerVocabSize)
model = load(savedModelDirectory + "/phishytics-model.joblib")
print("Model loaded...")
# Load document frequency dictionary
docDict = np.load(savedModelDirectory + "/phishytics-model-tfidf-dictionary.npy", allow_pickle=True).item()
print("Document frequency dictionary loaded...")
# Testing
print("Loading webpage...")
try:
request1 = requests.get(websiteToTest)
webpageHtml = str(request1.text)
webpageHtml = webpageHtml.replace("\n", " ")
except Exception as e:
print('\n',e)
print("\nAn error occurred, exiting now... ")
exit()
# Convert text into feature vector
output = tokenizer.encode(webpageHtml)
outputDict = collections.Counter(output.ids)
# Apply tfidf weighting
totalFilesUnderConsideration = docDict["totalFilesUnderConsideration"]
array = [0] * tokenizerVocabSize
for item in outputDict:
if len(docDict[item]) > 0:
array[item] = (outputDict[item]) * (math.log10( totalFilesUnderConsideration / len(docDict[item])))
predictionProbability = model.predict_proba([array])[0][1]
print("\n****************************\n--> Probability that the website is phishing: %.2f" % (predictionProbability * 100))
prediction = "NOT PHISHING"
flag=0
##dec[0]=0
if predictionProbability > threshold:
prediction = "PHISHING"
flag=1
##dec[0]=1
print("--> Based on your threshold of %.2f, this website is +++'%s'+++" % (threshold, prediction))
print("****************************")
result = phishing_detection.getResult(urlname)
print(result)
if result=="Phishing Url":
flag=1
'''
dec[1]==1
else:
dec[1]==0
'''
#if (dec[0]==1)or(dec[1]==1):
if flag==1:
predicted_value=1
else:
predicted_value=0
#print(predicted_value)
if predicted_value == 0:
value = "Legitimate"
return render_template("home.html",error=value)
else:
value = "Phishing"
return render_template("home.html",error=value)
