app.py

#pip install GoogleNews
#pip install --upgrade GoogleNews

import streamlit as st
from GoogleNews import GoogleNews
from sklearn.feature_extraction.text import TfidfVectorizer
from sklearn.metrics.pairwise import cosine_similarity
import pandas as pd
import numpy as np
import string
import re
from nltk.corpus import stopwords
from sklearn.metrics.pairwise import cosine_similarity
import sklearn
import time


googlenews = GoogleNews()
googlenews = GoogleNews(lang='ar')
googlenews.clear()



st.write("""
Arabic Fake News Detection System
A system designed as a part of master project
done by Reem AlFouzan
Supervised by : Dr, Abdulla al mutairi
""")
#df = pd.read_csv('News.csv')
text_input = st.text_input (''' **Enter the text** ''')
if len(text_input) != 0:
    inputt = []
    inputt = pd.DataFrame([text_input])

    googlenews.search(inputt.iloc[0,0])
    googlenews.get_news(inputt.iloc[0,0])

    result_0 = googlenews.page_at(1)
    print("Data")
    print(result_0, "data 2")
    # time.sleep(100)
    if len(result_0) == 0:
      desc_1 = ['لا يوجد نتائج للخبر ']
      link_1 = ['لا يوجد مصدر']
    if len(result_0) != 0:
        desc_1 = googlenews.get_texts()
        link_1 = googlenews.get_links()
        for i in list(range(2, 70)):
    
            result = googlenews.page_at(i)
            desc = googlenews.get_texts()
            link = googlenews.get_links()
    
            desc_1 = desc_1 + desc
            link_1 = link_1 + link
    
    column_names = ["text", 'link']
    df = pd.DataFrame(columns = column_names)
    
    df['text'] = desc_1
    df['link'] = link_1
    
    for letter in '#.][!XR':
        df['text'] = df['text'].astype(str).str.replace(letter,'')
        inputt[0] = inputt[0].astype(str).str.replace(letter,'')
        
    arabic_punctuations = '''`÷×؛<>_()*&^%][ـ،/:"؟.,'{}~¦+|!”…“–ـ'''
    english_punctuations = string.punctuation
    punctuations_list = arabic_punctuations + english_punctuations
    
    def remove_punctuations(text):
        translator = str.maketrans('', '', punctuations_list)
        return text.translate(translator)
    
    def normalize_arabic(text):
        text = re.sub("[إأآا]", "ا", text)
        text = re.sub("ى", "ي", text)
        text = re.sub("ة", "ه", text)
        text = re.sub("گ", "ك", text)
        return text
        
        
    def remove_repeating_char(text):
        return re.sub(r'(.)\1+', r'\1', text)  
            
    def processPost(text): 
        
        #Replace @username with empty string
        text = re.sub('@[^\s]+', ' ', text)
            
        #Convert www.* or https?://* to " "
        text = re.sub('((www\.[^\s]+)|(https?://[^\s]+))',' ',text)
            
        #Replace #word with word
        text = re.sub(r'#([^\s]+)', r'\1', text)
        
        # remove punctuations
        text= remove_punctuations(text)
            
        # normalize the text
        text= normalize_arabic(text)
            
        # remove repeated letters
        text=remove_repeating_char(text)
            
        return text  
            
        
    df['text'] = df['text'].apply(lambda x: processPost(x))
    inputt[0] = inputt[0].apply(lambda x: processPost(x))
        
    st.markdown(f"my input is : { inputt.iloc[0,0] }")
    #input=input.apply(lambda x: processPost(x))
        
        
    vectorizer = TfidfVectorizer()
    vectors = vectorizer.fit_transform(df['text'])
        
    text_tfidf = pd.DataFrame(vectors.toarray())
        
    traninput = vectorizer.transform(inputt[0])
    traninput = traninput.toarray()
    cosine_sim = cosine_similarity(traninput,text_tfidf)
    top = np.max(cosine_sim)
        
        
    if top >= .85 :
        prediction = 'الخبر صحيح'
    elif (top < .85) and (top >= .6) :
        prediction = 'الخبر مظلل '
    elif top < .6 :
        prediction = 'الخبر كاذب '
        
        
    st.markdown(f"most similar news is: { df['text'].iloc[np.argmax(np.array(cosine_sim[0]))] }")
    st.markdown(f"Source url : {df['link'].iloc[np.argmax(np.array(cosine_sim[0]))]}")
    st.markdown(f"Credibility rate : { np.max(cosine_sim)}")
    st.markdown(f"system prediction: { prediction}")
    df.to_csv('Students.csv', sep ='\t')    
    
    
st.sidebar.markdown('مواقع اخباريه معتمده ')
st.sidebar.markdown("[العربية](https://www.alarabiya.net/)")
st.sidebar.markdown("[الجزيرة نت](https://www.aljazeera.net/news/)")
st.sidebar.markdown("[وكالة الانباء الكويتية](https://www.kuna.net.kw/Default.aspx?language=ar)")
    
    #st.markdown('test')