app.py

import streamlit as st
from transformers import pipeline
from scraper import scrape_website
from googletrans import Translator
from gtts import gTTS
import speech_recognition as sr
from langdetect import detect
from io import BytesIO
from pydub import AudioSegment
import os

# Initialize components
translator = Translator()
qa_pipeline = pipeline("question-answering", model="distilbert-base-uncased")

# Function to process text-based interaction
def process_text(user_input, context, lang):
    try:
        # Translate to English for the model
        translated_input = translator.translate(user_input, src=lang, dest='en').text
        answer = qa_pipeline({'question': translated_input, 'context': context})['answer']
        # Translate answer back to user's language
        translated_answer = translator.translate(answer, src='en', dest=lang).text
        return translated_answer
    except Exception as e:
        return f"Error: {e}"

# Function to process voice input
def process_voice(file, context):
    recognizer = sr.Recognizer()
    audio = sr.AudioFile(file)
    with audio as source:
        audio_data = recognizer.record(source)
    user_input = recognizer.recognize_google(audio_data)
    lang = detect(user_input)
    return process_text(user_input, context, lang)

# Streamlit UI
st.title("Multilingual Chatbot with Web Scraping")
st.sidebar.title("Options")

# Scraping options
url = st.sidebar.text_input("Enter a website URL to scrape:")
if st.sidebar.button("Scrape Website"):
    context = scrape_website(url)
    st.sidebar.success("Website content scraped successfully!")
else:
    context = "No content available. Please scrape a website first."

# Interaction options
st.subheader("Chat with the Bot")
mode = st.radio("Choose interaction mode:", ["Text", "Voice"])

if mode == "Text":
    user_input = st.text_input("Enter your query:")
    if st.button("Ask"):
        lang = detect(user_input)
        response = process_text(user_input, context, lang)
        st.write(f"Bot: {response}")

elif mode == "Voice":
    uploaded_file = st.file_uploader("Upload a voice file (WAV format):", type="wav")
    if uploaded_file and st.button("Ask"):
        response = process_voice(uploaded_file, context)
        st.write(f"Bot: {response}")

# Voice Output
if st.button("Play Response as Voice"):
    if response:
        tts = gTTS(text=response, lang=lang)
        audio_file = BytesIO()
        tts.write_to_fp(audio_file)
        st.audio(audio_file.getvalue(), format="audio/mp3")
    else:
        st.warning("No response to play!")