AI-based Voice Assistant

Published on 01 Apr 2021. Written by Vardaan Raj Singh

Voice assistants come to some degree, in small modules and can play out an assortment of activities subsequent to hearing your command. They can turn on lights, answer questions, play music, put in online requests and do a wide range of AI-based stuff.

Read more..

SLNOTE


Voice-based gadgets/applications are growing a great deal. Today, there are Google Assistants, Alexa which accept our voice as information processes it and performs activities based on it. It utilizes the condition of craftsmanship process in Speech to Text, Natural language, deep learning and Text to speech.

Voice assistants are not to be mistaken for virtual helpers or assistants, which can work remotely and can, along these lines, handle a wide range of errands. Voice assistants are technology-based. Today voice assistants are becoming increasingly popular, and their applications in personal and professional worlds are increasing

SLLATEST
Technology Used

System requirements: Python 2.7, macOS Mojave(version 10.14), Spyder IDE

Google and Microsoft give API based administrations that can be effortlessly incorporated with applications. Google likewise offers voice activities which are an API based support to perform activities inside application flawlessly utilizing voice inputs.

Likewise, for ascertaining numerical articulation WolframAlpha API can be utilized.

Play sound Package is utilized to play the saved mp3 sound from the framework.

Project Implementation

Make a Google cloud account.

Click on "Select a project" to make a project in Google Cloud. Click on "New project" and give a name.

Type "Cloud Speech API" on the project search page. We have to empower this API to utilize the Speech to Text API administration. What's more, we have to give credit/debit or ledger subtleties to utilize the free API administration. There is no auto-charge after the free trial finishes. So please give all the subtleties and empower the API. There are constraints in the free preliminary use which is given in the Google discourse API documentation.

Now go to the API page to click on the "Credentials" and then click on "Create Credentials".

Now we will have to download the Json key and for that, an assistance account is needed. In the search bar, type "administration accounts" and make another assistance account. Give a name and dole out "Proprietor" for the Project Role. Click on "Furnish another private key" and check the "JSON" item and click "Spare".

This will download the API key in a JSON group.

For the Python-based part, the following libraries will be required

import speech_recognition

 import os

 import sys

 import re

 import webbrowser

 import smtplib

 import requests

 import subprocess

 from pyowm import OWM

 import youtube_dl

 import vlc

 import urllib

 import urllib2

 import json

 from bs4 import BeautifulSoup as soup

 from urllib2 import urlopen

 import wikipedia

 import random

 from time import strftime

So our initial step is to make the function that will decipher client voice reaction.

On the off chance that you have said the word hi in your command, at that point contingent upon the time, the bot will welcome the client. In the event that the time is in excess of 12 early afternoons, the bot will react "Hi Sir. Good evening", similarly if the time is more than 6 pm, the bot will react "Hi Sir. Goodbye". What's more, when you give the command as shutdown, sys.exit() will be called to end the program.

The client will give any command to open any website and the command ought to be "Hello ! Would you please open Google ". just the italic strong expression ought to be utilized all things considered. You can utilize any sort of prefix, simply deal with the italic bold expression.

On the off chance that you have said the expression open google in your command, at that point it will scan for google name in the client command utilizing re.search(). .The Web browser module gives an elevated level interface to permit showing Web-based reports to clients.

Now suppose you utter 'Email' then at that point the bot will request recipient, suppose If my reaction is raju, If my reaction is raju, the bot will utilize pythons smtplib library. The smtplib module characterizes a SMTP customer meeting object that can be utilized to send mail to any Internet machine with a SMTP or ESMTP audience daemon.

Sending mail is finished with Python's smtplib utilizing a SMTP server. First it will initiate Gmail SMTP utilizing smtplib.SMTP(), at that point, distinguishes the server utilizing ehlo() function, at that point encrypting the session starttls(), after that point login to your mailbox utilizing login(), and send the message utilizing sendmail().

The code snippets for all the above-defined functions can be easily found on the net , but I’d advise to not just blatantly copy the code, first understand it then implement it based on the above model.

SLDYK
Kit required to develop AI-based Voice Assistant:

• No kit required

Technologies you will learn by working on AI-based Voice Assistant:

• Python
• AI
• Machine Learning

Any Questions?