Every day, we visit so many applications, be it messaging platforms like Messenger, Telegram or ordering products on Amazon, Flipkart, or knowing about weather and the list can go on. And we see that these websites have their own software program for initiating conversations with human beings using rules or artificial intelligence. Users interact with the software program via a conversational interface through written or spoken text. This can be referred to as an assistant.
Example:
Input: what is coding
Output: What/If (stylized as WHAT/ IF) is an American thriller web television miniseries, created by Mike Kelley, that premiered on May 24, 2019, on Netflix. The series stars Renée Zellweger, Jane Levy, Blake Jenner, Daniella Pineda, Keith Powers, Samantha Ware, Dave Annable, Saamer Usmani, John Clarence Stewart and Louis Herthum.
== Premise ==
What/If is a neo-noir thriller that explores “the ripple effects of what happens when acceptable people start doing unacceptable things.
User:
What you need to know ?
An assistant can be made using Natural Language Processing(NLP) one of the most promising fields of artificial intelligence that uses natural languages to enable human interactions with machines.
There are two main approaches to NLP:
- rule-based methods: That follow some pre specified rules and answer as per those rules.
- statistical methods: i.e. methods related to machine learning that learn on their own according to user inputs.
And this article will learn about how to create your own assistant using statistical methods.
To create an assistant, we’ll use the Python programming language, for it has all the modules required for building it. Secondly, Python is easy to understand, even if you don’t have much experience with programming.
Let’s get started
Before, getting into the actual code, we need to understand, that this chatbot or assistant will be voice-based, so we need to import the following modules.
- pyttsx3: A python package that supports common text to speech engines on Mac OS, Windows and Linux.
- speech_recognition: Library for performing speech recognition, with support for several engines and APIs, like CMU Sphinx, Microsoft Bing Voice Recognition, Google Cloud Speech API etc.
- wolframalpha: Python’s support library for WolframAlpha computational intelligence .
- wikipedia: Python’s library that makes it easy to access and parse data from Wikipedia.
To install the above modules on your system, use the following :
pip install pyttsx3 pip install SpeechRecognition pip install wolframalpha pip install wikipedia
Creating a WolframAlpha Account
Since, this chatbot uses WolframAlpha API to find answers, the user must create a free account by signing up at . Follow the steps required to set up a student account that is free and is for personal use only. Generate your app-id and copy it for further reference.
Concepts Followed
- Speech Recognition: Speech recognition is the ability of a machine or program to identify words and phrases in spoken language and convert them to a machine-readable format.
- TTS – text to speech: a form of speech synthesis that converts text into spoken voice output.
- Computational Knowledge Integration: Integrating your bot with computational knowledge intelligence using Wolfram|Alpha.
Below is the implementation.
# Python package supporting common text-to-speech engines import pyttsx3 # For understanding speech import speech_recognition as sr # For fetching the answers # to computational queries import wolframalpha # for fetching wikipedia articles import wikipedia # Function to search the query # that is either entered or spoken # by user def search(query): # try is used for searching with wolframAlpha try: # Generate your App ID from WolframAlpha app_id = "Your WolframAlpha App ID here" client = wolframalpha.Client(app_id) res = client.query(query) answer = next(res.results).text print(answer) SpeakText("Your answer is " + answer) # If the query cannot be searched using # WolframAlpha then it is searched in # wikipedia except: query = query.split(' ') query = " ".join(query[0:]) SpeakText("I am searching for " + query) print(wikipedia.summary(query, sentences = 3)) SpeakText(wikipedia.summary(query, sentences = 3)) # Function to convert text to # speech def SpeakText(command): # Initialize the engine engine = pyttsx3.init() engine.say(command) engine.runAndWait() # Driver's code # input query from the user by # typing or by voice query = input() query = query.lower() # if query is blank then user # is prompted to speak something. if query == '': r = sr.Recognizer() # uses the default microphone # as the source to record voice with sr.Microphone() as source: print("Say Something ") # reduces the background disturbances # and noise for 2 seconds r.adjust_for_ambient_noise(source, 2) # listening to source audio = r.listen(source) try: speech = r.recognize_google(audio) search(speech) # Handling Exceptions if speech # is not understood. except sr.UnknownValueError: print("Google Speech Recognition could not \ understand audio") # Couldn't handle requests, occurs # mainly because of network errors except sr.RequestError as e: print("Could not request results from Google \ Speech Recognition service;{0}".format(e)) else: search(query) |
Output:
Note: The voice output of the output text is also generated.
