Speech Analysis.

Developed as a part of 24hr "Ode to Code" hackathon conducted by Riskcovry and BMSCE,CSE Dept.

Won 2nd Place in the Competetion .The problem statement is given in the "Problem Statement.pdf" file.

Contains a web app and an attempted API.

Uses Flask,Google-Speech-Web API.

Takes the following input : question id,options and audio.The answer is returned.The Project can be viewed in action in the "Demo.mp4" file.

Team Members :

Niharika C

Phoebe P

Ramya B S

Vivek M R

INFO/STEPS TO RUN :

1.Create a virtual environment if necessary and install the requirements mentioned in requirements.txt.

Note : If installation from the txt file fails,Install the following modules manually :

flask

SpeechRecognition

requests

textblob

pipwin

pyaudio

2.Run app.py

3.Web App is hosted at "localhost:5000".Works fine.

4.API can be accessed at "localhost:5000/api" using "POST" method and sending the data in JSON format.

5.API Testing UI can be access at "localhost:5000/api/test" using "GET" method.This was supposed to take inputs from the user and call the API and pass the inputs in form of JSON.There seems to be a problem in recieving the data as JSON in the API.

Tasks Done :

1.UI to upload files and transcribe

2.Takes input in either audio wav file or from microphone.

3.Question 1 is working with good accuracy.

4.Question 2 is working with excellent accuracy.

5.Question 3 is working with excellent accuracy.

6.Unprocessed Transcribed Audio can also be obtained.

7.API Endpoint created to handle requests

8.API Testing Tool is created.Sends JSON data to the API.

Tasks Left :

1.Discovered bugs in the API. Doesn't recieve JSON data correctly.Need to work on it.