Skip to content

A web application that takes audio and display labeled audio with NER & Speaker Segmentation

Notifications You must be signed in to change notification settings

OscarSong2003/MeetingNoteAnalyzer

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

22 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Audio/Meeting Notes Analyzer

  • A simple interface that allows the user to record one/multiple voices
  • Uses Django as backend framework, and Forte to conduct processing. Also uses google speech to text recognition API to transcribe text.
  • Conducts speaker segmentation to identify speakers.
  • Uses Stave to display the annotations
  • Front end is built using React, React-Bootstrap, React-voice-recording library, and Chokra UI

To Start:

  • Clone Repository
  • Set up virtual environment: https://virtualenv.pypa.io/en/latest/user_guide.html (Command: virtualenv venv p=python3.8.5) NOTE: make sure to specify python version
  • Navigate to frontend and pip install libraries specified in requirements.txt
  • Navigate into backend and npm install packages from package.json
  • In frontend directory: run npm run start to start server on localhost:3000
  • In backend directory: run python manage.py runserver to start backend

Restrictions

  • Currently the analysis cannot handle very large/long audio files as it would require the use of cloud storage and asynchronous requests (user can implement this on top of current implementation if necessary)
  • Feel free to checkout the branch google-api-method if you wish to use your own api key (currently uses default generated by python library.
  • Additional optimizations can be made on recognition and NER.

Analysis Tools Used

Video Demonstration (Informative Purpose Only)

File Locations

  • notes/media/documents contains both audio processing and processing (files for speaker segmentation and speech recognition)
  • notes/media/nameEntityRec contains file for NER and running the forte pipeline and stave processor
  • notes/ contains files for django (views.py consists of rest framework)
  • frontend/components contains recording component along with basic page components.
  • frontend/ also ctontains react basic setup and chokra setup

About

A web application that takes audio and display labeled audio with NER & Speaker Segmentation

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published