Neural-Network-for-Recognizing-Features-in-Music

Bachelor Thesis - Spring 2022

Info

• Authors:
  • Sindre Eiklid ([email protected])
  • Rickard Loland ([email protected])
  • Maren Skårestuen Grindal ([email protected])
• Final report: NTNU Open
• Website address: localhost:3000/
• Endpoints:
  • Main API
  • Model API

Instructions

0. Download Docker
1. You need to create a serviceAccountKey.json file in the API folder and add the service key provided by your firebase project to this file.
2. Run the command docker compose up to build the images and run the containers.
   • To stop the containers run the command docker compose stop
   • To start the containers run the command docker compose start

Note: The application is programmed to start when the server starts. To stop this, edit the docker-compose.yml file by removing the restart: always lines. You will have to build the containers again by running docker compose down and then docker compose up.

Usage

Open the website at localhost:3000/ and utilize the functionality through the UI.

Main page

localhost:3000/ brings you to the main page, here you can input a youtube link to analyze. When you submit a song it will say Parsing link..., once it is finished it will say Song successfully analyzed, the result is uploaded to the results page.

Results page

http://localhost:3000/results brings you to the result page, here you can see all the results from the song analysis. It allows you to edit the result if you wish and approve the song. Once it is approved, the results are locked in. If you wish to edit it again, you can un-approve the songs. If you want to remove any results from the database, you can click the trashcan icon.

Status page

http://localhost:3000/status brings you to the status page, where you can see the current songs that are being analyzed, failed songs, and the status of each endpoint.