Music_Recommender_System-Angular-Node.js-MySQL

A simple Music Recommender System developed using Angular, Node.js and MySQL and Recommendations logics are explored, using different filter techniques

Technologies Used

• Node.js v14.15.3 (Download)
• NPM v6.14.9
• MySQL v5.7+
• Angular CLI v11.0.5 (Angular CLI Installation)

External Libraries

Angular Application Libraries (Frontend)

• @ng-bootstrap (Latest)
• Bootstrap (Latest)
• JQuery (Latest)
• Ngx-Spinner (Latest)

Node.js Application Libraries (Backend)

• Knex (Latest) (Knex Package)
• MySQL (Latest) (MySQL Package)
• Natural (Latest) (Natural Package)
• Stopword (Latest) (Stopword Package)
• Striptags (Latest) (Striptags Package)
• Vector-Object (Latest) (Vector-Object Package)

Instructions to Run the Application

Setting up the Database

Users Table

• Name: users
• Purpose: To store user details and filter results of users

Songs Table

• Name: songs
• Purpose: To store song details and similar content of songs

Liked Table

• Name: liked
• Purpose: To store liked songs of users

Download and run the provided SQL files to fill the tables in the selected database. Update database connection details in backend/config.js.

Example:

DB_NAME: "music_recommender",
DB_HOST: "project.cqvbqklh6tez.ap-south-1.rds.amazonaws.com",
DB_USER: "admin",
DB_PASSWORD: "admin123"

Running Backend

1. Navigate to the backend project folder and execute the following command to install dependencies:

npm install

2. Start the Node.js server using the following command:

npm start

3. Once you see "Listening on port 3001," the application is running on port 3001.

4. To stop the server, press Ctrl + C.

Running Frontend

1. Install Angular CLI from Angular CLI Installation.

2. Navigate to the frontend project folder and execute the following command to install dependencies:

npm install

3. Start the Angular Application using the command:

ng serve

4. Once you see "Compiles successfully," the application is running on port 4200.

Component-wise Description

Helpers Folder

Contains reusable code snippets, including functions for data training and response formatting.

Controllers

Main functions where the request and response cycle (ExpressJS) ends.

Repositories

Holds Knex queries, a query helper wrapped around MySQL.

Routes

Contains all the routes of the Express.js server.

App.js

Imports every single component to make it available as an API. Also contains server management.

Content-Based Recommendation (backend/helpers/trainer.js)

• train(): Calls various functions for content-based recommendation.
• validateDocuments(): Validates and preprocesses input documents.
• preprocessDocuments(): Completes document preprocessing, including tokenization.
• produceWordVectors(): Generates word vectors using tf-idf algorithm.
• calculateSimilarities(): Calculates document similarities for recommendations.

Uncomment line 11 in backend/controllers/contentRecommender.js to run the content-based trainer function.

Collaborative Recommendation (backend/helpers/cftrainer.js)

• trainDataForAllUsers(): Trains collaborative recommendation data for all users.
• fetchUserList(), fetchSongList(), fetchLikeList(): Fetches necessary data.
• createCoMatrix(): Creates co-occurrence matrix.
• getRecommendations(): Calculates recommendations using co-occurrence matrix.

Uncomment line 12 in backend/controllers/contentRecommender.js to run the collaborative trainer function.

To run the collaborative filter for the current user, uncomment line 101 in backend/controllers/contentRecommender.js.