raghavm1/Sentiment-Song-Search

🎶 A search engine for songs, where you can filter by sentiment and lyrics

Sentiment-Song-Search

Project Report

For the project report that was submitted, refer to Report.pdf

Data crawling

data_collection.ipynb

Fetches for all song data like lyrics, artist info, etc

Song_data_prep.ipynb

For cleaning lyrics by removing info like [Verse 1], [Intro ], [Chorus ] etc.

Upload the data_collection notebooks on Google Colab

Click on each cell and press Shift + Enter to run

Explicit Word Rating

Install nltk library for python

pip install --user -U nltk

Ensure the dataset named Song_data_prep-L.csv is present in the same directory as explicit_word_removal.py file

To run the python scrip to generate explicit ratings

python explicit_word_removal.py

Sentiment Analysis

Run the jupyter notebooks in the sentiment analysis folder on google colab in the order

• data_analysis_and_preprocessing.ipynb
• sentiment_analysis_ouput_generation.ipynb

Indexing

No installation is required, the only pre-requisite is that JDK must be set up on your machine.

To run Solr

Run the following command to start the Solr server:

cd solr
bin/solr start

To check if the Solr server is running, run the following command:

cd solr
bin/solr status

To access the Solr admin dashboard, visit http://localhost:8983/solr

Now queries to the database can be made by first selecting Core Selector > songs and then clicking on the Query tab.

Client

Setup

We use virtualenv to make sure the same python libraries are being used. To install virtualenv, run the following command:

pip3 install virtualenv

Once installed, run the following command to create and activate the virtual environment:

virtualenv env
source env/bin/activate

You should now be inside the virtual environment and can download the required libraries:

pip install -r requirements.txt

Starting the client server

Run the following command and the client will be available at http://127.0.0.1:8051:

cd client
streamlit run app.py