/LyricBnokey

Primary LanguageJupyter NotebookMIT LicenseMIT

LyricB

To use: recommend.py 'artist 1' 'song 1' ['artist 2' 'song 2' ...] where artist1 song1 is the last track listened to And the remaining ones are recommendations to choose from. If only first pair is given it matches against songs already in the database.

To update database: in /back, in order, execute: Download data.ipynb - changing artist list whose songs to download ELMo data.ipynb MakePCA.ipynb - possibly adjusting n components, depending on the plot, currently no PCA at all.

Use example: python recommend.py 'slipknot' 'the devil in i' 'slipknot' 'psychosocial' 'red hot chilli peppers' 'otherside' 'chase atlantic' 'devilish' 'ariana grande' 'thank u, next' 'red hot chilli peppers' 'californication'