MixWithTSP

A utility to help you to find out the best sequence of songs in a mix using IBM's Cplex optimizer

What it is

MixWithTSP takes as input a .txt file containing data (BPM, Tonality) about songs as exported by rekordbox. It then computes the best sequence of these songs according to mixing theory by trying to minimize a given distance.

How to use it

You'll need the CPLEX omptimizer python API installed on your computer to use it. git clone this repo, cd into it and run the following command:
python mixtsp.py <your file> <shifts allowed>

Current Features

• Allow some tracks to be shifted by one or more semitones (is penalized by the objective function)

• Allow the optimizer to not include one or more songs in the output mix

More to come

• Finding out the ambiance of a given sound, and take it into account for even smoother mixes
• Using AI, Feature Extraction and rhythm detection to actually create a mix, with the transitions (and maybe even some effects)
• Remove proprietary solver CPlex in the dependencies of this project and use GLPK or CVXPY instead