Auto-Playlist

AI-based automatic DJ setlist maker.

• If you use this system for DJ, please cite this page or one of my socials below.
• このシステムをDJで使用した場合、こちらのページのリンク、又は、以下にある私のSNSをタグ付けしてください。
• Yuga SNS Links - Instagram, Twitter, Facebook

First of all...(Citation)

• This work uses 'Kapre' from this paper. Much love and appreciation to Keunwoo Choi for making such a wonderful API!
• Kapre GitHub

How to use

It may be better to make a new python environment for using this.

• This will be worked with Standalone Max 8.1.3 or later (checked with Max 8.1.3).
• It may be a little long preparation but I hope you'd try!

Preparation

0. Please copy all the tracks you have to a new empty directory

1. Clone this repository and install python requirements

git clone https://github.com/yuga-k-dj/auto-playlist.git
pip install -r requirements.txt

2. Make your tracks avialable for Auto-Playlist

• NOTE! You will have 30sec ".wav" files each tracks after this step and your tracks in the direcotory will be changed into ".wav" files. Please be sure to have enough storage (e.g. 5523tracks == you would need 29.54GB empty storage, and another bigger storage for the original directory) before using the code below.
• Please try this code with ~1000 tracks if you don't have enough storage.

cd auto-playlist
python scanplaylist_for_auto.py 'Path to the directory where you copied all the tracks at step 0'

If this worked succesfully, you would have a directory named 'track_in_number' with 30s ".wav" files and one ".csv" file named 'tracknames.csv' at current directory.

• If the error "wave.Error: unknown format: 3" happens, please delete the "NOW" track and try the code above again.

Next, you have to make your own Annoy model for extracting features of the tracks.

python make_your_annoy.py

If this worked succesfully, you would have a file named 'auto_playlist_annoy.ann' at 'models' directory.

3. Making another ".csv" for avoiding choosing same tracks

python make_played_csv.py

You would have a ".csv" file named 'played_tracks.csv' at current directory.

4. Node.js (In case the Auto-Playlist doesn't work with only the steps above)

4.1. Installing nodebrew

curl -L git.io/nodebrew | perl – setup
export PATH=$HOME/.nodebrew/current/bin:$PATH

Checking nodebrew installed succesfully or not

nodebrew -v

4.2. Install Node.js using nodebrew

nodebrew install-binary stable

4.3 Make node.js usable

nodebrew use v*

(* == "the node.js version which you downloaded at 4.2")

4.4 Install npm

npm install -g [email protected]

Checking npm installed succesfully or not

npm -v

Finally you are ready to make a setlist

• Open the "auto_playlist_1.maxpat"!

1. You need to type your python path at python.js file.

1.1 Unlock presentation mode and double click the "node script python.js" patch.

1.2 Then, the python.js would open (type your python path, showing below)

2. Back to presentation mode and use it!! (How to GIF below!)

2.1 Push "On" button at section1.

• System status would change to "Running".

2.2 Drop start and goal track from the 'track_in_number' directory.

• Check the tracknames from 'tracknames.csv'.

2.3 Push "Load" button.

• If it doesn't work, try "No.4 Node.js" above. If it still doesn't work after that, let me know at GitHub Issues.