Situational_Session_Generator

This is the code for the paper 'Exploiting device and audio data to tag music with user-aware listening contexts'. If you use the code in your research, please cite the paper as:

Karim M. Ibrahim, Elena V. Epure, Geoffroy Peeters, and Gaël Richard. 2022. Exploiting device and audio data to tag music with user-aware listening contexts. [Under Revision]

Instructions

We recommend using Docker for reproducing the results of the experiments.

Clone the repository then follow these instructions.

1. To build the docker image run:

docker build -t sit-gen-image docker

This will automatically clone the repository inside the docker container

2. Run the container with:

nvidia-docker run -ti --rm --memory=20g --name=sit-gen-container -p 8888:8888 sit-gen-image

Note: you might need to adjust the memory or port according to your machine.

3. Download the dataset available at Zenodo.

4. Download the audio previews inside the docker using the Deezer API. Then compute the melspectrograms (recommended using librosa) in the direcory "/src_code/repo/spectrograms/" with the following parameters:

"n_fft": 1024,
"hop_length": 1024,
"n_mels": 96

5. From the downloaded dataset, create the splits using the 'fold' variable and put them accordingly in the 'groundtruth'. Note: the subdirectories are empty and are meant to guide on how the directories structure should be.

6. To reproduce our experiments, either run the python file:

python run_all.py

The model evaluation results exist in the output directory in 'src_code/repo/experiment_results/'

Or go through the code block by block in the accompanied notebook 'Notebook.ipynb'. To run jupyter use:

jupyter notebook --allow-root --ip=$(awk 'END{print $1}' /etc/hosts) --no-browser --NotebookApp.token= &

Then access the notebook through the designated port and run all code blocks in order of the following notebook

Items

This repository contains the following item:

• 'run_all.py' the script for rerunning all experiments in the paper and producing the complete results.
• 'Notebook.ipynb' Same as the script, but allows for running the code block by block and access the intermediate results.
• 'requirements.txt' contains the required packages to run the code. Only needed in case docker is not used.

The repository contains two directories:

• 'Docker' contains the script for building the docker image and installing the requirements.
• 'groundtruth contains the subdirectories in the same format needed to rerun the code. Note: Dataset need to be downloaded seperatly before splitting and moving to the corresponding subdirectory.

Acknowledgment

This work has received funding from the European Union’s Horizon 2020 research and innovation programme under the Marie Skłodowska-Curie grant agreement No. 765068.