Create dataset v0.1

[sertansenturk]

One-line summary (optional) Create the first version of the dataset according to the investigation done in #21.

Reason We need the dataset to make structured experiments.

Design

Modify the Jupyter notebook created for #21 to generate the dataset. Fetch the relevant scores in all formats except pdf. The folder structure should be the same as SymbTr for convenience.

We may create a new git repo for the dataset, and link it to the main repo as a submodule (is there a better way?)

Acceptance Criteria

The scores selected and saved to a new repo.

Outcome

• Jupyter notebook to create the dataset
• The music score dataset v0.1

[sertansenturk]

The dataset repo is created: https://github.com/sertansenturk/otmm_generation_dataset/

• The Jupyter notebook walks through the score selection and the dataset creation steps. Currently, we are not uploading any music scores as we still haven't figured out the needed formats.

• metadata.json stores all the selected scores, as well as some additional metadata.

@magdalenafuentes, @hedonistrh : could you review the pull request: https://github.com/sertansenturk/otmm_generation_dataset/pull/1