Refactor datasets.py module

[ninamiolane]

What?

• Create a class OSFProject in ioSPI/datasets.py that refactors the code in:
  • the class Project in ioSPI's PR#70
  • the class OSFUpload in ioSPI/datasets.py module
  • the function upload_dataset_from_files of simSPI's tem_tutorial.ipynb
• Remove the function upload_dataset_from_files of simSPI's tem_tutorial.ipynb
• Adapt the tutorial ioSPI/notebooks/download_and_upload_with_osf.ipynb from ioSPI that upload or download data using the datasets.py module.
• Add corresponding unit-tests.
• Use pre-commit (see README) to make sure that your code follows style guidelines.

Why?

There are several reasons for this task:

• Functionality in ioSPI and simSPI should not be entangled. The fact that they are, i.e. simSPI is using ioSPI, complicates the development: as any PR on ioSPI potentially requires a refactoring PR on simSPI.
• Functionalities that download or upload data with OSF use different logics that are hard to grasp for newcomers: this task creates one Python class only, i.e. one single point of entry with one logic.

The library ioSPI can only be useful if users can understand it quickly. This task is urgent and very much needed.

Where?

The class OSFProject should go into the ioSPI/datasets.py module.

This class replaces the classes Project and OSFUpload from ioSPI/datasets.py.

How?

The new class OSFProject should have (at least):

• a method ls: listing the files of a OSF project
• a method download
• a method upload

and follow the logic that one object instantiated from this class corresponds to one OSF project.

[bongjinkoo]

I also added remove method. Btw, osf ls does not give us all items in the project. Somehow it doesn't list ones created by the obsolete OSFUpload. I suspect it is because it was added as a child node and osf ls cannot list them, or something. Move on for now, as we will only use osf upload?