Develop the `Dataset` class in context.py

[ubdbra001]

The Dataset class is currently a stub and needs additional work, specifically:

• modalities method - This should walk the tree, identify different modalities in the dataset (e.g. anat, func, eeg, etc), and return them as a list
• datatypes method - This should return the different modalities (e.g. mri, eeg, etc) as a list

Two questions for @effigies:

1. Should the datatypes method use the modalities and a mapping to generate its result, or some other way?
2. Have I missed anything here?

[effigies]

Swap modalities and datatypes. anat is a datatype, mri is a modality. Otherwise, that sounds right.