246 update python functions to adhere to simpler standards and pre format data

[ymahlich]

Addresses all points raised in #246

Refactor / rewrite of DatasetLoader into Dataset

cd.Dataset contains the following functions:

• dataset.format(data_type, ...): returns a formatted version of the datatype of interest. Currently supports ('transcriptomics', 'mutations', 'copy_number', 'proteomics', 'experiments', 'combinations', 'drug_descriptor', 'drugs', 'genes', 'samples')
• dataset.train_test_validate(args): splits the dataset into train, test & validation sets, and returns a @dataclass Split object containing all three datasets
• dataset.types(): returns a list of data types present in the dataset
• dataset.save(): saves the dataset object into a pickle file

rewrite / addition of functions in coderdata:

• cd.download(name, ...): refactor of cd.download_by_prefix that also allows for local_path & exist_ok arguments defining the directory the files should be downloaded into, and whether they should be overwritten if they already exist.
• cd.load(name, ...): returns a cd.Dataset object based on the dataset name given as argument. Also accepts parameters directory (defines which directory contains the datafiles to be loaded) & from_pickle (determines if a pickled Dataset, e.g. from Dataset.save() should be loaded).
• cd.list(): returns a list of available datasets (based on datasets.yml in the root directory of coderdata - potentially will be changed to a yaml that is stored in figshare in future builds)
• cd.version(): returns version strings of the package and dataset