NLP Dataset Downloader

[alanakbik]

Some NLP datasets, such as the Universal Dependencies corpora, are freely available online. To make experimentation easier, add a feature for automatically fetching these datasets and putting them in a default folder structure.

Perhaps the existing NLPTaskDataFetcher can be extended such that:

corpus = NLPTaskDataFetcher.fetch_data(NLPTask.UD_ENGLISH)

When this is called, checks first if the dataset is already there. If not, triggers a download action in which the corresponding UD corpus gets downloaded and upacked into default folder structure.