CLI to deal with data categories

[jelmervdl]

• List category names
• List all datasets in a category
• Give concatenated version of all datasets in a category

[graemenail]

So we don't have to do this:

python -c 'import json; ds=json.load(open("../../bitextor-mt-models/hin-eng/data/raw/categories.json"))["mapping"]["clean"]; print(" ".join(ds))'

[jelmervdl]

jq --raw-output .mapping.clean[] <categories.json