* It should take list of pmids as an input and return the JSON of datasets descriptions by ids * It should use caching DB first and then fetch EMBL for additional information