Litter-getter

* Standardize author name cleanup across all import types
* Rename `authors_list` to `authors`; all return a list of authors
* Pubmed and HERO import fetch requests expect a list of integer values, not strings
* Pubmed and HERO ids returned in content are now integer instead of strings
* For HERO and RIS, don't save json field as string, but instead as JSON-compatible dictionary
* Python 3.6 or higher is required

* Make Pubmed API key optional

* Use new HERO URL
* Use Pubmed API keys
* Code style enforcement with black
* Simplify tooling dependencies (no more travis, readthedocs, etc)

* Fix bug in capturing title text with child elements

* Fix bug in capturing abstract text with child elements

* Bump RISparser version
* Prevent error if fetch called before count (breisfeld)