Tests are broken because of NCBI side metadata changes (again)

[bmpvieira]

Probably easy to fix (just update the tests data) but right now I can't look into it.
We could maybe as a long term solution use some mock framework to replicate the current NCBI expected behaviour and not rely and any specific metadata.

[bmpvieira]

This could be useful to mock NCBIs API:
http://unexpected.js.org/unexpected-mitm

It seems it can generate the mock automatically and also check that it's in sync with the real API.

[bmpvieira]

Using Flet/tape-nock instead to record and later mock NCBI connections.