Commit output data files to repo

[dfalster]

Currently we don't commit the generated database to git repo.

For multiple reasons I'm thinking we should revise this.

• can then use GitHub diff tool to see differences across versions, not only in raw but also also in output (these can then be linked to in logs)
• can be downloaded directly
• files will be archived in zenodo (doesn't accept outputs)

The downside is that the repo will be bigger.

Will & Remko do this in their respective data repos, e.g. taxonlookup and gmindatabase