Implement YAML-based curation

[edkerk]

Follow similar approach as is used in human-GEM, keep the YAML-file as the version to work on in the development branch. This greatly simplifies contributing to model curation, as you can directly change this file, instead of having to use RAVEN, cobrapy or other packages. Using such packages is still possible (and will often make life easier), but is not required.

[mihai-sysbio]

I believe this also implies the use of TSV files for annotation.