Use logging process to exclude datasets from consideration in subsequent runs

[bryanmgee]

In some (many?) cases, not all of the possible remediation steps will be implementable (especially around authors w/o ORCIDs). To limit pressure on the API, an initial search result of the entire repository should look at the date that a dataset was last modified, and if that date is before the date of the last assessment sweep (i.e. dataset has not been modified since the last assessment), automatically excluded unless a new assessment or flag has been added.

[bryanmgee]

Sort of blocked by #12 and #13 but not fully - those tickets are ready to be closed with the PR, but the issue is whether datasets that already underwent re-curation might undergo re-curation again with an expanded criteria set, so the log would need to be configured in a way that is easy to map to 'flags' or 'fixes' or another type of variable that could be used to include/exclude. This is future dev work.