Source data validation issues

[strogonoff]

Describe the issue

Stefano outlined discrepancies between data sources and relaton-py expectations here.

The discrepancies arise because (1) previously invalid YAML was updated to match Relaton RNC/LutaML data specification (where relaton-py has come to rely on YAML to be invalid), or (2) data specification updated, YAML updated accordingly, but relaton-py was not updated.

These discrepancies lead to issues like: #277, #250, #281.

Fronts on which they could be attacked:

• [ ] Normalizing abnormal source data in BibXML service (fixes partly implemented in #282)
• [ ] Making it so that abnormal source data doesn’t cause issues (not always realistic, but exploring options)
• [ ] Pinning Relaton gems by their minor version part in GHAs (and relying on gem maintainers to avoid introducing breaking data format changes within the same minor version; I was informed that’s how it already works)
• [ ] Updating Relaton specification to support versioning (medium term, we shouldn’t do it in rush)

Code of Conduct

• [X] I agree to follow the IETF's Code of Conduct