Consolidates the arXiv identifier of an entry onto the DOI field.

arXiv articles are automatically assigned a stable DOI with the prefix 10.48550/arXiv., followed by the arXiv ID. Replacing an article with a new version does not generate a new DOI; the metadata of the existing DOI is updated at DataCite instead. The DOI is therefore the canonical, version-stable identifier and is preferred over the (redundant) eprint field.

See the arXiv DOI documentation.

Rules:

1. eprint holds an arXiv ID and there is no valid DOI (absent or unparseable) -> create the DOI from the eprint.
2. the DOI field holds an arXiv DOI matching the eprint -> drop the redundant eprint fields.
3. otherwise (e.g. a real publisher DOI next to an arXiv eprint, or no arXiv data) -> keep both.

This job is self-contained and independent of cleanup job ordering. The consolidation reads the already-normalized eprint and doi fields. When no consolidation is possible from the fields as they are, the arXiv identifier might still live in a field such as url/journal/note, from where EprintCleanup would move it into the eprint field. In that case this job applies EprintCleanup itself, but only after probing (on a copy) that doing so actually enables a consolidation - an entry without arXiv data is left untouched, regardless of whether EprintCleanup runs separately or at all.