Metadata for new matches found during reprocessing
When re-processing with re-matching, new entries in published datasets might appear. Those will only have the initial metadata (e.g. the upload authors). It would make sense that they inherit some of the metadata from similar already matched entries. Metadata here refers to comment, references (URLs), datasets, and additional entry co-authors (only for old uploads).
- If the new matches are keys for an existing mainfile entry, we should simply copy the metadata from the mainfile entry
- If the new matches are new mainfiles ...
- we could copy metadata from the closest entry we find. Close here means, same dir, same parent dir, same parent, parent dir, ..., any other entry in the same upload.
- we could merge (set union, not sure how to merge comments) all metadata from all other entries in the same upload
- we could take the common metadata from all other entries (set intersection)
- we can apply different strategies to different metadata (e.g. treat references, comments, datasets, independently).