As user, I can refine entry's data using a DOI. See https://docs.jabref.org/finding-sorting-and-cleaning-entries/getbibtexdatafromdoi#completing-information-based-on-doi-or-isbn for an example.

Then, for any field, it could be possible that information is existing locally ("Original entry") and from the DOI ("Entry from DOI").

Example:

However, sometimes, the local data is really wrong. In this example, the "year" (1109) and the "Entrytype" (Misc).

The dialog should select the data which could be more right. Here, "Entrytype" and "Year" should be selected from "Entry from DOI":

Task

1. Find out how the Merge Entries Dialog is called
2. Modify it to know that it knows that is called from "get bibliographic data from DOI" (or not)
3. If called from "get bibliographic data from DOI", then it should go through all fields existing in both entries and select the ones more "right". Implement it for entry type (misc; != misc) and year. - Also use "DOI" from the "DOI" fetcher (if code-wise clean enough)
4. Implement a class "FieldValueValidityComparator" implementing a comparator. The score returned is -, +, 0 depending if the left,, right value are "better" and 0 if the fields are equal or the comparator cannot tell
5. Inherit from it for YearFieldValueValidityComparator.
6. Use these comparators in the Merge Entries Dialog

Details

The dialog currently prioritizes the entry from the library. Sometimes, the data from the library is really wrong.

Example:

@Misc{,
  year      = {1109},
  doi       = {10.1109/IPDPSW55747.2022.00093},
  publisher = {IEEE},
}

1. Wrong entry type
2. Wrong year

When the merge entries dialog is started, it should select the more propriate entry type (everything else as MISC is better) and the better year (if year differs more than 10 years, the "new" year should be taken)