Deduplicate a Type

Anybody have any pointers for deduplicator/combiner scripts? I am bringing my former system to current sync with Dgraph after initial import almost a year ago now. I have duplicates here and there that need to be fixed. I understand that a dedup script would be very custom as every dataset has unique requirements on what is and what is not a duplicate record. I would like to be able to define a few fields and edges to compare to see if the fields are equal.