Order match candidates by resolution priority

The single source of truth for the candidate sort: the fuzzy distance, whether the key has occurrence records, the remaining concept scores of score_candidates() in tier order, the occurrence count, then the nomenclatural-validity tiebreak, then the lowest taxonID. Pass group_col to sort within groups first, so the first row of each group is that group's best candidate.

Usage

candidate_order(candidates, scores = NULL, group_col = NULL)

Arguments

candidates

A data.frame accepted by score_candidates(), carrying a taxonID column and, when group_col is given, that column too.

scores

The score_candidates() output for candidates, when it has already been computed; recomputed when NULL.

group_col

Character or NULL. Column to sort by ahead of the scores.

Value

An integer permutation of seq_len(nrow(candidates)).