I am looking for best practices around scoring results being returned from dgraph full-text search. Is rolling your own score algo the norm?
1 Like
There’s an open feature request to add full-text scoring in Dgraph. Feel free to chime in on the thread.
https://github.com/dgraph-io/dgraph/issues/3211
As mentioned in the thread, Dgraph supports fuzzy matching based on Levenshtein distance with the trigram index.