Neptune Gremlin query to recommend users based on common AND positive rated content

Question

The Gremlin query pasted below returns an ordered list of friends of a given user - the ordering is descending, based on "how many of the same movies have we rated". I would like to only compare those movies where rated.score > 5 (on a 10 point scale) the goal is to sort the results by "how many of the same movies have we rated positively". Thanks in advance!

  g.V('a2661f57-8aa7-4e5c-9c89-55cf9bxxxxx').as('self').
  sideEffect(out('rated').store('movies')). 
  out('friended').
  group(). 
    by(). 
    by(out('rated').where(within('movies')).count()). 
  order(local). 
    by(values,desc). 
    unfold().
  select(keys).
  project('id','label','username').
    by(id).
    by(label).
    by('username')

Answer 1

I may be missing some context of the "score". Are you just using the count of ratings between common movies as the "score"? Or is this a property on the 'rated' edge?

If so, you would just add a where() clause to your query to only return the results that are greater than 5.

where(select(values).where(is(gt(5)))).

g.V('a2661f57-8aa7-4e5c-9c89-55cf9bxxxxx').as('self').
  sideEffect(out('rated').store('movies')). 
  out('friended').
  group(). 
    by(). 
    by(out('rated').where(within('movies')).count()). 
  order(local). 
    by(values,desc). 
    unfold().
  where(select(values).where(is(gt(5)))).   //added filter
  select(keys).
  project('id','label','username').
    by(id).
    by(label).
    by('username')

If you wanted to filter on a ratings score that is included on a ratings edge, you would need to filter for this condition earlier in the query using:

where(outE('rated').values('score').is(gt(5))).

g.V('a2661f57-8aa7-4e5c-9c89-55cf9bxxxxx').as('self').
  sideEffect(out('rated').store('movies')). 
  out('friended').
  group(). 
    by(). 
    by(outE('rated')
            .where(values('score').is(gt(5)))  //filter on score
       .inV()
           .where(within('movies')).count()). 
  order(local). 
    by(values,desc). 
    unfold().
  select(keys).
  project('id','label','username').
    by(id).
    by(label).
    by('username')