deal with request producing very large answers
The following request makes the server crash:
[21_01_13_11_50_38] <searchPatternInGraphs#2> project_id=[Naija_Gold_Reference] pattern=[pattern { e: GOV -> DEP}] clusters=[e.label; GOV.upos; DEP.upos]
The answer contains one conll graph for each edge of the corpus and so is much bigger than the corpus itself.