a little performance improve on k-shortest

[BlankRain]

https://discuss.dgraph.io/t/performance-is-not-good-when-finding-the-paths-between-2-nodes-with-k-shortest/3533/2
https://discuss.dgraph.io/t/shortest-query-runs-out-memory/3695

If the data has cycle , for example
A to B , to is a reversed edge. so B ~to A.
if there is a path A to B to C to D .
it may add some route like A to B ~to A to B to C to D.
the cycling in A to B ~to A wastes memory.
and make the priorityQueue growing without stop.
that's one of the reasons for run out of memory.
In Djikstras algorithm , a DAG is better.
so if we break some cycles in adjacencyMap, we can make our path-finder go forward.
and save half of memory if A B both connected as A to B, B ~to A

---

This change is 

[CLAassistant]

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you all sign our Contributor License Agreement before we can accept your contribution.
1 out of 2 committers have signed the CLA.

✅ harshil-goel
❌ BlankRain
_{You have signed the CLA already but the status is still pending? Let us recheck it.}

[github-actions]

This PR has been stale for 60 days and will be closed automatically in 7 days. Comment to keep it open.

[github-actions]

This PR has been stale for 60 days and will be closed automatically in 7 days. Comment to keep it open.