* Don't re-use gremlin query for edges, fixes#184
Moved the deduping and size limiting to the server
Now querying specifically for the IDs of the vertices from the gremlin query
a) Size we're limiting size in server, we only query up to max vertices
b) Nevertheless we still need to chunk up the query
c) Added retry logic for RequestRateTooLarge, which can happen because we're querying for a lot of edges in multiple queries in parallel
This cleans up and simplified the client-size code
Also means the edge query should run much faster for many cases
* Add some graph tests
* Add some simple tests
* minor refactoring
* PR comments
* Fix comments for Vertex
Eric Jizba
Neelima Potharaj
Stephen Weatherford (MSFT)
Stephen Weatherford
Prashanth
Sandeep Somavarapu
Microsoft GitHub User