JanusGraphVertexDeserializer discards vertices without properties and edges

The `JanusGraphVertexDeserializer` has the following logic in the `readHadoopVertex` method that causes vertices without edges and properties to be ignored.

```
/*Since we are filtering out system relation types, we might end up with vertices that have no incident relations.
This is especially true for schema vertices. Those are filtered out. */
if (!tv.edges(Direction.BOTH).hasNext() && !tv.properties().hasNext()) {
log.trace("Vertex {} has no relations", vertexId);
return null;
}
```

```
gremlin> graph = JanusGraphFactory.open('conf/janusgraph-cassandra.properties')
==>standardjanusgraph[cassandrathrift:[127.0.0.1]]
gremlin> g = graph.traversal()
==>graphtraversalsource[standardjanusgraph[cassandrathrift:[127.0.0.1]], standard]
gremlin> g.V().count()
19:56:44 WARN org.janusgraph.graphdb.transaction.StandardJanusGraphTx - Query requires iterating over all vertices [()]. For better performance, use indexes
==>100
gremlin>
gremlin> graph = GraphFactory.open('conf/hadoop-graph/read-cassandra.properties')
==>hadoopgraph[cassandrainputformat->gryooutputformat]
gremlin> g = graph.traversal().withComputer(SparkGraphComputer)
==>graphtraversalsource[hadoopgraph[cassandrainputformat->gryooutputformat], sparkgraphcomputer]
gremlin> cnt = g.V().count().next()
==>0e 0:===================================> (3 + 2) / 5]
gremlin> cnt
==>0
```

Start in JanusGraphVertexDeserializer.readHadoopVertex and inspect the guard that returns null for vertices without edges or properties. Reproduce the discrepancy using conf/janusgraph-cassandra.properties and conf/hadoop-graph/read-cassandra.properties, then verify that the Hadoop traversal no longer drops those vertices and that the vertex count reflects the source graph.

Written by the indexing model from the issue text.