You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I found the offset/index stuff rather confusing and redundant. I think this could be reorganized a bit in order to make it more straight forward. Right now the index information is stored as pairs of offset+lenght as well as ranges in GraphStruct and again decentralized in the EdgeData/VertexData. Here are some ideas:
Each vortex and (in contrast to LighGraphs) each edge is identified by a unique index. Since the GraphStruct should not depend on the concrete underlying data structure. It should only allow quick access to
index of src and dst vertex for given edge
indices of all outgoing/incomming edges for given vertex
I found the offset/index stuff rather confusing and redundant. I think this could be reorganized a bit in order to make it more straight forward. Right now the index information is stored as pairs of
offset
+lenght
as well asranges
inGraphStruct
and again decentralized in theEdgeData
/VertexData
. Here are some ideas:Each vortex and (in contrast to LighGraphs) each edge is identified by a unique index. Since the
GraphStruct
should not depend on the concrete underlying data structure. It should only allow quick access tosrc
anddst
vertex for given edgeDuring the initialisation of
GraphData
all of the offsets could be calculated explicitly very cheap like... and once the 'views' aka
EdgeData
/VertexData
are constructed they are cached there.The text was updated successfully, but these errors were encountered: