Closed F-I-D-O closed 4 years ago
I don't really see where the parameter is specified. Is it on the command line? The command line arguments are identical regardless of the input type (it is just a different Loader
class that is used for data loading, and all of them only take string inputFile
as constructor input - please see here, here and here).
Maybe the parameter is contained in the input files themselves?
I checked that and it appears that the precision can be set when importing the Geojson graph to xengraph format. The same should be possible in the C++ processor part for the adjacency matrix, or for all input types.
Does this mean we want to support floating point weights by rounding them based on this precision parameter?
Added in e1a1de8. I hope this is what you imagined. This line is the core of the change. Each weight is simply divided by the provided (optional) integer parameter.
Ok, it looks like a correct implementation.
the parameter is present in other methods/input formats