meelgroup / weighted-to-unweighted Goto Github PK

Hi,
I'm using this tool but the results of the computation is not clear to me. Consider the program shown in the README

$ cat my.cnf
p cnf 2 1
c ind 1 2 0
1 2 0
w 1 0.9
w 2 0.5

I convert it with the python script of this repo and I get as normalization factor 2^8.
After the conversion to UMC, I get (using dsharp as model counter) 2^9 models.
So, 2^9/2^8 = 2.
However, If i use a weighted model counter, such as ADDMC on the original WMC task, I get as result 0.95, wich is the value I expect. What are the reasons of such different results? The fact that the conversion from WMC to UMC is approximate? Or am I missing something? Even if I increase the precision with the flag --prec, I still get 2.

Thanks