meelgroup/barbarik

The first efficient procedure to test whether a sampler is uniform

barbarik's Issues

No module named WAPS

From your last commit you said you deleted the WAPS directory. When I try to run
python3 barbarik.py --seed 1 --sampler 2 tests/uniform/blasted_case110.cnf

it get the error:

python3 barbarik.py --seed 1 --sampler 2 tests/uniform/blasted_case110.cnf 
Traceback (most recent call last):
  File "/Users/victor.miller/Programming/Git/barbarik/barbarik.py", line 10, in <module>
    from interfaces.cnf import *
  File "/Users/victor.miller/Programming/Git/barbarik/interfaces/cnf.py", line 8, in <module>
    from WAPS.waps import sampler as samp
ModuleNotFoundError: No module named 'WAPS'

Problems to run barbarik

Hi,

I've just downloaded barbarik. Following your instructions, I've tried to run it:

./barbarik.py --seed 1 --sampler SAMPLER_TYPE blasted_case110.cnf

Then, the following error message appeared:

Experiment: 0     of     1
constantFactor:38   boundFactor: 10.219408528680095   logBoundFactor: 3.3532397943494647  
tj: 1789   totalLoops: 2     beta: 0.8181818181818182 epsilon: 0.6 
numSolutions: 389   loThresh:56     hiThresh: 332
Traceback (most recent call last):
  File "./barbarik.py", line 734, in <module>
    barbarik()
  File "./barbarik.py", line 710, in barbarik
    ok, breakExperiment = exp.one_experiment(experiment, j, i, numExperiments, tj)
  File "./barbarik.py", line 606, in one_experiment
    **chainFormulaSetup = chainFormulaSetup(sampleSol, unifSol, self.numSolutions)**
**UnboundLocalError: local variable 'chainFormulaSetup' referenced before assignment**

The python version installed in my Ubuntu computer is 3.7.3

Thank you so much.

Best regards,
Ruben

Case for unique solution sampler

I have a question on the case of unique solution.
If a sampler only outputs unique solutions(which means one valid solution can only appears once in the output), whether barbarik works correctly? In other words, whether the output data of barbarik is meaningful?
Thanks : )

Combine with Barbarik2.

The Barbarik2 algorithm is designed to work on all weighted samplers and hence subsumes Barbarik. Thus we can replace the code of Barbarik with Barbarik2.

Code Issue

Those code seems to make the for loop end in the first loop each time. Is this a bug or not?

barbarik/Verifier.py

Lines 430 to 433 in f412674

	if (solMap[key] >= loThresh and solMap[key] <= hiThresh):
	return True
	else:
	return False