Deterministic Volume Sampling
ipendlet opened this issue · 1 comments
We need a way to sample volumes in the code deterministically in order to have an expected code output. This should be implemented either as a fixed value (ideally specified by either the minimum or maximum of the portion) or as a seed state of the numpy random sampler. The code will likely replace the existing (rather underwhelming) optunity sampler with a numpy random number generator in qrandom.
There is a case to be made for both the fixed volume (which would help to resolve issue #56) and for the random seed state (would look more like an actual plate of experiments).
In either case this issue (issue #57) should be quickly followed up with a test case to validate the non-googleio processes locally during a debug run. As such the variables/toggles for fixed sampling should likely live in devconfig.py.
this can be accomplished by using manual sampling for an expect output. This DOES NOT test mathematica or the other sampling scripts. A likely resolution is to test those modules separately while using the rest of the funcationality in line with manual runs.