Hi there, I noticed that when running DNest4 via the Python interface, all files except sampler_state.txt are properly saved into the basedir set when initializing the backend. For consistency with the rest of the DNest4 output, I think it would be good to have this file also saved into basedir, unless I am overlooking some reason this cannot be done.
For details on how I am using DNest4 -- I initialize a CSVBackend with basedir set, pass that to DNest4Sampler when creating the sampler object, run it with the sample method (and the enumerated for loop, as in examples), then call the sampler object's postprocess method, which looks to just call the postprocess function in analysis.py with the sampler object's backend.
Looking in the code, the CSVBackend sets basedir for all of the output files mentioned in the class, which doesn't include sampler_state.txt (see here). That file shows up here but that is the only mention of it that I find.
Hi there, I noticed that when running DNest4 via the Python interface, all files except sampler_state.txt are properly saved into the
basedir
set when initializing the backend. For consistency with the rest of the DNest4 output, I think it would be good to have this file also saved intobasedir
, unless I am overlooking some reason this cannot be done.For details on how I am using DNest4 -- I initialize a CSVBackend with
basedir
set, pass that to DNest4Sampler when creating the sampler object, run it with the sample method (and the enumerated for loop, as in examples), then call the sampler object's postprocess method, which looks to just call the postprocess function in analysis.py with the sampler object's backend.Looking in the code, the CSVBackend sets
basedir
for all of the output files mentioned in the class, which doesn't include sampler_state.txt (see here). That file shows up here but that is the only mention of it that I find.