Open nhpackard opened 3 years ago
Not sure of my diagnosis.
I don't see where the code is fixing Experiments1.csv. Don't see this happening in daptics.random_experiments_with_responses
.
The problem: in
expsim = daptics.simulate_experiment_responses(experiments)
daptics.export_csv(csv_experiments_file, expsim, True)
The return value expsim
does not have the same form as experiments
, and so can't be exported with daptics.export_csv
.
Instead, needs to be
expsim = daptics.simulate_experiment_responses(experiments)
expsim = expsim['simulateResponses']['table']
daptics.export_csv(csv_experiments_file, expsim, True)
If no objections, I am going to change 'daptics.simulate_experiment_responses(experiments)to return an experiments dictionary of the same form as the arg
experiments`
@nhpackard If you want to put up a pull request with these changes I will review.
Will do. Now distracted by workshop.
(cf. #32)
To generate simulated responses and export them I execute:
This works to the extent that csv_experiments_file has the right simulated results
But then when I request next gen design with
I get error
My diagnosis is that the call
daptics.random_experiments_with_responses
installs the random experiments inExperiments1.csv
, and then there is a skew between those and the simulated response experiments incsv_experiments_file
.Proposed solution: dispense with current
daptics.simulate_experiment_responses
, create newdaptics.simulated_experiments_with_responses
that is exactly likedaptics.random_experiments_with_responses
(including installation ofExperiments1.csv
on backend), except using simulated responses instead of random.Actually better names would be:
I will work on this in new branch...