run_compute() in multi-processor mode with 2 data sets of different size causes an error #1078
Description
issuing a run_compute() in multi-processor mode, letting phoebe do the mpi'ing, with 2 data sets of different size causes an error (seems like numpy throws the error.
in the following:
timesR = np.linspace(0, 1, 201)
timesV = np.linspace(0, 1, 202)
b = phoebe.Bundle.default_binary()
b.add_dataset('LC', time=timesR, dataset='lc01R')
b.add_dataset('LC', time=timesV, dataset='lc01V')
b.run_compute(model='tesParRV201')
if only one of the datasets is added, phoebe runs without issue.
if both are added with the same number of data points there is no issue
if both are added with different number of data points and we let phoebe do the multi-processor parsing
phoebe (numpy) throws an error
> ValueError: setting an array element with a sequence. The requested array has an inhomogeneous shape
> after 1 dimensions. The detected shape was (401,) + inhomogeneous part.
I have not tried w/ > 2 datasets nor using one LC and one RV.
Have tried run_solver() in a multi-processor environment. That has no issues.
a minimal script is attached as is the error output.