llnl / sora Goto Github PK

Readme describes sympy as optional, program still runs but errors at the end returning this.

Traceback (most recent call last):
File "../sora.py", line 16, in
main()
File "/home/david/Desktop/sora/SoRa/sora/sora.py", line 256, in main
logger.printPopulation(0, hof)
File "/home/david/Desktop/sora/SoRa/sora/printLogger.py", line 91, in printPopulation
import sympy
ImportError: No module named sympy

Hi,

I tried to use SoRa after a fresh install in a virtualenv in Linux. Here is the setup of the virtual env:

• Python 2.7.15
• Numpy 1.14.3
• Sympy 1.1.1
• deap 1.2
• mpi4py 3.0.0.

When trying to run the two examples in the /test_sr/ folder, it raises errors related to an overflow in numpy.

This problem is solved with using Numpy 1.9 to 1.13.

Here is the error message obtained when using Numpy 1.14:

Traceback (most recent call last):
  File "../sora.py", line 16, in <module>
    main()
  File "/home/charles/sora/SoRa/sora/sora.py", line 232, in main
    toolbox.algorithm(pop)
  File "/home/charles/sora/lib/python2.7/site-packages/deap/gp.py", line 1045, in harm
    record = stats.compile(population) if stats else {}
  File "/home/charles/sora/lib/python2.7/site-packages/deap/tools/support.py", line 237, in compile
    record[name] = stats.compile(data)
  File "/home/charles/sora/lib/python2.7/site-packages/deap/tools/support.py", line 208, in compile
    entry[key] = func(values)
  File "/home/charles/sora/lib/python2.7/site-packages/numpy/core/fromnumeric.py", line 2957, in mean
    out=out, **kwargs)
  File "/home/charles/sora/lib/python2.7/site-packages/numpy/core/_methods.py", line 70, in _mean
    ret = umr_sum(arr, axis, dtype, out, keepdims)
FloatingPointError: overflow encountered in reduce

Current SoRa requires MPI. This is to handle demes/islands. I would rather have MPI be optional. In my opinion it would be OK if the demes feature is just optional based on the availability of MPI.