Pandas/Numpy conflict causing issues in casm-learn

[jcthomas]

Was contacted by a user with the following problem, so I'm opening an issue. It seems related to a known pandas/numpy conflict. The user was able to resolve by updating pandas, which also updated numpy. I haven't been able to recreate the issue myself, but at least wanted to document it if anyone else encounters the same.

When I run casm-learn -s fit_1_ga.json the first time it works and writes out all files. I can then plot my predicted and DFT hull from them. When I run it a second time with the previously generated fit_1_ga_halloffame.pkl I get
Loading fit_1_ga.json
# Hall of Fame size: 25

Loading Hall of Fame: fit_1_ga_halloffame.pkl
Traceback (most recent call last):
File "/usr/local/bin/casm-learn", line 119, in
hall.update(existing_hall)
File "/usr/lib/python2.7/dist-packages/deap/tools/support.py", line 523, in update
self.insert(population[0])
File "/usr/lib/python2.7/dist-packages/deap/tools/support.py", line 551, in insert
item = deepcopy(item)
File "/usr/lib/python2.7/copy.py", line 190, in deepcopy
y = _reconstruct(x, rv, 1, memo)
File "/usr/lib/python2.7/copy.py", line 334, in _reconstruct
state = deepcopy(state, memo)
File "/usr/lib/python2.7/copy.py", line 163, in deepcopy
y = copier(x, memo)

... lots more deepcopy errors...

y.append(deepcopy(a, memo))
File "/usr/lib/python2.7/copy.py", line 174, in deepcopy
y = copier(memo)
ValueError: Iteration of zero-sized operands is not enabled
To refit I then have to delete the fit_1_ga_halloffame.pkl and then I can rerun, is that intended? This error occurs even when I run the same fit twice, i.e. not changing anything in the input file for casm-learn.

I suspect we might not be able to do anything about it beyond advising people to update to the latest version of pandas an numpy.