fix another CombinedData bug

[htz1992213]

Summary

Include a summary of major changes in bullet points:

• Fix 1: The CombinedData constructor previously creates new ff dict only according to the keys in the first mol in the mols llist, which may cause the ff information in the remaining mols lost. The new implementation create the ff dict using a set of keys containing all ff keys in mols
• Fix 2: The CombinedData constructor previously traverses the topo and ff info of mol in mols list without checking None type, which may cause error. The new implementation checks for None before iteration, and set ff/topo of the CombinedData to be None if all mol in mols list has None ff/topo.

Tests have been created accordingly.

[coveralls]

Coverage decreased (-0.6%) to 83.082% when pulling 8a30eab on htz1992213:master into c724e7b on materialsproject:master.

[mkhorton]

Thanks @htz1992213. Rather than edit the pylintrc with this specific case, I think you can add an in-line comment (something like # ignore).

Can you comment on what the issue was with ComputedStructureEntry._composition? This is a protected member anyway, so shouldn't be access outside the class.

[htz1992213]

Thanks for the advice @mkhorton! The access comes from the following line, I am not familiar with these classes so I will change the code with the # ignore workaround.

pymatgen/pymatgen/entries/computed_entries.py

Line 700 in 463e4f1

entry._composition /= factor

[rkingsbury]

Thanks @htz1992213. Rather than edit the pylintrc with this specific case, I think you can add an in-line comment (something like # ignore).

Can you comment on what the issue was with ComputedStructureEntry._composition? This is a protected member anyway, so shouldn't be access outside the class.

@mkhorton I will note that ComputedStructureEntry._composition seems to be causing a 1pylint failure in another PR as well, in which the changes are completely unrelated. Does this test pass in master?
https://github.com/materialsproject/pymatgen/pull/2135/checks?check_run_id=2754058779

[htz1992213]

Hey @mkhorton, could you please let me know if there are other things that I should address for the PR to be merged? Thanks!

[mkhorton]

This is all good, thank you Tingzheng!