Properly parse energy states and occupations

[pmrv]

Due to a faulty regex previously we threw away the signs of both energy states and occupations parsed from sphinx logs.

[coveralls]

Pull Request Test Coverage Report for Build 3402601774

• 3 of 3 (100.0%) changed or added relevant lines in 1 file are covered.
• 1 unchanged line in 1 file lost coverage.
• Overall coverage decreased (-0.007%) to 68.645%

---

Files with Coverage Reduction	New Missed Lines	%
pyiron_atomistics/lammps/structure.py	1	82.03%

Totals
Change from base Build 3401874913:	-0.007%
Covered Lines:	12085
Relevant Lines:	17605

---

💛 - Coveralls

[niklassiemer]

Now this only works on negative quantities?!

[niklassiemer]

wouldn't this allow, but not require a - sign? (Not tested!)

Suggested change

	np.array(re.sub("[^-0-9\. ]", "", "".join(fa)).split())
	np.array(re.sub("[^-?0-9\. ]", "", "".join(fa)).split())

[pmrv]

No, this regex matches everything [] that is not ^ a -, 0-9, \. (plain dot, \ is for escaping) or and replaces it by the empty string. This leaves just numbers and spaces, which are then "parsed". Because the regex previously also matched the - the parsed number were effectively only the absolute value.

Actually the whole regex business is too complicated, because in the sphinx log the lines are of the form some text: number number number. So removing that prefix would be enough.

[niklassiemer]

Well in that case, it souds like an easier and more stable solution to replace everything until a colon (being greedy it will be the last colon) by the empty string and parsing the numbers. Right now, a single - in the string would again flip the sign of the first number...