`KeyError` happened when use `Cp2kBaseWorkChain` to do a NEB calculation

[xjf729]

Environment

aiida-core version 1.2.1
aiida-cp2k version 1.1.0

Error Report

When I use Cp2kBaseWorkChain for a NEB calculation, error happened in parse step.

*** 1 LOG MESSAGES:
+-> REPORT at 2020-06-04 03:17:35.625606+00:00
 | [1474|Cp2kCalculation|on_except]: Traceback (most recent call last):
 |   File "/anaconda3/envs/aiidav1/lib/python3.7/site-packages/plumpy/process_states.py", line 220, in execute
 |     result = self.run_fn(*self.args, **self.kwargs)
 |   File "/anaconda3/envs/aiidav1/lib/python3.7/site-packages/aiida/engine/processes/calcjobs/calcjob.py", line 263, in parse
 |     exit_code = execmanager.parse_results(self, retrieved_temporary_folder)
 |   File "/anaconda3/envs/aiidav1/lib/python3.7/site-packages/aiida/engine/daemon/execmanager.py", line 439, in parse_results
 |     exit_code = parser.parse(**parse_kwargs)
 |   File "/anaconda3/envs/aiidav1/lib/python3.7/site-packages/aiida_cp2k/parsers/__init__.py", line 39, in parse
 |     returned = self._parse_trajectory()
 |   File "/anaconda3/envs/aiidav1/lib/python3.7/site-packages/aiida_cp2k/parsers/__init__.py", line 91, in _parse_trajectory
 |     return StructureData(ase=Atoms(**parse_cp2k_trajectory(output_string)))
 |   File "/anaconda3/envs/aiidav1/lib/python3.7/site-packages/ase/atoms.py", line 208, in __init__
 |     self.new_array('numbers', symbols2numbers(symbols), int)
 |   File "/anaconda3/envs/aiidav1/lib/python3.7/site-packages/ase/symbols.py", line 20, in symbols2numbers
 |     numbers.append(atomic_numbers[s])
 | KeyError: '.E+'

Possible Reason

I think the reason maybe in def parse_cp2k_trajectory code.

Because you parse the COORD section in *-1.restart file, however, the COORD section in *-1.restart of NEB don't have information of chemical symbols, then E+ is regarded as chemical symbol. The COORD section in *-1.restart of NEB like follows.

&COORD
    6.7740185951757121E+00    5.9999999787312079E+00    6.2143687594147634E+00
    5.2259808444671396E+00    6.0000000393828854E+00    6.2143688063089808E+00
    4.8129448743603360E+00    5.9999999500883883E+00    5.1990521657080473E+00
    4.8126645585662970E+00    6.8792360991874624E+00    6.7218712281724748E+00
    4.8126646095750898E+00    5.1207639721373717E+00    6.7218712678906591E+00
    7.1870544024378509E+00    5.9999999775529664E+00    5.1990520561417144E+00
    7.1873347582418425E+00    6.8792360937843435E+00    6.7218711966487747E+00
    7.1873347372392553E+00    5.1207638983179260E+00    6.7218712597601877E+00
&END COORD

I think parse the last frame of *-pos-1.xyz file maybe a better choice.

[dev-zero]

I tend to agree here: whenever we can let CP2K write a (more or less) standardized file format we should do that and use a parser for the specific file. The other option would be to use the cp2k-input-tools to parse the restart file (which is an input file).