Occasional CClib parsing error in QM using Mopac

[connie]

This error has popped up a few times in my jobs, but is not easily reproducible.

Using the cclib parser on the output file of interest it works correctly, and when I rerun the job keeping the original QM files it also seems to run fine the second time around. However, I have seen this error come up in other jobs.

The best fix may be to rerun MOPAC through a second attempt?

---

2 radicals on C1CC(CCC1(C)[O])[C](C)C exceeds limit of 0. Using HBI method.
Output file /home/connieg/QMfiles/CMLYGGFIXXLYQT-UHFFFAOYSA.out does not (yet) exist.
Trying MopacMolPM3 attempt 1 of 10 on molecule C1(CCC(C(C)C)CC1)(C)O.
Successful MOPAC quantum result found in /home/connieg/QMfiles/CMLYGGFIXXLYQT-UHFFFAOYSA.out
Traceback (most recent call last):
  File "/home/connieg/Code/RMG-Py/rmg.py", line 136, in <module>
    cProfile.runctx(command, global_vars, local_vars, stats_file)
  File "/usr/lib/python2.6/cProfile.py", line 49, in runctx
    prof = prof.runctx(statement, globals, locals)
  File "/usr/lib/python2.6/cProfile.py", line 140, in runctx
    exec cmd in globals, locals
  File "<string>", line 1, in <module>
  File "/home/connieg/Code/RMG-Py/rmgpy/rmg/main.py", line 329, in execute
    self.initialize(args)
  File "/home/connieg/Code/RMG-Py/rmgpy/rmg/main.py", line 317, in initialize
    self.reactionModel.enlarge([spec for spec in self.initialSpecies if spec.reactive])
  File "/home/connieg/Code/RMG-Py/rmgpy/rmg/model.py", line 693, in enlarge
    spec.generateThermoData(database, quantumMechanics=self.quantumMechanics)
  File "/home/connieg/Code/RMG-Py/rmgpy/rmg/model.py", line 124, in generateThermoData
    tdata = database.thermo.estimateRadicalThermoViaHBI(molecule, quantumMechanics.getThermoData)
  File "/home/connieg/Code/RMG-Py/rmgpy/data/thermo.py", line 788, in estimateRadicalThermoViaHBI
    thermoData = stableThermoEstimator(saturatedStruct)
  File "/home/connieg/Code/RMG-Py/rmgpy/qm/main.py", line 147, in getThermoData
    thermo0 = qm_molecule_calculator.generateThermoData()
  File "/home/connieg/Code/RMG-Py/rmgpy/qm/molecule.py", line 252, in generateThermoData
    self.qmData = self.generateQMData()
  File "/home/connieg/Code/RMG-Py/rmgpy/qm/mopac.py", line 205, in generateQMData
    result = self.parse() # parsed in cclib
  File "/home/connieg/Code/RMG-Py/rmgpy/qm/mopac.py", line 144, in parse
    qmData = CCLibData(cclibData, radicalNumber+1)
  File "/home/connieg/Code/RMG-Py/rmgpy/qm/qmdata.py", line 79, in __init__
    molecularMass = (cclib_data.molmass,'amu'),
AttributeError: 'ccData' object has no attribute 'molmass'

[rwest]

f31693b may help with the debugging, but I can't quite see why it would be failing this way, unless it's an i/o problem, eg. the file has not finished being written yet when you start to parse it. Are you running this on a networked file share, eg. from your home user space on a cluster? I wonder if there's a way to wait until a file is no longer open by any other process, before parsing it.

Seeing this also, when importing mechanims.

  File "/home/goktekin.e/Code/RMG-Py/importChemkin.py", line 2162, in <module>
    mm.main()
  File "/home/goktekin.e/Code/RMG-Py/importChemkin.py", line 1484, in main
    self.limitEnlarge(self.speciesDict_rmg[labelToProcess])
  File "/home/goktekin.e/Code/RMG-Py/importChemkin.py", line 1317, in limitEnlarge
    spec.generateThermoData(database, quantumMechanics=rm.quantumMechanics)
  File "/home/goktekin.e/Code/RMG-Py/rmgpy/rmg/model.py", line 149, in generateThermoData
    thermo0 = quantumMechanics.getThermoData(molecule) # returns None if it fails
  File "/home/goktekin.e/Code/RMG-Py/rmgpy/qm/main.py", line 147, in getThermoData
    thermo0 = qm_molecule_calculator.generateThermoData()
  File "/home/goktekin.e/Code/RMG-Py/rmgpy/qm/molecule.py", line 221, in generateThermoData
    self.qmData = self.generateQMData()
  File "/home/goktekin.e/Code/RMG-Py/rmgpy/qm/mopac.py", line 216, in generateQMData
    result = self.parse() # parsed in cclib
  File "/home/goktekin.e/Code/RMG-Py/rmgpy/qm/mopac.py", line 146, in parse
    qmData = CCLibData(cclibData, radicalNumber+1)
  File "/home/goktekin.e/Code/RMG-Py/rmgpy/qm/qmdata.py", line 87, in __init__
    raise e
AttributeError: 'ccData' object has no attribute 'rotcons'

and

  File "/home/goktekin.e/Code/RMG-Py/importChemkin.py", line 2162, in <module>
    mm.main()
  File "/home/goktekin.e/Code/RMG-Py/importChemkin.py", line 1484, in main
    self.limitEnlarge(self.speciesDict_rmg[labelToProcess])
  File "/home/goktekin.e/Code/RMG-Py/importChemkin.py", line 1317, in limitEnlarge
    spec.generateThermoData(database, quantumMechanics=rm.quantumMechanics)
  File "/home/goktekin.e/Code/RMG-Py/rmgpy/rmg/model.py", line 130, in generateThermoData
    tdata = database.thermo.estimateRadicalThermoViaHBI(molecule, quantumMechanics.getThermoData)
  File "/home/goktekin.e/Code/RMG-Py/rmgpy/data/thermo.py", line 803, in estimateRadicalThermoViaHBI
    thermoData = stableThermoEstimator(saturatedStruct)
  File "/home/goktekin.e/Code/RMG-Py/rmgpy/qm/main.py", line 147, in getThermoData
    thermo0 = qm_molecule_calculator.generateThermoData()
  File "/home/goktekin.e/Code/RMG-Py/rmgpy/qm/molecule.py", line 221, in generateThermoData
    self.qmData = self.generateQMData()
  File "/home/goktekin.e/Code/RMG-Py/rmgpy/qm/mopac.py", line 216, in generateQMData
    result = self.parse() # parsed in cclib
  File "/home/goktekin.e/Code/RMG-Py/rmgpy/qm/mopac.py", line 144, in parse
    cclibData = parser.parse()
  File "/home/goktekin.e/Code/RMG-Py/external/cclib/parser/logfileparser.py", line 221, in parse
    self.extract(inputfile, line)
  File "/home/goktekin.e/Code/RMG-Py/external/cclib/parser/mopacparser.py", line 142, in extract
    self.inputatoms.append(symbol2int(broken[1]))
  File "/home/goktekin.e/Code/RMG-Py/external/cclib/parser/mopacparser.py", line 20, in symbol2int
    return t.number[symbol]
KeyError: '143'

This is also somewhat random and only when running on a networked file system (on an HPC cluster).

[rwest]

I think this is easily (though not very beautifully) addressed by inserting a 1 second pause to wait for the file to be written to disk, as @keceli tried in 56ca670 (did it work, Murat?)

[keceli]

It was working 6 months ago, I haven't checked it since then. While it is
as ugly as it gets, I don't know any other portable solution. There are two
related things I'd suggest: 1) Use a folder (to store thermo files) in a
scratch disk that no other RMG simulations access 2) Use thermoEstimator.py
(the one in scoop branch is parallel), after generating the species list
that you want to run QM calculations. I can try to update this branch with
the latest version of scoop and RMG next week.

Murat