grimme-lab
Cannot suppress output for a GFN-FF calculation in API
Moved here from xtb-python by @awvwgk
Describe the bug
print_level=0 for a GFN-FF calculation with ASE still prints calculations output, as opposed to print_level=0 for a GFN1-xTB calculation with ASE, where the output is suppressed.
Detailed description
Setup: version = "20.1" API_VERSION = "1.0.0" Python 3.8.3
Command in python script: ase_mol.set_calculator(XTB(method="GFN-FF", max_iterations=250, solvent="", print_level=0)) relax = LBFGS(ase_mol, logfile=None) relax.run(fmax=fmax, steps=max_steps)
Error:
Calculation output is not suppressed:
Expected behaviour
No output, just like running method="GFN1-xTB" with the same settings.
The GFN-FF C-API is currently experimental and the implementation has not been checked thoroughly for API usability. There are numerous issues that might interfere with API usage, like file IO for the topology and reading from files in the working directory (like charges). Printing to STDOUT is only a minor annoyance here.
The Python API offers access to the GFN-FF on the same terms as the C-API. Therefore, we better move this issue to the source upstream.
We can use https://github.com/minrk/wurlitzer to mute the output of GFNFF.
