Add option to only perform a fine optimization about arc HOT 2 CLOSED

Thanks! I agree this is necessary.
@dranasinghe once suggested we make it completely modular, so users can specify the exact job types they'd like and the order they'd like to have it executed in.

We already have the job_types argument, I think we should make its functionality such that the following specs will do what this issue requests:

job_types =  {
    'conformers': False,
    'opt': False,
    'fine_grid': True,
    'freq': True,
    'rotors': True,
    'sp': True
    'orbitals': False,
},

And if we'd like to go really fancy, we could also expand specific_job_type to encode the above dict, see something similar here.

I think the only required change to achieve the original goal requires only slight modifications in scheduler.run_opt_job
Perhaps do something like:

fine = self.job_types['fine'] and not self.job_types['opt']

and add fine=fine to the self.run_job() call

from arc.

Sorry I just missed this post before I opened the PR. I can implement something like what you are describing here, where fine=True, opt=False only runs one optimization at a fine level and fine=True, opt=True does the default behavior.

As for the future going forward, I agree with you and @dranasinghe that it would be a good idea to make it completely modular so that the user can decide which jobs to run in what order. Perhaps there is a way to balance this so that common or recommended approaches are called by default (or easily) if the user doesn't specify an order explicitly.

from arc.