Improve the selection strategy in convert()

[dprada]

When the input argument of convert() is used... the initial molecular system is converted twice sometimes to first get the atom list corresponding to the selection, and second, make the conversion. This is notable when:

import molsysmt as msm
msm.convert('181L', selection='molecule_type not in ["water", "ion"]')

In this last case, the PDB ID 181L is downloaded twice:

• One to make the selection and get the atom indices list
• One to make the conversion, invoking the method form.string_pdb_id.to_molsysmt_MolSys

This makes no sense, and it should be improved. The selection could be done inside form.string_pdb_id.to_molsysmt_MolSys.