[DO NOT MERGE] Pre-generate residue templates

[zhang-ivy]

When instantiating KinaseInhibitorTestSystem, this version of Perses currently loads a ffxml containing Imatinib (MOL0), Dasatinib (MOL1), Nilotinib (MOL2), and Bosutinib (MOL3). This is located at data/clinical-kinase-inhibitors.xml or https://github.com/choderalab/perses/blob/residue_templates/perses/data/clinical-kinase-inhibitors.xml

However, even though the ffxml seems to contain the right residue templates, the templates cannot be found/matched:

Traceback (most recent call last):
  File "test_ivy.py", line 9, in <module>
    testsystem = KinaseInhibitorsTestSystem()
  File "/lila/home/zhangi/perses/perses/tests/testsystems.py", line 1829, in __init__
    super(KinaseInhibitorsTestSystem, self).__init__(**kwargs)
  File "/lila/home/zhangi/perses/perses/tests/testsystems.py", line 1736, in __init__
    modeller.addSolvent(forcefield, model='tip3p', padding=9.0*unit.angstrom)
  File "/home/zhangi/miniconda3/envs/cenv/lib/python3.6/site-packages/simtk/openmm/app/modeller.py", line 376, in addSolvent
    system = forcefield.createSystem(self.topology)
  File "/home/zhangi/miniconda3/envs/cenv/lib/python3.6/site-packages/simtk/openmm/app/forcefield.py", line 1139, in createSystem
    raise ValueError('No template found for residue %d (%s).  %s' % (res.index+1, res.name, _findMatchErrors(self, res)))
ValueError: No template found for residue 1 (MOL0).  This might mean your input topology is missing some atoms or bonds, or possibly that you are using the wrong force field.

I'm not sure if this is because the ffxml was generated incorrectly or if there is an issue with template matching.

To modify the code so that it pre-generates templates again (instead of loading the ffxml):

1. In SmallMoleculeLibraryTestSystem of testsystems.py, uncomment lines 1717-1724
   https://github.com/choderalab/perses/blob/residue_templates/perses/tests/testsystems.py#L1717-L1724
2. In SmallMoleculeLibraryTestSystem of testsystems.py, modify lines 1713-1714 such that app.Forcefield only takes in gaff_xml_filename and tip3p.xml

Note: Use the latest commit from this openmoltools PR: choderalab/openmoltools#282
Note: The data/clinical-kinase-inhibitors.csv in this version of Perses contains only a subset of all the clinical kinase inhibitors found in master.

[jchodera]

Did you mean to uncomment this code and pre-generate the parameters, or did you not want to pre-generate the parameters?

[zhang-ivy]

That code is commented so whoever is looking at it can reproduce the error I was getting when loading the ffxml I already generated (https://github.com/choderalab/perses/blob/residue_templates/perses/data/clinical-kinase-inhibitors.xml). However, I wasn't sure if the problem was related to loading the template or generating the template. Therefore, to better understand the problem (and debug), I kept the code I used to generate the templates in there. Feel free to uncomment it if you think the problem occurs while generating the templates.

[jchodera]

Marking this as "DO NOT MERGE" so we can pull changes into a new PR.