Add initial attempt at a plugin detector.

[uppittu11]

Here is an attempt at a writing a plugin detector (fixes #290). See foyer.plugins:collect_plugins()

Functions named foyer.forcefields.load_{plugin_name} are detected as plugins, but there is also a plugin_names argument to supply a list of plugin function names that don't fit that mold. The version, xml_file, and load_function are collected into a dict and returned.

I still don't know how to get the version. Maybe we could add it as an attribute of the plugin forcefield object when we load it up:

def load_OPLSAA():
    ff = get_forcefield(name='oplsaa')
    ff.version = "0.0.1"
    return ff

[codecov]

Codecov Report

Merging #292 into master will decrease coverage by 0.79%.
The diff coverage is 25%.

@@            Coverage Diff            @@
##           master     #292     +/-   ##
=========================================
- Coverage   86.55%   85.76%   -0.8%     
=========================================
  Files          14       15      +1     
  Lines        1220     1236     +16     
=========================================
+ Hits         1056     1060      +4     
- Misses        164      176     +12

[mattwthompson]

Any initial thoughts on how we should store this information? I see we want it to be accessible from the loader but I'm not sure how to go about doing that - maybe these functions should be expanded into (lightweight) classes?

[uppittu11]

Check out my plugins branch for the PFE repo. To demonstrate how FF versioning would work, I added some dummy forcefield versions and file organization for adding forcefields versions. Version and xml_path are returned when the forcefield is loaded, and can be easily accessed by the plugin detector in foyer:

>>> import foyer
>>> pfe = foyer.forcefields.load_PFE()
>>> pfe.version
'1.0.0'
>>> pfe.xml_path
['/Users/parashara/Documents/devel/git/forcefield_perfluoroethers/perfluoroethers/xml/current/PFE.xml']
>>> pfe = foyer.forcefields.load_PFE("0.0.2")
>>> pfe.version
'0.0.2'
>>> pfe.xml_path
['/Users/parashara/Documents/devel/git/forcefield_perfluoroethers/perfluoroethers/xml/versions/0.0.2/PFE.xml']

[mattwthompson]

Is there a way to semi-intelligently glob recursively, i.e. to get both '{}/xml/*xml' and '{}/*xml'? We probably don't need arbitrary recursion, maybe handling 2-3 cases is sufficient.

[rmatsum836]

For me this is what I get:

(Pdb) plugin_dir
'/Users/raymatsumoto/science/forcefields/forcefield_perfluoroethers/perfluoroethers/perfluoroethers.py'
(Pdb) 

So I'm unable to glob this path.

[uppittu11]

Ideally all of this information (like the xml file location and version) will be stored in the forcefield class, so we don't have to do any globbing. I think this is something @mattwthompson is working on