Updating odds for decision making inside SMIRKY

[bannanc]

@davidlmobley and @camizanette

Here is the update to SMIRKY I'm using to run the new simulations. I realized I had some changes in my branch from before I left for the summer that I hadn't pushed. None of these were major.

Here is a list of changes in this pull request:

• Removing the file The-SMIRFF-force-field-format.md (addressed in issue #255 )
• Renaming some instances of "AlkEtOH" to "AlkEthOH" for consistency (discussed in openforcefield issue #36 )
• Changing temperature 0 to cause only increases in score to be accepted instead of the relative_temperature=1 (We discussed this in person last June (or earlier), however I can't find an issue related to it).
• Odds updated so "add new atom" happens less frequently
  • This also requires the atom being added to has at least one OR or AND decorator
  • It also gaurentees the new atom gets an ORbase
  • These requirements were there before. It should be noted that decorators can be removed which might have lead to some of the odd patterns I saw in the last set of simulations.

[bannanc]

I know this isn't passing tests, I have it fixed now I think, I'll push the new changes in a minute

[davidlmobley]

I unpinned the conda-build version, which helped. Looks like now I need to update the encrypted OpenEye license as the one which was used (Chodera lab version I believe) has expired.

[davidlmobley]

Tests pass. Off to the races.