Plan going forward about perses HOT 7 CLOSED

We should do a bit of prep work to get this in a shape where multiple people can contribute at once.

Here's what I think we need to do:

• Rename the repo to be a useful Python package name (perses?)
• Reorganize repo to create a directory/module structure, like:

perses/
   perses/
      annihilation/ - alchemical factories for annihilation
      dualtopology/ - direct dual-topology alchemical transformation
      multitopology/ - core-based multi-topology alchemical transformation
      rjmc/ - reversible-jump based methods
      tests/ - automated tests
   devtools/
   examples/
   setup.py
   travis.yml
   README.md

• Add basic documentation in README.md just to describe where stuff is
• Add setup.py to make this Python installable
• Add devtools/ and travis.yml and enable automated testing [JDC can do this]
• Add examples/ with most basic working example of each type

After that, I think we're all set to divide up the work and flesh out some details of the rjmc approach:

• geometry engine [PBG]
• small molecule topology proposal engine [PBG] based on simplest approach (picking similar compounds from a specified compound library using MCSS?)
• protein topology proposal engine [JMB]
• NCMC switching using alchemical protocols [JDC]

I'll also clean up the multitopology and alchemical scripts to the point where we can have Julie figure out how to use them for residue mutations.

from perses.

Does this sound reasonable?

from perses.

That definitely sounds reasonable to me. Do you want me to organize the repo as you described above, or did you want to do it as you indicated in the pull request? Either is fine with me, I just don't want to leave the repo in a weird state.

from perses.

Can you take a stab at it? Feel free to do some direct commits if that is easiest, and raise an issue if you want any input/feedback!

from perses.

Sounds good!

from perses.

OK, we've stripped out the alchemical stuff and moved it to examol.

We might need to add in some additional tests and automatic running of the examples to make sure they work.

After that, we can get to work on dividing up the actual implementation!

from perses.

I'm going to close this for now to avoid issue clutter.

from perses.