[WIP] Control where additional nprop steps are added in the lambda protocol

[sgill2]

Description

Adds a propRegion option to the integrator to allow focusing the additional propagation steps either in the sterics (in the middle) or the electrostatics (at either ends of the lambda protocol)

Todos

Notable points that this PR has either accomplished or will accomplish.

• Adds a propRegion argument to the integrator
• Updates TestNCMCSteps to account for the change in region
• Adds tests to ensure that the steps are being added to the appropriate region

Status

• Ready to go

[codecov]

Codecov Report

Merging #156 into master will increase coverage by 1.36%.
The diff coverage is 79.16%.

@@            Coverage Diff            @@
##           master    #156      +/-   ##
=========================================
+ Coverage   59.53%   60.9%   +1.36%     
=========================================
  Files           7       7              
  Lines        1478    1504      +26     
  Branches      266     272       +6     
=========================================
+ Hits          880     916      +36     
+ Misses        498     491       -7     
+ Partials      100      97       -3

Impacted Files	Coverage Δ
blues/reporters.py	41% <100%> (+1.76%)	⬆️
blues/utils.py	54.23% <70.58%> (+5.16%)	⬆️
blues/integrators.py	89.52% <82.14%> (-1.49%)	⬇️
blues/simulation.py	76.9% <0%> (+1.57%)	⬆️

[davidlmobley]

@sgill2 - what needs to be done to finalize this? I believe it's working for @bergazin ?

Also this would INCREASE code coverage which is good. :)

[bergazin]

Yes, and then perhaps including some textfile that lists what changes to the alchemical function someone could use to control how much time is spent on the sterics/electrostatics ?

[davidlmobley]

In other words the documentation needs updating... :)

[bergazin]

ie
40% electrostatics, 60% sterics (Default)
Lambda sterics: min(1, (1/0.3)abs(λ-0.5))
Lambda electrostatics: step(0.2-λ) - 1/0.2λstep(0.2-λ) + 1/0.2(λ-0.8)*step(λ-0.8)

65% electrostatics, 35% sterics
Lambda sterics: min(1, (1/0.175)abs(λ-0.5)
Lambda electrostatics: step(0.325-λ) - 1/0.325λ step(0.325-λ) + 1/0.325(λ-0.675)*step(λ-0.675)

90% electrostatics, 10% sterics
Lambda sterics: min(1, (1/0.05)abs(λ-0.5)
Lambda electrostatics: step(0.45-λ) - 1/0.45λstep(0.45-λ) + 1/0.45(λ-0.55)*step(λ-0.55)

will be usefull if someone needs to use it for water....

[stevenayoub]

@davidlmobley
This can be valid merge request - the PR adds propRegion is a simple but useful extension to AlchemicalExternalLangevinIntegrator - to allow focusing the additional propagation steps either in the sterics (in the middle) or the electrostatics (at either ends of the lambda protocol).

I was able to merge it with my local master branch without any conflicts.