Skip to content

Minutes 2017 11 13

Jump to bottom
Oliver Beckstein edited this page Nov 14, 2017 · 4 revisions

Virtual Meeting 2017-11-13

All discussion by e-mail. Below an edited summary.

@orbeckst

Let me briefly summarize what happened since the last call. This might be sufficient and we could skip the call on Monday, unless you want to discuss in person or if there are other developments to be shared.

  1. Roadmap: https://github.com/alchemistry/alchemlyb/wiki/Roadmap was pretty much agreed upon (feel free to correct me or the roadmap)

  2. Shuai did a lot of work!

  3. alchemtest: made licensing requirements for contributing test data explicit: we require an open license http://opendefinition.org/licenses/#recommended-conformant-licenses but effectively CC0 or CC-BY; see https://alchemtest.readthedocs.io/en/latest/ and https://github.com/alchemistry/alchemtest/pull/7 (approved by David and Michael)

  4. decided to use logging (with Python logging library) throughout the library https://github.com/alchemistry/alchemlyb/issues/34 (allows calling code more control over what is output and makes it easy to capture output in a file)

  5. decided to implement strict merging from PRs into master, i.e., everyone has to submit a PR in order to get code into the library. The code MUST be reviewed by a developer with write access and it MUST pass status checks: tests and testing coverage. See https://github.com/alchemistry/alchemlyb/wiki/Developer-Guide#pull-request-review-merge-model

Apologies if I missed anything, please add.

Doing the Amber parser was a good way to test and consolidate the development process and make some general decisions (licensing, strict merging). If you want to see the process in action, look at PR #32 https://github.com/alchemistry/alchemlyb/pull/32.

I think we are moving steadily forward with Amber, just needs a few more tests to get merged. We are relying on Michael for the Gromacs extended ensemble (https://github.com/alchemistry/alchemlyb/issues/14) but IIRC Michael said he would be super-busy until middle of November.

If we talk again in a month then it would be good to have the Amber parser fully functional, and initial steps taken towards the Gromacs XE parser, i.e., (1) tests in alchemtest and (2) an initial PR on alchemlyb that can be productively discussed.

@mrshirts

The Shirts group (through graduate student Travis) will be contributing the GROMACS expanded ensemble and replica exchange tests, using the formats for the gromacs test committed so far. The timeline for this is somewhere in the next week, and @mrshirts will help make sure things are going OK. The replica exchange test are expected to pass just fine on first pass, and then the others might not work initially – @mrshirts can help discuss what needs to be done once those are in.

They also have a set of NAMD files that they are trying to get working with alchemical-analysis.py – that might be something to look at after the GROMACS files are in. There's a possible beta tester who will be willing to beta test out NAMD functionality. I think that can wait a little bit.

@davidlmobley

  • Nice job, Shuai — and nice to see you learning along the way. :)
  • I just want to also point out that the link Oliver gives here also is part of a very nice (and brief) Developer Guide that we should all read. :) Thanks for doing this, Oliver. Travis, if you are new to this approach, this is a good place to start getting an idea what’s involved. (BTW, this also means we now have two “Travis”’s involved in the process — Travis-CI, which runs the continuous integration testing, and Travis J., who is a human. :) )
  • Travis, make sure to note that if you need additional data to test on, that goes into alchemtest, and then you write tests/make code changes in alchemlyb (partly because not everyone who has alchemlyb will need the test data, which may be large).
  • NL in my group has a custom modified version of alchemical-analysis (or a separate spinoff) which handles NAMD analysis, and he and VL in my group have both done some free energy calculations with NAMD (VL on an ongoing basis). So it’s likely we can help some with this. I’ll reach out to them. NL might be interested in contributing a little on the coding front.
Clone this wiki locally