search

choderalab / fahmunge

Tools for Munging Folding@Home datasets
MIT License
4 stars 6 forks source link

Migrate away from `omnia`? #55

Open sukritsingh opened 1 year ago

sukritsingh commented 1 year ago

Given the migration towards conda-forge for other OpenMM/lab repositories, and the perennial usefulness of this repo for post-processing Folding@home trajectories, would it be worth migrating this at some point to conda-forge if that was possible? Or at least away from omnia?

This is more of a question than anything to be clear. Seems to work fine with omnia based installation for the moment.

@mikemhenry do you have thoughts on what the best direction here would be (or are we fine as we are)? Just want to make sure this repo doesn't get "left behind" something

mikemhenry commented 1 year ago

Yes we should, is there anything you need in omnia that isn't on conda-forge?

sukritsingh commented 1 year ago

I took a look at the list of packages on Omnia (https://anaconda.org/omnia/repo?sort=time.modified&sort_order=desc), but I can't really find a list of the fahmunge dependencies (although they have to be somewhere right?).

Either way, it looks like any package listed on Omnia that we may use (particularly like MDTraj) has been migrated toconda-forge already, so at the moment I don't think I need anything in omnia that isn't already in conda-forge

mikemhenry commented 1 year ago

So we should be good to migrate, do we just need to get fahmunge on conda-forge?

sukritsingh commented 1 year ago

I think so! I guess we could keep it on Omnia for now but I feel like omnia is deprecated. fahmunge is still a go-to way to post-process OpenMM MD datasets run on FAH, so any other labs being able to conda install it would help for other folks to adopt it (and identify other issues/improvements in our post-processing pipeline)

That said - it'll probably be good to run a quick test after (or implement any kind of testing, lol). Happy to help with the testing side of things!