1.10 release planning about mdtraj HOT 10 CLOSED

I've downloaded this release candidate and quickly am running it on my analysis/MSM scripts and so far no crashes! Thanks for this heroic effort. I'll take a look at some of the other PRs and see if I can help merge/review!

(Also apologies for the big hiatus! I got award a grant and have only just now had life return to some degree of normal pace lol.)

from mdtraj.

Yep, that's my understanding as well

from mdtraj.

Depending on how much work is a part of making MDTraj compatible with NumPy 2.0, we should either pin to 1.x or add 2.0rc1 testing ASAP

#1873

from mdtraj.

Should this release add compatibility for MacOSX + python 3.12? From reading around the commits on the feedstock, It looks like the issue was related to the .tng code, which I think was removed in the latest PR.

from mdtraj.

I tagged a 1.10.0rc1 release which should fix a handful of issues that have popped up over the past year, though not all PRs are merged through.

Try it with

$ pip install mdtraj==1.10.0rc1

or

$ micromamba install mdtdraj -c conda-forge/label/mdtraj_rc -c conda-forge

and let us know if anything is surprising. If no feedback this will become a full release in a few weeks

from mdtraj.

Just wanted to add that the numpy<2 pin is not yet included in the rc, so it might be worth adding into 1.10 before release (at least until mdtraj supports numpy 2).

#1873

from mdtraj.

Good idea - do I do that here? https://github.com/mdtraj/mdtraj/blob/1.10.0rc1/pyproject.toml#L5-L6

from mdtraj.

I'm like 80% sure that {{ pin_compatible('numpy') }} handles this in conda world but I'd be fine making that pin more explicit in feedstock builds

from mdtraj.

I made https://github.com/mdtraj/mdtraj/releases/tag/1.10.0rc2

from mdtraj.

#1890

Wheels (#1833) would be nice, too, if we can figure out how to set that up

from mdtraj.