-
Notifications
You must be signed in to change notification settings - Fork 20
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Running ./scripts/setup_dev_env.sh
fails at MDAnalysis build
#81
Comments
Installing Cython 0.29.36 into a fresh conda-based env before running the script again produced a different error, so if downgrading the version is a possible workaround, that version isn't old enough:
|
Regarding the pin, I've read #75 (comment) and do agree, but you should also know that since many projects bring in Cython as a transitive dependency those projects don't necessarily pin Cython versions, their upgrade from from 0.29 to 3.0 broke a lot of code. |
Thanks @berquist for this report and all of the details. @JFRudzinski: I remember that you were trying out updating the MDAnalysis version: Did you update the mda version in the issue you are working on: #1643? In principle, I don't see any problems updating our pinned version to e.g. 2.5.0. |
@lauri-codes Actually, I had updated to 2.5.0 some time ago (not sure which issue or branch). However, in my tests for #1643 I found that 2.5.0 has some issues reading certain versions of .trr and .xtc files. So, I was planning to revert now back to 2.1.0. However, once I am done with my dev, I can do some more testing if it would help to see what the root of the issues are with 2.5.0 in my case? |
@JFRudzinski: Yes, it would be great if you could debug the issue in reading .trr and .xtc files. If the problem is with MDA itself, then you should search for any existing issues in their GitHub page, and possibly file a new one if nothing exists already. It is a bit unfortunate that MDA has a cython build step: the official recommendation is to ship the resulting .c files in a release and use them primarily even if the user has cython installed. But looking at their setup.py, they seem to prefer the cython compilation if cython exists in the installation environment. |
Ok, np, I will look into it and get back to you here |
@lauri-codes I am fairly confident now that the only issue with MDA version 2.5.0 is something to do with a sort of advanced usage that I am performing when calculating the msds. Additionally, this is only happening for 1 trajectory file obtained from a user. (As I said before, I had been using 2.5.0 for a while without any issues on many other data). I would say then that you can pretty safely go ahead and update the MDA version. I will address my own specific issue in the mean time, but I don't really think you need to wait for this, as it is so specific (and possibly unique to this single example). |
The MDA version has now been updated to v2.5.0. |
This is part of the JOSS review.
Following https://nomad-lab.eu/prod/v1/docs/develop/setup.html#install-nomad, running the script after performing
gives the following
pip
traceback, failing to compile MDAnalysis 2.1.0:I haven't confirmed, but I wouldn't be surprised if it's one of the many issues related to the new Cython 3.0.0 release, since it's probably picking up my system install, and https://github.com/MDAnalysis/mdanalysis/blob/2bf55dd714292bbf7c12023a28153a6443a52ec6/package/CHANGELOG#L86 shows this may not have been fixed until MDAnalysis 2.5.0.
Two questions:
The text was updated successfully, but these errors were encountered: