Julianus Pfeuffer
Julianus Pfeuffer
Yes if you don't care about if an ID or pam was associated to a feature, I would count it as well. Is it because you use best:spectrum_per_peptide in idfilter?...
I think setup.py needs to require numpy 2 as well then.
Ideally setup py would require the same major version it was built with. Setup.py can and currently already is easily configured with cmake.
Ah great, I didn't know! Then definitely let's do >=1.23.5!
I think he meant to install numpy 2 (or any latest) for the build but require numpy>=1.23.5 in the setup.py
@ypriverol What is the status here? Do we still not support tmt18? I would be surprised and I think we should urgently add this if not yet available.
Yes I think it is by far superseded by quantms now. It could be kept for simplicity of lfq focussed analyses but being behind in tool versions and likely not...
Where do you see this? the last run of this was 3 weeks ago and the CDash logs are expired.
I think the second one we are not doing yet but was proposed as a solution to indicate version compatibility
I think the cleanest solution is to trigger CMake with setuptools. See my efforts in better_pyopenms_build. You could quickly try if you can add `--no-binary pyopenms` to skip building binaries...