On Wed, Dec 04, 2024, Osadchey Brown, Reilly via AMBER wrote:
>
>I experienced this issue as well in both pdb4amber and parmed. I was able
>to fix this error and get parmed to run by altering all the specific
>commands related to the np.array lines (for parmed case, 1 line in each .in
>./parmed/structure.py & ./parmed/amber/_amberparm.py). An example is shown
>here:
>
>coords = np.array(value, dtype=np.float64, copy=False, subok=True)
>becomes
>coords = np.asarray(value, dtype=np.float64)
>
>I installed Amber and Ambertools 24 with Python 3.12 and NumPy 2.1.1 for
>reference.
Thanks for the suggested patches. My recollection(!?) is that there are a
number of places in AmberTools where numpy 2 will fail. A better solution
(for most folks) is probably to pin numpy to version 1.26.4. But maybe the
changes are less extensive that I had thought. I'll try to get someone to
look into this.
If you ask CMake to download miniconda, it will create a python inside the
installation tree, which will be only used for Amber. In that way, other
python apps that you have (using NumPy 2.1.1) won't be affected.
...thx...dac
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Wed Dec 04 2024 - 09:03:01 PST
