You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Just a note regarding the release of pymbar 4.0, which has breaking API changes, as documented here. We'll likely need to update the analyse_freenrg script if we want to use the new API, or take advantage of the new features. (I know that BioSImSpace will be moving to alchemlyb, with pymbar as a backup in future.) For now I'll pin pymbar to version 3 in our conda recipe. (pymbar is actually missing at present following the 2023.0.0 merge, so I'll re-add.)
The text was updated successfully, but these errors were encountered:
Just to add on to this, pymbar is a dependency for alchemlyb, as alchemlyb does end up feeding the extracted data into pymbar during the MBAR analysis. This is also currently not able to use pymbar 4.0.
Thanks for this. I had also noticed this the other day when I checked their recipe having discovered that version 4 pulls in cudnn and jaxlib, which come to over a GB in size. If/when we do support version 4, I really hope that we can avoid the overhead of these packages if we don't need their functionality.
Just a note regarding the release of
pymbar 4.0
, which has breaking API changes, as documented here. We'll likely need to update theanalyse_freenrg
script if we want to use the new API, or take advantage of the new features. (I know that BioSImSpace will be moving toalchemlyb
, withpymbar
as a backup in future.) For now I'll pinpymbar
to version 3 in our conda recipe. (pymbar
is actually missing at present following the 2023.0.0 merge, so I'll re-add.)The text was updated successfully, but these errors were encountered: