Unable to import BioSimSpace after mamba installation
akalpokas opened this issue · 12 comments
Hello,
After installing biosimspace using mamba I am unable to import it in python.
I have installed it using:
mamba create -n biosimspace23 -c conda-forge -c michellab biosimspaceI tried importing it using:
import BioSimSpace as BSSWhich gives:
---------------------------------------------------------------------------
AttributeError Traceback (most recent call last)
Cell In[1], line 1
----> 1 import BioSimSpace as BSS
File ~/miniconda3/envs/biosimspace23/lib/python3.9/site-packages/BioSimSpace/__init__.py:209
206 del _shlex
207 del _subprocess
--> 209 from . import Align
210 from . import Box
211 from . import FreeEnergy
File ~/miniconda3/envs/biosimspace23/lib/python3.9/site-packages/BioSimSpace/Align/__init__.py:43
1 ######################################################################
2 # BioSimSpace: Making biomolecular simulation a breeze!
3 #
(...)
23 # https://github.com/MobleyLab/Lomap, which is released under the MIT
24 # license.
26 """
27 .. currentmodule:: BioSimSpace.Align
28
(...)
40 viewMapping
41 """
---> 43 from ._align import *
File ~/miniconda3/envs/biosimspace23/lib/python3.9/site-packages/BioSimSpace/Align/_align.py:79
77 from .._Exceptions import AlignmentError as _AlignmentError
78 from .._Exceptions import MissingSoftwareError as _MissingSoftwareError
---> 79 from .._SireWrappers import Molecule as _Molecule
81 from .. import IO as _IO
82 from .. import Units as _Units
File ~/miniconda3/envs/biosimspace23/lib/python3.9/site-packages/BioSimSpace/_SireWrappers/__init__.py:40
1 ######################################################################
2 # BioSimSpace: Making biomolecular simulation a breeze!
3 #
(...)
19 # along with BioSimSpace. If not, see <http://www.gnu.org/licenses/>.
20 #####################################################################
22 """
23 .. currentmodule:: BioSimSpace._SireWrappers
24
(...)
37 SearchResult
38 """
---> 40 from ._atom import *
41 from ._bond import *
42 from ._molecule import *
File ~/miniconda3/envs/biosimspace23/lib/python3.9/site-packages/BioSimSpace/_SireWrappers/_atom.py:34
30 __all__ = ["Atom"]
32 from sire.legacy import Mol as _SireMol
---> 34 from ..Types import Coordinate as _Coordinate
35 from ..Types import Length as _Length
36 from ._sire_wrapper import SireWrapper as _SireWrapper
File ~/miniconda3/envs/biosimspace23/lib/python3.9/site-packages/BioSimSpace/Types/__init__.py:44
1 ######################################################################
2 # BioSimSpace: Making biomolecular simulation a breeze!
3 #
(...)
19 # along with BioSimSpace. If not, see <http://www.gnu.org/licenses/>.
20 #####################################################################
22 """
23 .. currentmodule:: BioSimSpace.Types
24
(...)
41 Volume
42 """
---> 44 from ._angle import *
45 from ._area import *
46 from ._charge import *
File ~/miniconda3/envs/biosimspace23/lib/python3.9/site-packages/BioSimSpace/Types/_angle.py:35
31 from sire import units as _SireUnits
33 from ._type import Type as _Type
---> 35 class Angle(_Type):
36 """An angle type."""
38 # A list of the supported Sire unit names.
File ~/miniconda3/envs/biosimspace23/lib/python3.9/site-packages/BioSimSpace/Types/_angle.py:42, in Angle()
39 _sire_units = ["radian", "degree"]
41 # Dictionary of allowed units.
---> 42 _supported_units = { "RADIAN" : _SireUnits.radian,
43 "DEGREE" : _SireUnits.degree }
45 # Map unit abbreviations to the full name.
46 _abbreviations = { "R" : "RADIAN",
47 "D" : "DEGREE" }
AttributeError: module 'sire.units' has no attribute 'radian'BioSimSpace version: 2023.0.0
Sire version: 2023.0.3 (011b6f6edb831eeb5447b6d4e2a015123a99b2a2/GitHub/2023-01-10T10:14:37+00:00 | UNCLEAN)
I have tried this by installing the development version too using:
mamba create -n biosimspace23-dev -c conda-forge -c michellab/label/dev biosimspaceWhich gives the same error when importing as well. If I force mamba to install BioSimSpace version 2022.2.1 then it works.
I hope this helps, thanks.
Yes, it looks like we've incorrectly specified the Sire version range for our 2023 pre-release version of BioSimSpace. You should also be able to fix things by specifying the Sire version. For example, I think the 2023.0.2 development version of Sire should work with the latest devel conda package.
mamba create -n biosimspace23-dev -c conda-forge -c michellab/label/dev sire=2023.0.2 biosimspace
Yup, that works. Thanks for your help!
No problem, thanks for reporting. Once the full 2023.1.0 release is available I'll delete the pre-release packages. This will avoid other people running into the same issue in future.
Hej, fyi: I have the same problem with your conda-forge version :(
Thanks, we don't have a package on conda-forge. Does the solution posted above not work for you?
ah! my bad! but yes the the fix works (I was installing it like this: conda install -c conda-forge biosimspace - that's where my confusion came from)
Hi, I have installed biosimspace on this way. However when I am analyzing the relative free-energy value it gives an error.
AnalysisError Traceback (most recent call last)
Cell In[12], line 1
----> 1 pmf_free, overlap_matrix_free = BSS.FreeEnergy.Relative.analyse(f'o_xylene_benzene/free')
2 pmf_bound, overlap_matrix_bound = BSS.FreeEnergy.Relative.analyse(f'o_xylene_benzene/bound')
3 freenrg_rel = BSS.FreeEnergy.Relative.difference(pmf_bound, pmf_free)
File /cluster/ddu/cmmartinez001/miniconda3/envs/biosimspace-dev/lib/python3.9/site-packages/BioSimSpace/FreeEnergy/_relative.py:431, in Relative.analyse(work_dir)
429 # SOMD.
430 if len(data) > 0:
--> 431 return Relative._analyse_somd(work_dir)
433 # Now check for GROMACS output.
434 else:
435 data = glob(work_dir + "/lambda*/gromacs.xvg")
File /cluster/ddu/cmmartinez001/miniconda3/envs/biosimspace-dev/lib/python3.9/site-packages/BioSimSpace/FreeEnergy/_relative.py:605, in Relative._analyse_somd(work_dir)
598 proc = _subprocess.run(
599 _Utils.command_split(command),
600 shell=False,
601 stdout=_subprocess.PIPE,
602 stderr=_subprocess.PIPE,
603 )
604 if proc.returncode != 0:
--> 605 raise _AnalysisError("SOMD free-energy analysis failed!")
607 # Re-run without subsampling if the subsampling has resulted in less than 50 samples.
608 with open("%s/mbar.txt" % work_dir) as file:
AnalysisError: SOMD free-energy analysis failed!
In a previosu version it worked without any problem. Not sure if the code has been updated. Thanks
Cesar
Thanks, I'll try to reproduce this later and get back to you.
Hmmm, it's working for me with some existing ethane/methanol solvation free energy data. Is it possible to share the data for which it's failing. You could also edit the file shown in the error message above to print the standard output and error, e.g.:
print(proc.stdout)
print(proc.stderr)(Do the above after adding text=True to the subprocess.run parameters.)
For reference, I did find an issue with our GROMACS analysis which resulted from the new (well, not so new) way we split command-line arguments for subprocess. This means that we now need to expand the glob for xvg files prior to running gmx bar. This is now fixed. For the few cases I've tested, SOMD analysis still appears to work just fine.
Also, when we figure out what the actual problem is I'll move it to a new issue thread, since it appears to have nothing to do with the original question. In future, please open a new issue when the query us unrelated. It makes book-keeping much easier.
Cheers.