Import error on Google Colab

[bcallen95]

Hi there! I've been trying to run the "Alchemical Free Energy Calculations" Tutorial on the OpenMM Page. When I try to import openmmtools I get the following traceback:

---

AttributeError Traceback (most recent call last)
in ()
----> 1 import openmmtools

2 frames
/usr/local/envs/openmm/lib/python3.7/site-packages/openmmtools/init.py in ()
6 """
7
----> 8 from openmmtools import testsystems, integrators, alchemy, mcmc, states, cache, utils, constants, forces, forcefactories, storage, multistate
9
10 # Handle versioneer

/usr/local/envs/openmm/lib/python3.7/site-packages/openmmtools/multistate/init.py in ()
63
64 from openmmtools.multistate.multistatesampler import MultiStateSampler
---> 65 from openmmtools.multistate.multistatereporter import MultiStateReporter
66 from openmmtools.multistate.replicaexchange import ReplicaExchangeSampler, ReplicaExchangeAnalyzer
67 from openmmtools.multistate.paralleltempering import ParallelTemperingSampler, ParallelTemperingAnalyzer

/usr/local/envs/openmm/lib/python3.7/site-packages/openmmtools/multistate/multistatereporter.py in ()
1752 # ==============================================================================
1753
-> 1754 class _DictYamlLoader(yaml.CLoader):
1755 """PyYAML Loader that reads !Quantity tags."""
1756 def init(self, *args, **kwargs):

AttributeError: module 'yaml' has no attribute 'CLoader'

Do you have any thoughts/advice on how to go about fixing this issue? Thanks!

[bcallen95]

It appears that my environment set-up was the issue. Alternatively setting up the Conda environment (in Google Colab) in the following way appears to fix all issues (with each separate "code block" representing a separate cell):

!pip install -q condacolab
import condacolab
condacolab.install()

import condacolab
condacolab.check()

!mamba install -q openmmtools

However, this does lead to a new issue in the "Analyzing the Data with MBAR" Section... Running the following code:

from pymbar import MBAR, timeseries
N_k = np.zeros([nstates], np.int32) # number of uncorrelated samples
for k in range(nstates):
      [nequil, g, Neff_max] = timeseries.detectEquilibration(u_kln[k,k,:])
      indices = timeseries.subsampleCorrelatedData(u_kln[k,k,:], g=g)
      N_k[k] = len(indices)
      u_kln[k,:,0:N_k[k]] = u_kln[k,:,indices].T
mbar = MBAR(u_kln, N_k)
[DeltaF_ij, dDeltaF_ij, Theta_ij] = mbar.getFreeEnergyDifferences()

Results in this traceback:

---

ParameterError Traceback (most recent call last)
in ()
10 # Compute the Free Energy Differences and Statistical Uncertainties
11 mbar = MBAR(u_kln, N_k)
---> 12 [DeltaF_ij, dDeltaF_ij, Theta_ij] = mbar.getFreeEnergyDifferences()

2 frames
/usr/local/lib/python3.7/site-packages/pymbar/utils.py in check_w_normalized(W, N_k, tolerance)
367 raise ParameterError(
368 'Warning: Should have \sum_k N_k W_nk = 1. Actual row sum for sample %d was %f. %d other rows have similar problems' %
--> 369 (firstbad, row_sums[firstbad], np.sum(badrows)))
370
371 return

ParameterError: Warning: Should have \sum_k N_k W_nk = 1. Actual row sum for sample 6 was 0.999696. 7 other rows have similar problems

[mikemhenry]

@bcallen95 Sorry I just saw this, that error you are getting now is from pymbar, see this issue: choderalab/pymbar#419