example_fabolas broken. Which george version?

I tried to run example_fabolas, but i get this error:

-
AttributeError Traceback (most recent call last)
in
101 # Start Fabolas to optimize the objective function
102 res = fabolas(objective_function, lower=lower, upper=upper,
--> 103 s_min=s_min, s_max=s_max, num_iterations=200)
104
105 x_best = res["x_opt"]

~/anaconda3/lib/python3.7/site-packages/RoBO-0.3.1-py3.7.egg/robo/fmin/fabolas.py in fabolas(objective_function, lower, upper, s_min, s_max, n_init, num_iterations, subsets, inc_estimation, burnin, chain_length, n_hypers, output_path, rng)
112 # Kernel for the environmental variable
113 # We use (1-s)**2 as basis function for the Bayesian linear kernel
--> 114 env_kernel = george.kernels.BayesianLinearRegressionKernel(log_a=0.1, log_b=0.1,
115 ndim=n_dims + 1,
116 axes=n_dims)

AttributeError: module 'george.kernels' has no attribute 'BayesianLinearRegressionKernel'

I am currently running george 0.3.1. Which george version should i use?

Start by running the example_fabolas entry point and inspect the fabolas call where george.kernels.BayesianLinearRegressionKernel raises the AttributeError. Compare the installed george 0.3.1 API with the example's expected API; done means the example runs successfully and the supported george version or required compatibility change is documented.

Written by the indexing model from the issue text.