GPyOpt is not working with numpy 1.24.0 or higher

[tnakazato]

GPyOpt doesn't work with the latest version of numpy (1.24.0 or higher) due to the "expred deprecations" of numpy 1.24.0, namely gh-22607, which removes some deprecated types including numpy.bool.

https://numpy.org/devdocs/release/1.24.0-notes.html#expired-deprecations

Here is an example traceback.

I would appreciate it if you could fix the problem.

File ~/pyenv/priism-py3.8/lib/python3.8/site-packages/priism-0.11.2-py3.8.egg/priism/core/imager.py:594, in SparseModelingImager._cv_bayesian(self, l1_list, ltsv_list, hp_scale, num_fold, imageprefix, maxiter, eps, clean_box, nonnegative, resultasinitialimage, bayesopt_maxiter)
    588 bounds = [
    589     {'name': 'var_1', 'type': 'discrete', 'domain': l1_list},
    590     {'name': 'var_2', 'type': 'discrete', 'domain': ltsv_list},
    591 ]
    593 problem = GPyOpt.methods.BayesianOptimization(__objective_function, bounds)
--> 594 problem.run_optimization(bayesopt_maxiter)
    595 # for debugging
    596 # problem.save_evaluations('cv_eval.txt')
    597 # problem.save_models('cv_model.txt')
    598 # problem.save_report('cv_report.txt')
    600 return self.CrossValidationResult(
    601     mse=result_mse, image=result_image,
    602     L1=result_L1, Ltsv=result_Ltsv
    603 )

File ~/pyenv/priism-py3.8/lib/python3.8/site-packages/GPyOpt/core/bo.py:137, in BO.run_optimization(self, max_iter, max_time, eps, context, verbosity, save_models_parameters, report_file, evaluations_file, models_file)
    134 while (self.max_time > self.cum_time):
    135     # --- Update model
    136     try:
--> 137         self._update_model(self.normalization_type)
    138     except np.linalg.linalg.LinAlgError:
    139         break

...

File ~/pyenv/priism-py3.8/lib/python3.8/site-packages/paramz/model.py:171, in Model.optimize_restarts(self, num_restarts, robust, verbose, parallel, num_processes, **kwargs)
    169 if not parallel:
    170     if i > 0:
--> 171         self.randomize()
    172     self.optimize(**kwargs)
    173 else:#pragma: no cover

File ~/pyenv/priism-py3.8/lib/python3.8/site-packages/GPy/core/__init__.py:80, in randomize(self, rand_gen, *args, **kwargs)
     78 x = self.param_array.copy()
     79 [np.put(x, ind, p.rvs(ind.size)) for p, ind in self.priors.items() if not p is None]
---> 80 unfixlist = np.ones((self.size,),dtype=np.bool)
     81 from paramz.transformations import __fixed__
     82 unfixlist[self.constraints[__fixed__]] = False

File ~/pyenv/priism-py3.8/lib/python3.8/site-packages/numpy/__init__.py:284, in __getattr__(attr)
    281     from .testing import Tester
    282     return Tester
--> 284 raise AttributeError("module {!r} has no attribute "
    285                      "{!r}".format(__name__, attr))

AttributeError: module 'numpy' has no attribute 'bool'

[apaleyes]

Hi! I am afraid GPyOpt is not supported any more, so your options are either to downgrade numpy or fork and fix.

[tnakazato]

Thanks @apaleyes. I didn't notice it since I have installed GPyOpt through pip. Sorry to hear that. I was just started using it recently...