Fixes for MultiDatasetTradingEnv

[sachetz]

In TradingEnv, the options parameter in the reset method is not used, so I kept that consistent for MultiDatasetTradingEnv as well.

For the dataset selection issue, the following selects the dataset that is used the least times at each step.

potential_dataset_pathes = np.where(self.dataset_nb_uses == self.dataset_nb_uses.min())[0]
random_int = np.random.choice(potential_dataset_pathes)
dataset_path = self.dataset_pathes[random_int]

I have tested this out for a personal project.

[sachetz]

These are implemented fixes for Issue #28