Add mechanism to manually do model selection

[MilesCranmer]

By passing, for example, model.predict(X, index=2), you will perform a prediction using the 2nd equation in the internal model.equations dataframe. You can view the index of each expression by running print(model), or simply looking at model.equations.

This allows the user to specify their own model selection strategy, and is quite a simple change to the codebase. It also works for the sympy, latex, pytorch, and jax export functions–simply by changing the argument of .get_best to have a row kwarg.