[Feature] Support string output path in `EvalAlgorithmInterface.evaluate(save="...")`

Today the `save` parameter of `EvalAlgorithmInterface.evaluate()` is [just a boolean](https://github.com/aws/fmeval/blob/220fdc090ee4f1db7174f6e7661a282a5987d1af/src/fmeval/eval_algorithms/eval_algorithm.py#L35).

From hunting around I found that the output path is [taken](https://github.com/aws/fmeval/blob/220fdc090ee4f1db7174f6e7661a282a5987d1af/src/fmeval/util.py#L47) either from an environment variable or else a [default under /tmp](https://github.com/aws/fmeval/blob/220fdc090ee4f1db7174f6e7661a282a5987d1af/src/fmeval/constants.py#L4)... And that it should also be overrideable by setting the obviously-supposed-to-be-private property `eval_algo._eval_results_path`.

IMO it's harder and less obvious than it should be for a developer using the library to save results in a folder they want. It'd be much easier if we could support `eval_algo.evaluate(save="my/cool/folder")` and ideally automatically create the provided folder if it doesn't already exist?

Start with EvalAlgorithmInterface.evaluate() in src/fmeval/eval_algorithms/eval_algorithm.py, then trace the output-path handling in src/fmeval/util.py and the default in src/fmeval/constants.py. Confirm that a string passed to save selects the requested folder, creates it when needed, and preserves the existing behavior for boolean values.

Written by the indexing model from the issue text.