Optuna Integration

[GoldenCorgi]

Hey @danthe3rd !

This should work for the integration mentioned in #145

Two key things I'm not too sure about (sorry again! pretty bad at open source contribution)

1.  ```def from_optuna(study: "optuna.study.Study") -> "Experiment":  ```
   

Is the type checking / type hinting here correct?
2. I added optuna as a dependency for dev & import it during the testing, is this correct?

If everything works out - I'm thinking of implementing this to the other popular hyperparam tuning libraries as well - like HyperOpt & Ray Tune

Thanks!

[danthe3rd]

Hey @GoldenCorgi - nice to see you contributing again :)

Is the type checking / type hinting here correct?

It looks good to me, let's see what the CI say about it

2. I added optuna as a dependency for dev & import it during the testing, is this correct?

Yes perfect. Indeed we don't want to add a dependency for end-users, but it's fine to add dev dependencies

I'll merge it when the CI is all green.

[danthe3rd]

Missing some typing here - let's not worry too much about tests typing tho, this should work

Suggested change

	def objective(trial):
	x = trial.suggest_float("x", -1, 1)
	return x ** 2
	def objective(trial: tp.Any) -> float:
	x = trial.suggest_float("x", -1, 1)
	return x ** 2

[GoldenCorgi]

Updated this in the latest commit!

Missing some typing here - let's not worry too much about tests typing tho, this should work

hiplot/hiplot/test_experiment.py

Lines 43 to 52 in 9501dbb

	def test_from_optuna() -> None:
	def objective(trial: "optuna.trial.Trial") -> float:
	x = trial.suggest_float("x", -1, 1)
	return x ** 2
	study = optuna.create_study()
	study.optimize(objective, n_trials=3)
	# Create a dataframe from the study.