Updated run in pdes.py for using TrialData and FullTrialData

[yb6599]

This changes the pdes.py code to use the TrialData and FullTrialData objects for showing the results.

[yb6599]

Hey Jake, I found the initial conditions for Burgers and Kuramoto Sivashinsky Equations that can give us better results by running

gen_experiments.pdes.run(42,
group= "burgers1D_periodic",
diff_params=gen_experiments.config.diff_params["test_axis"],
opt_params=gen_experiments.config.opt_params["test"],
feat_params=gen_experiments.config.feat_params["pde2"],
)

and

gen_experiments.pdes.run(42,
group= "ks_periodic",
diff_params=gen_experiments.config.diff_params["test_axis"],
opt_params=gen_experiments.config.opt_params["test"],
feat_params=gen_experiments.config.feat_params["pde4"],
)

[Jacob-Stevens-Haas]

Alright, thanks for doing that. Run pre-commit on your code (I've added some) so that you can pass the linting CI. Also, you need to remove the /, in pdes.run() function signature - mitosis doesn't allow experiments to have positional-only arguments.

BTW, the mitosis commands that are equivalent to what you posted:

mitosis gen_experiments.pdes -e seed=42 -F trials/debug -g burgers1D_periodic -p diff_params=test_axis -p opt_params=test -p feat_params=pde2

and

mitosis gen_experiments.pdes -e seed=42 -F trials/debug -g ks_periodic -p diff_params=test_axis -p opt_params=test -p feat_params=pde4

The next step is to recreate the experiments from the examples we looked at. It builds confidence that you're able to recover the correct equations in these cases, but ultimately, we want to be able to compare methods on the actual benchmark problems.