GUI: Improve first post-simulation plot

[asoplata]

Disclaimer: If you are new contributor, please do not work on this issue at this time. If you would like to contribute, please find a different issue, and make sure that there is no Pull Request already open for that issue.

When a simulation is run in the GUI and there is an existing loaded experimental dataset, have the autogenerated plot show the simulation, the experimental dataset, and the RMSE between them.

UPDATE: This needs to wait until after we perform a large refactor of the GUI's VizManager class.

[Tusharjamdade]

Hi @asoplata,

Could you please confirm whether this is the expected behavior when running the second simulation? The UI displays two dipole traces in different colors corresponding to the latest and previous simulations, along with an RMSE annotation indicating the difference between the two.

Please let me know if any further improvements or adjustments are required.

[dylansdaniels]

@asoplata are you talking about comparing simulations to each other, or showing the RMSE between loaded and experimental data for the last two simulations?

I'm not sure that we would want RMSE between the last two simulations as a default behavior. That would cause issues for some of our tutorials as well, when we are changing simulations but not necessarily directly comparing to the previous sim. We'd then have to instruct users how to remove the second simulation to get our 'default' graph with only one sim

[ntolley]

Really good point @dylansdaniels! I brought this up because I found that in the case of hand tuning, needing to regenerate the comparison plot every simulation takes a ton of time when you're doing many iterations. So it would be comparing new sim to loaded experimental.

[Tusharjamdade]

Thanks for the clarification @dylansdaniels and @ntolley, that makes sense. I agree sim-to-sim RMSE shouldn’t be the default. I can update the PR to compute RMSE between the latest simulation and the loaded experimental data instead. Let me know if that works.

[asoplata]

Discussed with @ntolley today and we agree with Dylan's original point: this does not make sense for sim-to-sim, and only makes sense for comparing simulations to loaded experimental data. However, because VizManager needs a complete refactor in order to discriminate between loaded data and simulated data, this issue is blocked until that work is done. Hopefully, we will be able to get help from CCV to do this GUI refactor over the next couple months.