[FEAT]: Sensitivity analysis in "Simulate" operator should support other "slice" options besides a specific timepoint

[liunelson]

Is your feature request related to a problem? Please describe.

Currently, we can only select a time point of interest to do a sensitivity analysis.

However, the Eval contained questions that required analysis of not a fixed timepoint for all samples, but the peak timepoint of each sample (i.e. max per sample). A nice-to-have would be to include a "argmax per sample" option too.

Describe the solution you'd like

Akin to the natural-language interface in "Optimize", the timepoint selection could be:

Consider how sensitive the [select "peak value | peak timepoint | value at timepoint"] [if "value at timepoint", show numeric text field for timepoint] of the outcome(s) of interest [multiselect with filter "state variable or observable"] is to the model parameters [multiselect with filter "model parameters"].

Pseudocode

If value at timepoint (currently):

# outcome = pandas.DataFrame(
#     {
#         "sample_id": [0, 0, 1, 1, 2, 2], 
#         "timepoint_value": [0.0, 1.0, 0.0, 1.0, 0.0, 1.0],
#         "value": [...]
#     }
# )
target = outcome[outcome["timepoint_value" == timepoint_of_interest]]["value"]

If peak value:

target = outcome[["timepoint_value", "value"]].groupby("timepoint_value").max()

If peak timepoint:

target = outcome[outcome.groupby("timepoint_value").idxmax()]["value"]