Fixing Infer Radius Method #153
Conversation
…e choice provided by the user which in many cases is the chosen evolutionary model.
|
Below is an example, where an evolutionary model was provided but when s.results is run, the inferred radius is not from the evolutionary model but from the a spt relation. |
|
Is the issue describing the problem you are trying to fix? #135 Ideally, you would write a test which demonstrates the "broken" behavior and then your fix would make that test pass. |
|
Maybe you need to add a test_infer_radius function to https://github.com/BDNYC/sedkit/blob/main/tests/test_sed.py |
|
the pseudocode @kelle , would be the following: |
…uld provide the user with a six digit radius when having an sed with substellar units.
…ts and another where we don't. It may be a bit much but I wanted to make sure that we were getting sensible values for both. Currently some tests fail, so I will continue to tweak.
Parametrize results too
…ven though the uncertainties are 0.001, in the results table they show up as zero.
| [ | ||
| ( | ||
| "sub_spec", | ||
| (0.973 * u.Rjup, 0.0 * u.Rjup, 0.0 * u.Rjup), |
There was a problem hiding this comment.
Is this what you really want, uncertainties of 0?
There was a problem hiding this comment.
Hi, yes and no, these zero uncertainties come because the radius in this case scenario is derived from the Lbol_sun in the dwarf sequence.txt and the interpolation the the code does ends up providing these uncertainties. I tried playing around a bit with this code Joe wrote but not super sure why it gives zero as the uncertainty.
There was a problem hiding this comment.
I think it's happening because of the round() statements on line 289 of relations.py. See my other comment.
sedkit/relations.py
Outdated
|
|
||
| if y_upper is not None: | ||
| return y_val, y_upper, y_lower, self.ref | ||
| return y_val.round(3), y_upper.round(3), y_lower.round(3), self.ref |
There was a problem hiding this comment.
I don't think you should round these values. instead, I think you should consider using %g formatting when displaying it to the user.
Check out the examples here: https://stackoverflow.com/a/3411731/4842634
Co-authored-by: Kelle Cruz <kellecruz@gmail.com>
…ferent formating to the evaluate function to have proper sig figs.
…se precision and it still provides 0 as the uncertainty simply because of the interpolation from the sequence. The zero is not from the formatting.
…e evalue the calculation in isochrone.py.
… fundamental parameters. Also, integrate the keyword for the age distribution in the evaluation lines.
…tements to account for how larger age ranges will give extremely similar results but not identical.
Currently, the checking infer_from loop is not properly preferring the choice provided by the user which in many cases is the chosen evolutionary model. I think changing this loop will change infer_from properly to evo_model.