Indicating model points on posterior corner plot

[gabrielastro]

Hello Tomas,

On the corner plot of plot_posterior(), it might be practical to show with points the parameter values of the models that were available during the fit to help evaluate easily, visually, whether the errorbars are credible or whether they are suspiciously small around a model point. (The first-order thing is to plot points for pairs of parameters of the grid, and the second-order thing would be to add lines for pairs e.g. involving mass, but the "first-order" option is probably enough.) Especially for the non-Teff parameters, one does not always know by heart the spacing/values (e.g. for metallicity), and the automatically-determined, often-not-quite-usual grid spacing (e.g. Delta T = 30 K) makes it a tiny bit more difficult.

By the way, is there a way of finding out what the parameter values are for parameters other than Teff, apart from reading the models in again to read the information off screen or, as I usually do, doing ls … in the data/ directory? The information is not in species/data/model_data/model_data.json and with database.list_content() I did not find it.

Thanks!

Gabriel

[tomasstolker]

For you second question, you can use get_points()

[gabrielastro]

Oh! great: In a directory in which some models are already stored in a database, doing e.g.:

from species.read.read_model import ReadModel
read_model = ReadModel(model='drift-phoenix')
read_model.get_points()

will show the models stored in the database there (not all models available!), e.g.:

{'teff': array([1200., 1300., 1400., 1500.]),
 'logg': array([3. , 3.5, 4. , 4.5, 5. , 5.5]),
 'feh': array([-0.6, -0.3, -0. ,  0.3])}

Maybe highlight this little function in a tutorial 😉? Thanks in any case.