Extra attributes in output GYRE files.

[ehsan]

Hi.
I think to facilitate output file exchange between different users or for archiving reasons, it will be very nice if the output frequency unit would also fly out with the HDF5 files as an extra attribute.
For almost the same portability reasons, Teff, log_g, Xc, Yc etc can also be appended to the HDF5/TXT files as well.
Let us know Rich what you think about this.

Cheers.
Ehsan.

[rhtownsend]

Hi.
I think to facilitate output file exchange between different users or for archiving reasons, it will be very nice if the output frequency unit would also fly out with the HDF5 files as an extra attribute.

That's an easy thing to add, and is now in the development version (which you can grab now, of course; or, wait for the next release).

For almost the same portability reasons, Teff, log_g, Xc, Yc etc can also be appended to the HDF5/TXT files as well.
Let us know Rich what you think about this.

That's something I could in principle do, but I'm concerned we'll end up with files that recreate the structure files in their entirety. As a compromise, what if the output files contained the name of the stellar model file?

[ehsan]

Thanks a lot Rich for the support with the freq_unit.

That's something I could in principle do, but I'm concerned we'll end up with files that recreate the structure files in their entirety. As a compromise, what if the output files contained the name of the stellar model file?

Maybe I did not explain well. Teff, e.g. is just a scalar which is unique for each input model, and can be read in from the input file at photosphere (say tau = 2/3).
My suggestion is to add Teff, log_g, etc as scalar attributes to the output HDF5 files, very similar to the current 'M_star' attribute which is retrievable from short-summary files.
The optimal idea is that each GYRE output file would have a minimalistic attributes to facilitate portability.
Then, we can exchange *.h5 files and read the Teff, log_g, X_c, Y_c, etc as attributes and locate the model on the HR diagram.
Simply, it is repeating what you have implemented for freq_unit, but this time for some critical physical quantities.
I never mean that the full profile of e.g. log_g be written to the short-summary file.

I hope I have managed to convey the idea well.

Best regards.
Ehsan.