The Asfgrid of dnu corrections has been updated (https://iopscience.iop.org/article/10.3847/2515-5172/ac8b12) with more points and an extended range. It can be downloaded from here: http://www.physics.usyd.edu.au/k2gap/Asfgrid/ . This could be quite an upgrade for the BaSTI library to complement the Serenelli corrections.

Update the code to use the new version -- REMEMBER that we patched the old version to work in BASTA. And that we added multiprocessing and therefore included some locks in the input to some routines.

Note: This is actually more of an issue for the building routines! I have made this issue just to make sure that we check to compatibility when the grids are updated.

When during frequency fitting I discovered that it was required of me to specify threepoint=False in the freqparams xml tag. That makes no sense as (i) that is for ratio fitting and (ii) it should be False by default.

I then discovered that _find_get could not handle the default value of threeport=False. This is a rewrite of that function.

While I was at it, I added defaults for fcor (cubicBG14), correlations (False), and dnufrac (0.15). When fitting frequencies, you can simply just specify

<freqparams>
    <freqpath value={path-to-frequencies}
</freqparams>

and add <freqs/> to fitparams and maybe set

<freqplots>
    <True/>
</freqplots>.

Implementation of epsilon-differences fitting (cf. our Plato-module). Major refactoring of the surface-independent fitting (i.e. ratios), especially the routines to do I/O and prepare the observations. Refactoring of plotting code (especially ratios) and added new plots.

New plots:

• Epsilon-differences diagnostic plots
• Correlation map {freqs, ratios, epsdiff}
• Plain echelle diagram
• Single-panel ratio plots

Other changes:

• Ratios are interpolated to the observed frequencies

This just refactors the code behind the echelle diagrams, which makes it more readable and less relient on the number of unique l values of the observations.

It also makes small aesthetic changes such as changing $l$ to $\ell$ and adding the Frequency normalised by dnu axis to all plots and not just the duplicateechelle plots.

Update packages to newest versions (to make it installable on MacOS). Fix bug in json example. Fix bugs and add default values in frequency fitting (PR #17). Update plotting strings to match newest version of the Sharma/Stello (asfgrid) corrections.

Two changes:

• Bump requirements to newest versions and add bottleneck.
• Switch to shipping the 16 Cyg A grid with new weights (based on Sobol volume). Reference examples updated to match; interpolation example still uses the old grid.

Add the option to use 3-point seperation ratios instead of the default 5-point. This is useful for reproducing older results or results from other groups.

Additionally, the default of dnufit_in_ratios is changed to False (as it ought to have been always).

Replaced the coefficients from DR2 with eDR3. The new coefficients follow the prescription of Casagrande and Vandenberg (2018). They are now the default Gaia coefficients

Bring main to version 1.1.3.

Calculation of bayesian weights when interpolating along a track is fixed -- before it used the base parameter, now it correctly uses seperation in age (or mass for isochrones). Details in #10 .

The Bayesian weights along the tracks/isochrones "dage"/"dmass" were accidentially computed from the chosen base parameter in the interpolation, instead of the age/mass as they should. This update fixes that, by separately computing the "dage"/"dmass" for each track/isochrone after interpolation of all other quantities. This should also ensure they are always computed.

Bring main to version 1.1.2.

Smoother KDE representations in the corner plots. Updated examples and documentation. Add published paper details.

Full details in #7 .