v0.8.0

Highlights

• Added lbsubs and an example of its usage (more accurate current densities and DMerc terms)
• $B_s$ on full instead of half grid (deviates from Fortran VMEC 8.52 but is more in line with how most codes use $B_s$ in practice)
• Separate resolution parameters for geometry, solve, and free boundary components
• Preliminary version of JAX differentiable wout quantities #861
• Bugfix for free boundary equilibria with p(1) edge pressure != 0 intentionally deviating from Fortran VMEC #830
• Flow control improvements (robustness, speed) #884 #828 #849 #933 #822

Version bump due to default changes in flow control and bsubs changing from half to full grid

What's Changed

• Carry signgs in the geometry of VmecModel by @CharlesCNorton in #881
• Size the parallel regions with num_threads clauses instead of the process thread count by @CharlesCNorton in #835
• Default tcon0 to 0.5 by @CharlesCNorton in #822
• Test lasym invariance under toroidal rotation and poloidal relabeling by @jurasic-pf in #879
• Fix free boundary for non zero boundary pressure Take the vacuum magnetic pressure alone as the total pressure outside the free boundary by @CharlesCNorton in #830
• Extrapolate the wout current-density ends from the interior alone by @CharlesCNorton in #865
• JAX port of the wout output post processing by @matildevalente-pf in #861
• Size the JAX output stage from the geometry it receives by @jurasic-pf in #900
• Add public VmecInput.resize() method by @jurasic-pf in #901
• Differentiate the adjoint with respect to the pressure, iota and current profiles by @jurasic-pf in #906
• Let the vacuum potential's Fourier cutoffs exceed the plasma's by @CharlesCNorton in #833
• Accept the mgrid mode N that simsopt writes and LIBSTELL assigns to mgrid files without a mode by @CharlesCNorton in #902
• Retry a bad initial Jacobian from a three-surface mesh by @CharlesCNorton in #884
• Implement the sum_cossq current profiles by @CharlesCNorton in #832
• Add an opt-in undamped lambda preconditioner by @CharlesCNorton in #820
• lgiveup by @CharlesCNorton in #845
• Carry the Jacobian sign in the Mercier volume derivative by @CharlesCNorton in #795
• add docling version of TNOV by @jons-pf in #526
• Check VMEC++ against analytic three-dimensional MHD equilibria in CI by @CharlesCNorton in #912
• Pin the m=1 gauge in VmecModel so the equilibrium and its Jacobian are functions of the boundary by @jurasic-pf in #849
• Extrapolate the full-grid ends from the interior alone by @CharlesCNorton in #864
• Hand every force iteration to a callback and watch a solve converge by @CharlesCNorton in #839
• Support simsopt indata conventions by @jurasic-pf in #920
• Fix editable installs when multiple venvs are available by @jurasic-pf in #923
• Back up the last evaluated state instead of the advanced one by @jurasic-pf in #933
• output_quantities: implement lbsubs full-grid B_s via radial force balance by @CharlesCNorton in #537
• Write the full-grid bsubsmns and wout version 9.0 by @jurasic-pf in #945
• examples: W7-X current density with lbsubs = False and True by @jurasic-pf in #948

New Contributors

• @matildevalente-pf made their first contribution in #861
• @omeneghini-pf made their first contribution in #911

Full Changelog: v0.7.5...v0.8.0