Migrate from numba to Cython

[Cadair]

fixes #39

I'd like someone to verify the performance delta here, from the speed the test suite runs this seems to be a good chunk faster than numba.

[STBadman]

For a standard non-trivial PFSS extrapolation using the sample_data gong map i get a 50-100% increase in runtime with this cython implementation vs the current numba release unfortunately. I will try to reproduce the test speed up you saw and see if we can understand the difference.

Timing test log for this PR with Cython :

(.sunkit-magex-cython) (base) sbadman@si393685:~/1_RESEARCH/GIT_REPOS/sunkit-magex$ ipython
Python 3.13.14 (main, Jun 11 2026, 12:30:58) [GCC 11.4.0]
Type 'copyright', 'credits' or 'license' for more information
IPython 9.14.1 -- An enhanced Interactive Python. Type '?' for help.
Tip: IPython supports combining unicode identifiers, eg F\vec<tab> will become F⃗, useful for physics equations. Play with \dot \ddot and others.

In [1]: import sunkit_magex; print(sunkit_magex.__version__)
   ...: 
   ...: import sunpy.map
   ...: from sunkit_magex import pfss
   ...: 
   ...: gong_fname = pfss.sample_data.get_gong_map()
   ...: gong_map = sunpy.map.Map(gong_fname)
   ...: 
   ...: nrho = 35
   ...: rss = 2.5
   ...: pfss_in = pfss.Input(gong_map, nrho, rss)
1.1.1.dev31+g9ff89ff6b

In [2]: %%timeit
   ...: pfss_out = pfss.pfss(pfss_in)
   ...: 
   ...: 
1.61 s ± 679 ms per loop (mean ± std. dev. of 7 runs, 1 loop each)

And with sunkit_magex 1.1.0 with current Numba implementation

(py312) sbadman@si393685:~$ ipython
Python 3.12.3 | packaged by conda-forge | (main, Apr 15 2024, 18:38:13) [GCC 12.3.0]
Type 'copyright', 'credits' or 'license' for more information
IPython 8.25.0 -- An enhanced Interactive Python. Type '?' for help.

In [1]: import sunkit_magex; print(sunkit_magex.__version__)
   ...: 
   ...: import sunpy.map
   ...: from sunkit_magex import pfss
   ...: 
   ...: gong_fname = pfss.sample_data.get_gong_map()
   ...: gong_map = sunpy.map.Map(gong_fname)
   ...: 
   ...: nrho = 35
   ...: rss = 2.5
   ...: pfss_in = pfss.Input(gong_map, nrho, rss)
1.1.0

In [2]: %%timeit
   ...: pfss_out = pfss.pfss(pfss_in)
   ...: 
   ...: 
879 ms ± 87 ms per loop (mean ± std. dev. of 7 runs, 1 loop each)