Recalculate halo mass function test values

[rrjbca]

#445 fixed a bug in the calculation of the Eisenstein & Hu matter power spectrum. This fix has now been merged into module/halos and is causing multiple unit tests and doctests to fail since they are calculating mass funtions and comparing them against values precomputed using the old bugged power spectrum module.

The following precomputed values need to be recomputed and updated:

• Docstring for halo_mass_function
• Docstring for ellipsoidal_collapse_function
• fsigma and massf arrays in test_mass.py

Test failures can be seen here: https://github.com/skypyproject/skypy/actions/runs/4032155320