Fixes & improvements on HSGP

Closes #7240

Working on this in concert with @bwengals for [this example](https://github.com/pymc-devs/pymc-examples/pull/

[x] Fix the bug in the computation of L
[x] Do mean-centering of X values under the hood, instead of requiring it from users
[x] Enable setting prior_linearized with m and c instead of only m and L
[x] parametrization is not documented
[x] Add option to pass in nicely named coords to HSGP for the beta coefficients
[x] Make ._m_star more accessible / user friendly (make a property with no leading underscore so existing code doesn't break), document it, and fix example in prior_linearized docstring.
[x] Get rid of numpy option, tl=np. It's subtly wrong when calculating the eigen stuff and its not used anywhere.
[x] Add the approx_params function to hsgp_approx file

Type of change