update perturbation example

[karlwessel]

This updates the perturbation example to use a more recent feature set of Symbolics.jl. Namely

• Symbolics.coeff replaces most of the code in collect_powers and
• symbolic_linear_solve is used instead of solve_for.

It also fixes some minor problems:

• the link to the perturbation example on differential equations was dead
• and the code on how to order the solution of the Kepler-equation by orders of $M$ was wrong.

If #1291 is fixed, one could also replace solve_coef completely with symbolic_solve using Groebner.

[karlwessel]

The failing integration tests don't seem to be related to the PR.

[ChrisRackauckas]

Oh sorry I thought I already merged this.