… module

… tests

…t matrix log, exp, and sqrt

…actly symmetric

Before: the primal output was computed in-line in the custom jvp
function. The advantage is that this avoids a second call to the
eigendecomposition. The downside is that this in-line computation
doesn't itself have a custom jvp, so its derivative can be wrong.

After: I re-compute the primal value through the base function
(e.g., log_symm), which has the custom jvp defined on it. The
eigendecomposition is repeated. We can refactor to eliminate this
later if profiling reveals it to be a performance bottleneck.

Put a leading underscore on functions menat for internal use.
Most Python tools will ignore these when reporting contents of
a module.

Replace all calls except in the new viscoelastic model. Changes are
about to merge there and I want to handle the conflicts separately.

Also: now a ValueError is emitted if a user asks for a quadrature
rule precision that is above what we have implemented. Before
the error was a bit obscure.

Improved the unit testing of the quadrature rules by making
subtests for every polynomial degree check. This way, if the test
fails, you can figure out which quadrature rules are broken.