a simpler / better shock detection algorithm

[zingale]

PR summary

The original algorithm was based on one in Flash but it had some normalizations in it that did not make sense. This now looks at

| (grad P - rho g) . v / |v| | / (P / dx) > eps div{U} < 0

This projects grad P in the direction of the velocity (v / |v|) and also removes the hydrostatic pressure, since only the pressure in excess of that is available to make a shock.

Here eps defaults to 2/3 and can be set at runtime via castro.shock_detection_threshold

PR motivation

PR checklist

• [ ] test suite needs to be run on this PR
• [ ] this PR will change answers in the test suite to more than roundoff level
• [ ] all newly-added functions have docstrings as per the coding conventions
• [ ] the CHANGES file has been updated, if appropriate
• [ ] if appropriate, this change is described in the docs

[zingale]

Here's the difference in the old (left) vs new (right) shock detection for Sedov

[zingale]

I still think that the shock region is too thick -- it should only be a single zone

[zingale]

Here's an updated comparison, the left is the latest version, which projects the gradient in the direction of motion.