Externalize discretized function interpolation implementations

[Kevin Milner]

Currently lots of interpolation code is duplicated between discretized function implementations. Generic implementations should be externalized to a utility class.

[Peter Powers]

see Interpolate.java. Methods of class will extrapolate, but this may not be what is desired in all cases. No data checking is done either.

[Kevin Milner]

I removed all of the duplicate code by creating abstract implementations in AbstractDiscretizedFunc? which pass through to the following methods:

org.opensha.commons.data.function.AbstractDiscretizedFunc?.getInterpolatedY(double, boolean, boolean)

org.opensha.commons.data.function.AbstractDiscretizedFunc?.getFirstInterpolatedX(double, boolean, boolean)

These methods, in turn, use Interpolate.java for the actual implementation, but don't defer fully to that class as turning the functions to arrays of x/y values would be computationally expensive.

Done in [10918], not closing until it is merged back to trunk. We have a pretty complete set of unit tests for these functions, and this new implementation passes those tests.

[Kevin Milner]

Fixes/updates merged to trunk in [10931], closing.