In certain cases the hard-coded value causes the algorithm to fail for
thin annulus samples. Using a parameter allows this to be controlled
by the user. The default value has been increased so that most users
will not need to know about it anyway.
Refs #20532

Re #20540

Re #20404

...and other generic cleanup of MatrixWorkspace to reduce its
dependence on blocksize()

This algorithm is specific to NR and relates to the slit resolution

Re #18842

Re #20404

Moved algorithm class private methods to a private namespace.

Re #20404

ThetaIn was still being used instead of theta from the log when correcting detectors

Re #20397

- Add a new property, ThetaInLogName, to ReflectometryReductionOneAuto, which is used to find ThetaIn from the given log value.
- Note that if the ThetaIn property is also specified, that will take precedence.
- Add another new property, CorrectDetectors, to make it optional whether detectors are moved to ThetaIn (whether specified via ThetaIn or ThetaInLogValue). The default is true to maintain the previous behaviour when using ThetaIn.
- Pass through the overridden ThetaIn value to ReflectometryReductionOne. If ThetaIn is not specified, theta is calculated using SpecularReflectionCalculateTheta, and this is also now passed through for consistency. The latter does not work for some IDF setups but in those cases ThetaIn should be overridden with one of the above properties (and this no longer needs to affect the detector positions if CorrectDetectors is set to false).
- Disable the DetectorCorrectionType drop-down if CorrectDetectors is false.
- Disable the ReductionType drop-down if SummationType is SumInLambda.
- Improve an error message to report angles in degrees rather than radians.

Re #20397

Jacobian.h moved from CurveFitting to API and renamed to BasicJacobian
since a Jacobian implementation is needed by
CalculatePolynomialBackground.

Re #20404

- This commit fixes a bug where summation in Q was giving incorrect
results for point detectors due to an incorrect calculation of the pixel
size.
- Summation is not really applicable for point detectors so we could have disabled it. However, it is useful
for debug purposes.

Re #20399