GenericAbsorptionCorrection to allow a Gauge Volume to be defined

[Nick Draper]

Allow another shape, a gauge volume to be defined, only scatering in this area needs to be considered and accepted, but the whole shape needs to be considered for path lengths.

We should create another version of CreateSampleShape to define the guage volume and shape, and store this along with the sample object.

[Russell Taylor]

(In [6671]) Add the concept of a gauge volume to the generic Absorption Correction algorithm. Re #1688. Some other minor tidying up.

[Russell Taylor]

(In [6865]) Generalise Cylinder & FlatPlate absorption algorithms to use the sample position and not just assume the sample is at the origin. Re #1688.

[Russell Taylor]

(In [7139]) Add 'CuboidGaugeVolume' absorption correction algorithm. Re #1688.

[Russell Taylor]

(In [7161]) Check properly for sample object not being set. Re #1688.

[Russell Taylor]

So I've actually done this in 2 ways, the one described above which requires running CreateSampleShape and DefineGaugeVolume to define the sample and active volumes respectively, then AbsorptionCorrection.

After doing this, I discovered that the gauge volume is always a cuboid for the instruments of interest so I realised that this could be done an easier way making use of the FlatPlateAbsorption code. So now there's a CuboidGaugeVolumeAbsorption algorithm that requires the sample to have been defined using CreateSampleShape and then takes in parameters defining the size of the gauge volume a la FlatPlateAbsorption. This was very last minute though, so it should be tested carefully.

[Martyn Gigg]

I've run into a crash with CuboidGaugeVolumeAbsorption correction. Running the attached script I get a crash which when I ran in debug showed up something wrong in FlatPlateAbsorption. Given nothing happens in CuboidGaugeVolumeAbsorption however this is not surprising I guess.

[Russell Taylor]

(In [7184]) Add protection against the gauge volume not being fully within the sample. Re #1688.

[Russell Taylor]

(In [7186]) Some clarifying comments. Re #1688.

[Russell Taylor]

The attached script has a bug in it - the variable slabWidthInCM was being passed twice in the call to CuboidGaugeVolumeAbsorption (and the slab thickness not at all). Still, this highlighted problems if the gauge volume was not fully enclosed by the sample and this is now protected against.

[Martyn Gigg]

I've corrected the script.

Fiddling with numbers in the script gives the expected agreements/disagreements. CylinderAbsorption tested in isolation and gives sensible results.

[Stuart Campbell]

This ticket has been transferred to github issue ​2535