Save single-crystal shoebox/skew integration on peak shape for subsequent overlap detection and two-step integration

[RichardWaiteSTFC]

Is your feature request related to a problem? Please describe.
Q-space integration methods save the peak shape (sphere or ellipsoid) on the peak.

For example PeakShapeEllipsoid here

mantid/Framework/DataObjects/src/PeakShapeEllipsoid.cpp

Line 16 in b7e0aaf

PeakShapeEllipsoid::PeakShapeEllipsoid(const std::vector<Kernel::V3D> &directions, const std::vector<double> &abcRadii,

Is set on each peak during the Q-space integration here

mantid/Framework/MDAlgorithms/src/IntegratePeaksMD2.cpp

Lines 621 to 624 in 94f78bf

	PeakShape *ellipsoidShape =
	new PeakShapeEllipsoid(eigenvects, peakRadii, backgroundInnerRadii, backgroundOuterRadii,
	CoordinatesToUse, this->name(), this->version(), translation);
	p.setPeakShape(ellipsoidShape);

This peak shape is currently used only for visualisation purposes - but could have other uses.

Describe the solution you'd like
It would be good to save the detector IDs and TOF bin indices in a similar shape e.g. DetectorTOFShape this could be used for:

• Overlap detection post integration (peaks could be excluded or flagged as combined)
• Two-step integration - re-integrate weak peaks using strong peak dimensions
• Eventual visualisation on instrument view (i.e. show bounding box of peak)

[RichardWaiteSTFC]

Integration algorithms that require peak shape to be saved on the Peak object:

• IntegratePeaksSkew
• IntegratePeaksShoeboxTOF
• IntegratePeaks1DProfile

[RichardWaiteSTFC]

Please leave IntegratePeaks1DProfile until after big refactor in #38439

[warunawickramasingha]

Reopening the issue since the merged PR #38433 was meant to partially fix this issue and another PR linked to the issue is still in progress.